585
DISCONNEXION SYNDROMES IN ANIMALS AND MAN1
BY
NORMAN GESCHWIND
PART II
{From the Aphasia Research Section, Neurology Service, Boston Veterans Administration
Hospital and the Department of Neurology, Boston University Medical School)
TABLE OF CONTENTS
(Sections I-in are in Part I)
PART II
Introduction
IV. The Agnosias
The problem of confabulatory response
Inability to identify colours
Classical visual agnosia
The lesions of classical visual agnosia
The handling of objects
The conditions for confabulatory responses
The problem of right parietal dominance
Visual imagery
V. The Mechanisms of the Apraxias
Disconnexion from the speech area
Extension of the theory of the apraxias
The apraxias of the supramarginal gyrus region
Facial apraxia
Whole body movements
Other bilateral movements
The problem of "motor" versus "cognitive" learning
VI. Other Aphasic Disturbances
Conduction aphasia
The case of Bonhoeffer
Echolalia
VII. Possible Objections andPitfalls
The results of Akelaitis and his co-workers
Vin. Philosophical Implications
The whole man
The unity of consciousness
The value of introspection
Language and thought
Summary
Bibliography
1
This work was supported in part by a grant (MH 084721, Professor Davis Howes,
Principal Investigator) from the National Institutes of Health to the Boston University
Medical School.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
586 NORMAN GESCHWIND
INTRODUCTION
THE first part of this paper (Geschwind, 1965) defined disconnexion
syndromes and discussed the anatomical bases of such disorders. Particular
emphasis was placed on the pattern of cortico-cortical connexions in
primates, and the disturbances resulting from lesions of these connexions.
This was followed by a discussion of possible developments in the pattern
of cortico-cortical connexions which had favoured the development of
language. Some examples were given of syndromes of disconnexion in
man. Part II will be devoted primarily to a discussion of some of the
disturbances of the higher functions in man.
IV. THE AGNOSIAS
The agnosias have always represented an unusually difficult group
of disturbances. Classically they were defined as disturbances of
' 'recognition'' without disturbances of elementary sensation. Someobservers
have simply denied the existence of disturbances of this type. There has
been the attempt on one side to reduce them to perceptual defects of
complex nature, such as continuously shifting thresholds. On the other
hand there has been the tendency to deal with such disturbances in terms of
rather complex psychological malfunctions such as loss of ability to
perceive Gestalts. Amore classical approach was that ofLissauer(1889)who
divided agnosias into apperceptive and associative, or disorders of primary
and secondary recognition. The theory that I will present here will
propose that the agnosias are indeed associative disorders; what will be
added to the classical notion will be an analysis of just what type of
associative disorder is involved and an explanation of certain clinical
features that have in the past acted as stumbling blocks to analysis of these
disturbances. I have already discussed in detail the agnosias of animals.
The human cases are also based on disconnexion but I hope to show that
the mechanisms are different in important ways from those operative in
animals.
Workers in this field have tended to use a certain set of criteria for calling
a disturbance an agnosia. First it is necessary to show either (1) that
elementary sensation is intact or (2) that if an impairment of elementary
sensation is present it is not sufficiently severe to explain the disturbance of
recognition. Assuming this to have been demonstrated, how has it been
possible to demonstrate that the failure of recognition was not simply a
failure of naming? There have been several criteria generally employed. I
will list them first and discuss their usefulness afterwards:
(1) The "agnosic" disturbance is a circumscribed one.—There is a common
view that aphasic misnaming cannot be confined to a single modality;
even more strongly it is assumed that it cannot be confined to a class within
a modality. From this point of view an aphasic misnames everything; by
contrast there can be a visual agnosia or even an agnosia for colours alone.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 587
(2) "Agnosic" errors differ in character from "aphasic" errors.—It is
often assumed that aphasic errors are of one type, agnosic of another; the
aphasic error is a "linguistic" one, the agnosic a "perceptual" one.
According to this view, "chair" for "table" is an aphasic error but "a small
package" for "a book" is an agnosic error; "something to eat with" for
"spoon" is an aphasic error but "pencil" for "knife" is an agnosic error.
(3) The "agnosic" cannot choose the correct response.—The aphasic who
has failed to name is usually described as being readily able to accept the
correct name when it is offered. The "agnosic" by contrast is said to be
incapable of selecting the correct name.
(4) The "agnosic" describes his perceptual difficulty.—The aphasic is
usually pictured as saying, "I know what it is but I can't find the name."
When asked about the appearances of objects he asserts that they appear
normal to him. The "agnosic," by contrast, informs the examiner that
things look different to him or that he has difficulty in seeing clearly.
Alternatively he may offer dramatic descriptions of bizarre perceptions,
thus apparently providing direct evidence that he is experiencing
distortions of his perceptions.
(5) The "agnosic" cannot show the use of the object.—This criterion is
actually the most complex. The aphasic is said to be able to show the use
of the object. The "visual agnosic" cannot show the use of the object but
can manipulate it correctly if it is actually placed into his hand. The
"apraxic" according to this classical analysis cannot handle the object
correctly even when it is placed in his hand.
As I hope to demonstrate, all of these criteria are actually inadequate for
the separation of patients into two classes of distinct physiological disturbance.
The fundamental difficulty has been in the acceptance of a special
class of defects of "recognition," lying somewhere between defects of
"perception" and of "naming." What indeed are the criteria for "recognition"
and is it a single function? I believe in fact that there is no single
faculty of "recognition" but that the term covers the totality of all the
associations aroused by any object. Phrased in another way, we "manifest
recognition" by responding appropriately; to the extent that any appropriate
response occurs, we have shown "recognition." But this view
abolishes the notion of a unitary step of "recognition"; instead, there are
multiple parallel processes of appropriate response to a stimulus. To
describe the behaviour correctly we must describe the pattern of loss and
preservation of responses to each particular type of stimulus.
Let us make this more concrete by referring to a particular patient's
behaviour. I will return for my example to the case of Geschwind and
Kaplan(1962)to which I have several times made reference. Let us consider
his difficulty in identifying objects placed in the left hand. He incorrectly
named objects placed in the left hand. He could, however, draw the object
39 BRAIN—VOL. Lxxxvin
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
588 NORMAN GESCHWIND
afterwards with his left hand or select it from a group tactilely or visually.
It is obviously correct to describe this patient as showing correct identification
by nonverbal means and incorrect identification verbally of objects
held in the left hand. It is difficult to see how this could in any way be
described as a "perceptual" disorder since perception must have been
intact to insure correct selection from a group or the drawing of the object.
We have interpreted this disturbance as a result of disconnexion from
the speech area. In addition the patient showed other disturbances of
identification not due to disconnexion from the speech area. Thus, he could
not select or draw with the right hand an object held with the left, concealed
from vision, or select or draw with the left hand an object held with
the right. But again "perceptual" disturbance in either hand is excluded
by the patient's ability to select or draw with a given hand an object that
had been held in the same hand. We can only regard the defect as associative—what
was perceived by the right hemisphere could not be relayed
to the left and vice versa.
Let us return for a moment to the failure to identify verbally with the
left hand objects held in that hand which we have attributed to disconnexion
from the speech area. It may be argued that we have evaded the
fundamental issue, that although we have shown that the patient could not
name the object placed in his left hand but yet could draw it or select it
afterwards from a group, we have not really shown whether the patient
had recognized the object. This question is based on two assumptions
which I believe are incorrect. The first assumption is that "the patient" is
an entity. But in someone whose two hemispheres operate as independently
as this patient's the word loses its ordinary meaning. There is no answer to
"Did the patient recognize?" There are, however, clear answers to "Under
what conditions did the right or the left hemisphere recognize?"
The second assumption is that recognition is more than the sum of the
individual acts of naming, drawing, handling, selecting, etc. If it is more
than these, how would one test for this faculty and how would one show
that this test established some superordinate function of recognition ? I do
not know of an answer to this question.
I have argued that the patient's failure to identify verbally objects held
in the left hand must be regarded as a result of separation of the right
somesthetic region from the speech area and can in no sense be regarded as
a perceptual defect. In the light of this analysis, let us now consider the
patient's responses from the point of view of the classical criteria listed
above which were used to distinguish "agnosic" from "aphasic" disturbances.
This analysis will demonstrate the inapplicability of these criteria.
(1) Our patient's disturbance was highly circumscribed to one modality
and indeed one-half of the body. Classically such circumscription was
often regarded as being incompatible with an aphasia. Yet as soon as one
admits the substantial restriction of speech to one hemisphere, the
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 589
possibility of cutting off one-half of the brain from speech must be
considered. There is no basis therefore for this classical distinction.
(2) The errors made by our patient were not "errors of naming" in the
sense of the commonly used criteria. He could call a "screw-driver" a
"rubber band," a "coin" a "cigarette lighter." These are not (without
meaningless extension of the term) errors in sphere nor are they errors
based on similarity of sound. They are, therefore, not "aphasic" errors
according to the usual criterion. But it is easy to see that if a particular
part of the brain is fully disconnected from the speech area, there is no
reason that an erroneous name should be related in any way to the correct
term. The more complete the disconnexion, the more random the errors
must be. The less complete the disconnexion, i.e. if some information gets
through to the speech area, the more the errors may be related to the
correct response. It should be added that in fact the classically anomic
patient often makes naming errors not related by sphere or sound.
(3) Our patient showed inability to choose the correct verbal response when
it was offered to him.—When he misnamed an object we would offer him.
several choices. The patient would generally not select the correct answer
from the group offered. If, however, the speech area is fully disconnected
from the right hemisphere, there is no reason why the speech area should
select the correct term. It is obvious that the ability to select the correct
word depends on there being some connexion between the site of perception
and the speech area. The ability to select the correct word from a
group when the patient cannot find it spontaneously indicates a lesser
degree of disconnexion. Phrased in another, diagrammatic way, one could
conceive that when one offers a word to a patient he in some way compares
the images or memories aroused by this word with the sensations he is
receiving from the object. If there is no site for such comparison, the
patient will not recognize the correct word when it is offered. I would like
to point out that in fact the obvious nominal aphasic often fails to accept
the correct word when it is offered. Furthermore, even when the patient
insists that "I know what it is but I can't tell you the name" he may fail to
accept the correct name from a group.
(4) The patient gave descriptions of his "perceptions."—Thus, when holding
a half-dollar, the patient said that it was a cigarette lighter. When
asked if it had corners he said, "Oh, yes, there's one, there's another." He
would describe incorrectly the object held in his hand in most instances.
The same objection applies here as to the previous categories. If the
patient's speech area is disconnected from a site of primary perception why
should his speech area be able to describe what is going on at the site of
the primary perception? Clearly, it should not. The fact that "the patient"
(i.e. the speech area) gives a description does not mean that we are getting
an actual description of the perceptions going on in another disconnected
part of the brain. We must again remember that we are dealing with more
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
590 NORMAN GESCHWIND
than one "patient" here. The "patient" who speaks to you is not the
""patient" who is perceiving—they are, in fact, separate.
(5) Our patient lacked only one of the classical criteria of "agnosia."
According to these classical criteria, a patient with "tactile agnosia" should
mishandle objects which he has not seen but should respond correctly to
the sight of the object (i.e. his behaviour should be opposite to that of the
visual agnosic). Our patient in fact handled objects perfectly correctly
when blindfolded. It will return in a later part of my discussion of the
agnosias to this question of the use of objects.
The Problem of Confabulatory Response
I have discussed in some detail a disturbance which might have been
called "agnosic" and tried to show that this disturbance is much more
meaningfully described as a naming defect resulting from disconnexion
from the speech area. There are certain further implications of this interpretation
which I would like to present now. One most important implication
is that the "introspections" of the patient as to his disability may be of
little or no use to the examiner. The patient cannot "introspect" about the
-activities of a piece of brain which has no connexion to the speech area.
What he tells you is of little value in elucidating the mechanism and may
indeed be actively misleading. Indeed, it becomes clear that many of the
patient's responses can only be described as confabulatory, i.e. they are
attempts to fill gaps in the information available to his speech area; phrased
in more conventional terms they are attempts to explain what the patient
•cannot understand.1
These confabulatory responses have been a major source of confusion
for many years, particularly in the assessment of difficulties of sensory
identification. It is curious that sophistication concerning confabulatory
ibehaviour has been much more widely developed in certain areas than
others. When the patient with a severe recent memory defect asserts that
the examiner spent the previous evening with him in a bar, we neither ask
ourselves why the patient had hallucinations of seeing the examiner the
night before nor why he is now having false memories. We see this instead
as the verbal fining in of a gap. We discount the attribution by the same
patient of his lack of knowledge of current events to the fact that he has
never been interested in politics. Similarly, if an aphasic patient tells us that
Ms trouble in speaking is the result of his ill-fitting dentures or a sore throat,
-we do not take the attitude that he is giving us any very useful insight into
tis illness but rather that he too is attempting to explain a gap in his performance.
An even more extreme example is that of patients with denial of
1
Dr Edwin Weinstein and his co-workers have been instrumental in stressing in
recent years in a series of careful papers the importance of confabulatory responses.
The stimulation of his work was a major factor in attracting my attention to this
problem.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 591
blindness who may be prepared to describe in detail many of the objects in
the room; once we know that the patient is blind, we do not attempt tomake
use of his "introspections" about his visual perceptions. We do not
take seriously this patient's protestations that he has failed in some task
because the light is poor or he is not wearing his glasses. Furthermore, we
do not assert as an alternative that the patient is really hallucinating.
The "higher" the function impaired, the more readily the patient may
succeed in fooling others with confabulatory responses. I recall one patient
with pure alexia without agraphia who asserted that his trouble in reading
was due to his "blindness in his right eye." The resident staff correctly
realized that he meant a right visual field defect when he spoke of a blind
right eye. However, they incorrectly accepted the patient's attribution of
his reading disturbance to this visual impairment. Only when the
discrepancybetweenreading and object-naming was pointed out to them did
they realize that they had too readily accepted the patient's "explanation"
of his disability.
There are many occasions in which patients with aphasia, dementia,
blindness, or recent memory defects produce confabulatory responses
which are recognized. There are probably equally numerous situations in
which such confabulatory responses are not recognized as such by the
examiner. Most of the conditions called "agnosias" fall into this category.
Much time has been spent in the vain attempt to analyse the supposed
disorder of perception which has been regarded as underlying the patient's
responses.
In summary, the theory here proposed is that most of the classical
agnosias are highly isolated disturbances of naming which are the result of
disconnexions from the speech area. The prominence of confabulatory
responses may incorrectly lead the examiner to believe that he is dealing
with a perceptual disorder.
Inability to Identify Colours
The patient of Geschwind and Fusillo (1964) (discussed in an earlier
section of the paper) showed pure alexia without agraphia and a failure to
identify colours. This disturbance could readily have been called an
"agnosia for colours." Yet, analysis of his responses makes it clear that his
difficulty could be correctly described as a disturbance in the naming of
colours. By all nonverbal criteria he identified colours correctly—he
matched and sorted correctly, did the Ishihara test perfectly and even
properly matched colours to pictures of objects, yet he failed consistently
in verbal naming. It is obviously appropriate to describe his disturbance
as a colour-naming defect since by no criterion could he be shown to have
a perceptual defect. His pattern of responses also illustrates the lack of
usefulness of the classical criteria. Thus he exhibited a very circumscribed
naming defect. Only once in many testing sessions did he misname an
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
592 NORMAN GESCHWIND
object, yet his colour-naming remained consistently extremely poor. He
generally failed to accept the correct name of the colour when it was
offered. Indeed, when we told him, for example, that something was not
grey but red, he said, "Well, it looks grey to me—maybe it's a reddishgrey."
On another occasion he said, "You say it's white but it looks tan
to me." Again the most parsimonious assumption is that he had a colournaming
defect with confabulatory response. Whenever we asked the
patient to give a verbal account of the colour experiences of his right
hemisphere, he produced a confabulatory response; if we tested the right
hemisphere nonverbally, we got evidence of perfectly normal colour
perception.
I would like to discuss the one instance of poor matching by this patient
since the careful analysis of this failure was most illuminating. The
patient was given a pile of colour chips which contained chips of two
different shades of each of the primary colours. He was shown a chip of
one of the two shades of red and told, "This is red. Now pick up all the red
chips." He picked each chip up slowly and examined it. He finally
separated out all of the chips of the selected colour as well as all the chips
of one of the two shades of green. When asked what colour the chips
were which he had selected he said, "These are red and these others
(indicating the green ones) are a different shade of red." He was then able
to separate these chips into two piles. When shown a chip of the other
shade of green and asked which of the two piles it was most like, he unhesitatingly
classified it with the pile of green chips.
There was clearly a sorting error here but full consideration of the
patient's behaviour seems to exclude the possibility that we are dealing with
a perceptual error. I believe the sequence of events to have been as follows.
We had asked the patient to "pick up all the red chips," rather than to
"pick up all the chips like this one." He had done the task slowly probably
because he did it by picking up each chip and naming it. Having misnamed
a chip of one of the shades of green as red, he had then selected the
remaining chips of this shade. We may consider this as a sorting error
resulting not from failuie of discrimination or perception (which his subsequent
behaviour in this task excluded, i.e. the rapid separation of the two
colours) but as a "secondary sorting error" which resulted from doing the
task verbally. Thus, even failures in sorting should be analysed carefully to
determine whether they are secondary to misnaming or are truly failures of
discrimination.1
Classical Visual Agnosia
There are many remarkable accounts of this disturbance in the literature
(e.g. Lissauer, 1889; Lange, 1936; Brain, 1941; Macrae and Trolle, 1956;
"^Pick (1931) pointed out that sorting errors may result from difficulties in verbal
mediation in patients with colour-naming troubles.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 593
Ettlinger and Wyke, 1961; Hecaen and Angelergues, 1963). In going
through these accounts one is immediately struck by the frequent evidence
of preservation of nonverbal visual identification while verbal identification
is impaired. One striking feature of many of these cases is the fact that the
patient's ordinary behaviour is in marked contrast to the supposed perceptual
disturbance. Thus, one may read an account of a patient who cannot
"identify" a glass of water and yet a few minutes later picks it up and
drinks from it. Much stress has been laid on the fact that a patient may
show normal behaviour in a "natural" situation but not in an "artificial"
one. Once one abandons the notion of a unitary process of "recognition,"
it is easy to see that the patient may fail to "identify" an object, i.e. fail to
give a verbal account of what is going on in his visual receptive regions and
yet be able to respond nonverbally to a nonverbal stimulus. Most of the
pathways involved in the nonverbal response may be quite different from
those involved in naming. For a thirsty man to respond appropriately to
the sight of a glass of water by grasping it and drinking from it is anatomically
a different task from that of responding to a verbal request for
identification. It is certainly reasonable that a naming defect should not
produce gross functional impairment; by contrast it would be very difficult
to understand how a "higher-order perceptual defect" could leave ordinary
perceptual function unimpaired. A similar preservation of function appears
in those cases who can draw the object which they cannot recognize (Lange,
1936) a behaviour obviously compatible with a naming difficulty.
The patient of Ettlinger and Wyke (1961) made many errors in visual
naming (e.g. "a box" for "a glove," "for cooking" for "a spoon") and
many fewer in tactile naming. This case can be used as further evidence
against the idea that an aphasic naming disorder must cut across modalities.
The data of these authors show that while "agnosic" errors (e.g. "box" for
"glove") were more common visually than tactilely, the same was true for
"aphasic" errors (e.g. "for cooking" for "a spoon"). A striking feature of
this case was the matching of outline shapes without error even though the
patient made nine errors in naming outline pictures of ten objects.
Certainly a naming defect is the simplest explanation of this disorder.
This case was similar in this respect to the earlier case of Brain (1941).
The remarkable case of Macrae and Trolle (1956) presents another
example of one of these striking discrepancies. The authors point out that
when the patient was pressed with regard to his method of finding his way
to and from work, "he could not visualize the street plan or the route."
Despite this he did in fact drive two miles a day to work without difficulty.
I would suggest that "visualization" here means either developing or describing
an internal picture in response to a verbal stimulus—but this will
not occur if the visual system is separated from the speech area. Yet, the
subject may respond nonverbally to the actual route.
Patients with classical visual agnosias have often been presumed to be
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
594 NORMAN GESCHWIND
suffering from "perceptual" disorder on the basis of their "introspections"
concerning their illness. The case of Lissauer (1889) illustrates how little
such "introspections" are to be trusted. Thus, Lissauer comments that
when the patient misidentified an object, he would, after learning its true
nature, excuse himself by saying that "his eyes had deceived him at first."
As Lissauer comments, this statement, coupled withthe fact that some of the
misidentifications were so curious, might have led the examiner to believe
that the patient had been having sensory illusions of some type. But if the
examiner then asked the patient whether the object, now that he knew
what it was, looked any different from the way it had before, he received a
vigorous denial. The patient in fact now asserted that the object had
looked the same before but that "his poor memory was responsible for the
fact that it had not occurred to him what the thing was called and what it
was." Had Lissauer pushed his analysis somewhat further here he might
well have been able to see that the "misidentification" was actually only
misnaming. Lissauer himself comments elsewhere in the paper, "The
patient succeeded . . . in drawing simple objects which he didn't recognize,
a striking proof that he was well able to perceive their forms." Similarly,
although the patient failed to name colours, he succeeded without any
difficulty in matching tasks.
The Lesions of Classical Visual Agnosia
At first it would be tempting to ascribe classical visual agnosia simply to
disconnexion of the visual region from the speech area. This interpretation,
however, runs into obvious difficulties. Thus, we have already explained in
an earlier section that pure alexia without agraphia (with its frequent
accompaniment of colour-naming difBculty) is the result of disconnexion
of the visual cortex from the speech area. In these cases there is usually
destruction of the left calcarine cortex. In addition there is a lesion of the
splenium of the corpus callosum which isolates the right visual region from
the speech area. It was particularly noted that these patients did not show
difficulties in object-naming and often showed retained ability to read
numbers. I said in the earlier discussion that it seemed likely that the
explanation for the preservation of object naming was that objects could
arouse somesthetic and other associations more anteriorly in the right
hemisphere; these would then traverse the callosum in the preserved
portions anterior to the splenium and thus reach the speech area. For the
development of a naming defect for objects, failure to arouse these tactile
associations (a visual-tactile disconnexion) is a necessary addition to the
disconnexion of visual regions from the speech area.
The lesions found in the majority of cases of visual agnosia support this
notion. Most commonly observed are very large bilateral parieto-occipital
lesions. There are probably no direct connexions between the visual and
the somesthetic cortex so that visual-somesthetic connexions must be
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 595
mediated via the intervening association cortex. A large parieto-occipital
lesion would effectively cut off such connexions. One can also see the
reason for the rarity of classical visual "agnosia." If the lesion is too
small, a large enough extent of association cortex is left to make the defect
a minor one. If the lesions are too large, the patient may show gross
aphasic impairment as well as marked disturbances of visual-motor connexions.
In this case his impairment of speech function may make analysis
exceedingly difficult. The cases with very extensive lesions might with some
justification be regarded as manifesting disturbances of the general ability
to manifest visual "recognition" since all outflow channels may be cut off
except possibly some primitive and not very extensive pathways descending
subcortically.
There is another but much smaller group of visual "agnosias" who have
unilateral lesions. This group shows destruction of the left visual cortex.
In addition there are always lesions of other structures. One group of
these patients shows involvement of the splenium. Others have such extensive
involvement of the white matter or of the association cortex in the left
occipital and parietal lobes as to have destroyed either the fibres coming
from the splenium or their terminations in the left hemisphere (Lange,
1936; Nielsen, 1937).i The beautifully described case of Lissauer (1889)
turned outto have such a lesion (Hahn, cited by Meyer, 1895-1896) although
he himself had expected bilateral lesions.
Why do these cases with unilateral lesions manifest more than pure
alexia? It is difficult to be certain and I would hope that more careful
study of cases in the future may provide the answer. One possibility is that
in some people visual stimuli in the right hemisphere fail to arouse tactile
associations or do so weakly. These patients should therefore develop a
more extensive naming defect when the right visual regions are cut off from
the speech area. Another possibility is that in these cases the lesion in the
1
A brief note is appropriate here on Nielsen's interesting and important review. In
his introduction he points out that many authors have stressed the bilaterality of lesions
in most cases of this disorder. He then goes on in his study to confine himself only to
the unilateral cases on the grounds that "if in any case a unilateral lesion produced
visual agnosia, there can be no point in citing cases of bilateral lesion of the occipital
lobes." This argument is of course based on the reasoning that a bilateral case must
only be a unilateral case with an extra unnecessary lesion. The fact that probably the
majority of cases of visual "agnosia" have had bilateral lesions militates against this
thesis. Furthermore it should always be kept in mind that many different combinations
of lesions may lead to the same effect when the physiological mechanisms are kept in
mind. He also neglects the possibility of individual differences. The paper of Nielsen
is therefore a contribution to the anatomy of the unilateral group of cases but not
necessarily to that of the bilateral cases. Many of Nielsen's cases show the discrepancies
between naming and other functions that I have cited above although his interpretation
of much of the material is different from mine because he accepts the idea of
defects of "recognition." He frequently uses as one of his criteria the failure of the
patient to behave correctly in relation to the "unrecognized" stimuli; at the end of this
section on the agnosias I will discuss this problem briefly.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
596 NORMAN GESCHWIND
splenium extends further forward than it does in the cases of pure alexia
without agraphia. The lesion would thus cut off not only fibres from the
visual association cortex but also fibres from the somesthetic association
cortex. In the case of Trescher and Ford (1937) the surgeon is said to have
divided the posterior half of the corpus callosum; it seems likely that the
lesion must have involved more than the splenium. The description of the
case would suggest that the patient in addition to being alexic in the left
half-field, also could not name objects in that field. This case therefore
provides suggestive evidence for the hypothesis that lesions extending
somewhat anterior to the splenium may cut off impulses from the somesthetic
association cortex and thus prevent visual stimuli from reaching the
speech area via this detour.
The Handling of Objects
It was pointed out earlier that one of the classical criteria for a visual
agnosia was failure of the patient to demonstrate the use of objects shown
to him. It was pointed out that according to this scheme the "aphasic"
could show the use of an object shown to him; the "visual agnosic" could
not do this but could manipulate the object correctly if it was placed in his
hand; the "apraxic" could not perform even with the object in his hand.
I have reserved this criterion for special discussion since it appears
at first glance to provide a very reasonable basis for separating disturbances
of recognition.
This criterion, however, also fails as the mark of a separate category of
"recognition" disturbances. In the first place, as I have already commented,
some patients who are severely "agnosic" by the other classical
criteria listed earlier can in fact respond correctly to the sight of objects. If
one simply interpreted agnosias as deficits of "recognition" there should be
no reason why correct demonstration of the use of a seen object should be
preserved in the presence of a severe disturbance of this type. Furthermore,
as I will show in the section on motor disturbances, there are severely
apraxic patients who fail to respond to the sight of an object yet handle the
actual object correctly when it is placed in the hand. By the classical
criteria these patients should be called agnosic. Yet these same patients
may be able to name the object and describe its use, behaviour certainly
incompatible with a loss of "recognition."
As I will point out in the section on the apraxias, they are also the result
of disconnexions. The lesions producing these disturbances separate
sensoryfrommotorregions. Itisobvious that the kindof lesionwhichcauses
classical visual agnosia will frequently although not invariably result
in the disconnexion of the visual regions not only from the speech area and
from the somesthetic areas but also from the motor region.
This discussion illustrates again the principle that division of phenomena
into aphasic, agnosic or apraxic often obscures underlying mechanisms. It
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 597
is far better in general to specify the classes of response which are lost or
are preserved in response to each class of stimuli.
The Conditions for Confabulatory Response
I have advanced a theory of classical "agnosic" disturbances which
views them as naming disturbances which are frequently overlaid with
confabulatory responses. Why do these confabulatory responses occur?
Weinstein, Cole, and Mitchell (1963) have stressed certain features of this
type of response. Let me paraphrase some of their major conclusions: (1)
Confabulation does not occur in the absence of a defect; it is, indeed,
clearly an attempt to fill in a gap of performance of some type. (2)
Confabulation is less marked in the presence of aphasia; this seems
reasonable since it clearly is a response of the speech areas. (3)
Confabulation is more likely in the presence of some over-all impairment of
awareness, such as generalized dementia or clouding of awareness.1
To
these three conditions I would add a fourth (Geschwind, 19636): Confabulation
is much more likely in the presence of disease of association cortex
or association fibres (either commisural or intrahemispheric) than it is in
cases of damage of the primary sensory pathways up to and including the
primary sensory cortex. We can combine this with the third condition
above to make the following statement: The more demented or clouded
the patient, the more likely he is to show confabulatory response in the
presence of any defect of whatever origin; the less dementia and clouding
are prominent, the more confabulation depends on lesions of association
cortex or fibres; the two mechanisms interact frequently. These, of course,
are not the only conditions; the influence of personality prior to illness may
also be highly important.
Many examples come to mind to support this contention. My personal
experience of several cases of denial of blindness illustrates this rule. The
first case was that of a man suffering from a septicaemia, markedly febrile,
drowsy and confused, who was blind as the result of infection of the
anterior chambers of both eyes. Another was a man with a large subfrontal
meningioma with advanced optic atrophy and a classical severe
frontal syndrome with dementia. In these cases denial of blindness was
related to marked general impairment of function of the nervous system.
I have seen a confused patient who gave bizarre confabulatory responses
when asked to name objects held in the hand. He could not, however,
select the objects correctly from a group afterwards. This therefore must
have represented the confabulatory response to an actual sensory loss.
The same patient gave confabulatory responses to visual field testing when
there was no stimulus in the field but gave correct answers to stimuli within
the bounds of a normal visual field.
1
Although I have leaned heavily on the work of Weinstein and his co-workers in
these preceding statements, I would by no means wish to suggest that Dr. Weinstein
would have used the phraseology I use here.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
598 NORMAN GESCHWIND
In contrast with these cases, patients with bilateral occipital infarcts
(which probably usually involve much visual association cortex) are likely
to manifest denial of blindness without marked dementia. Blindness from
peripheral lesions is far less likely to produce such confabulatory denial in
the absence of dementia or clouding. Confabulatory denial of a lost limb
is uncommon; although the patient frequently has a phantom limb, he
nearly always treats it as a phantom. But denial of hemiplegia is common
and probably more common in the presence of accompanying lesions of
association areas; even if it were insisted that all of these patients
show some dementia or alteration of awareness, it is obvious that it is far
less than the degree of dementia needed for the denial of a peripheral defect.
I have, of course, already cited many examples of confabulatory response
in the earlier sections of this part of the paper.
There is some suggestive but not definitive experimental evidence for
this hypothesis. Warrington (1962) carried out experiments on visual
completion in patients with hemianopias. She found that those patients
who were unaware of a hemianopia were likely to "complete" forms across
a field defect while those who were aware of their hemianopias did not.
Thus, the first group would report a segment of a circle in the good field as
"a circle" while the second group would report it as an incomplete circle.
She found that this tendency to complete forms across field defects was
strongly associated with the presence of a parietal lobe lesion, i.e. in our
sense a lesion of association cortex. The anatomical and other data are
however not fully convincing. Thus, of the 11 patients who showed completion,
6 had tumours and 3 of the remaining 5 showed mental deterioration.
Thus, 9 of 11 had evidence either of widespread deterioration or of a
lesion likely to cause more than focal disturbance. (It must be admitted,
however, that the six tumours were all parietal in location.) In the group
of 9 patients with little completion only 2 had tumours (neither of them
parietal) and only one other case of this group was said to have shown
deterioration. Thus, only 3 of 9 patients in this group had evidence of
dementia or of a lesion with possible widespread effects. It is therefore
possible to use Warrington's data in support of a different thesis from the
one I have advanced—that patients with dementia or tumours are more
likely to develop unawareness of a hemianopia and to complete forms
across a field defect. Further studies of this type with greater description
of the degree of dementia should help to resolve the issue in favour of one
or the other interpretation.
Another paper by Kinsbourne and Warrington (1962) is also suggestive
evidence for the thesis I have presented. They found patients with right
hemisphere lesions who tended to misread the left halves but not the right
halves of words ("novel" for "level," "cucumber" for "encumbei"); 4 of
these cases showed complete left hemianopias but 2 did not. The authors
thought that all of these patients had right parietal disease. It would be
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 599
most interesting if it could be shown that this type of confabulatory reading
error resulted from right parietal association area lesions, but this would
require a control group of patients matched for dementia and with proven
nonparietal right-sided lesions who would not show the defect.
Why should such confabulatory responses be less common in lesions of
primary sensory cortex and more common in lesions of association cortex?
I cannot answer with assurance, but I can propose a highly hypothetical
mechanism which is at least suggestive of how such a situation could arise.
Since the speech area has in the normal no direct contact with the visual
cortex proper, destruction of the calcarine cortex leaves the speech area
still innervated by visual association areas. The association areas in this
case receive no stimulus from the visual cortex and send the "message" to
the speech area that there is no visual "message," i.e. that all is black. This
presupposes that the association areas never fail to send a "message" to
the speech area and that they always send positive messages of light or
darkness. The destruction of association cortex or fibres corresponds to a
totally unphysiological state—one in which no message is received by the
speech area. Since this is not interpreted (not being one of the normal
states) the speech area begins to react to random messages received from
subcortical pathways or indeed responds to its own spontaneous firing. It
may respond to incomplete information coming over subcortical pathways
which are inadequately extensive to carry all the information about the
seen object; this may lead to errors which are less bizarre than those in
which no information at all gets to the speech cortex. As I have stressed
this is a highly speculative mechanism, but certainly not an impossible one.
The Problem of Right Parietal Dominance
It may be useful to speculate briefly on another problem of major
interest in the current neurology of the higher functions, the question of
the dominance of the right parietal lobe in spatial functions. This problem
has been a perplexing one, perhaps because there is said to be a range of
disturbances from right parietal lesions unlike those seen with left-sided
lesions. On the one hand there is the fairly clear-cut syndrome of inattention
to the left side of space, manifested by a tendency to bisect lines to the
right, a tendency in copying to omit features in the left half of the picture,
neglect of the left side of the body as shown by failure to dress it or even
denial of its existence. The corresponding syndrome of neglect of the right
half of space seems to be less common. At the other extreme one sees
patients with right parietal lesions with the grossest impairments of
function apparently notconfined tothe left half of space. Thesepatientsmay
show gross disturbances in drawing figures or copying involving much
more than neglect of the left half of the object and difficulties in dressing
not confined to the left half of the body. A common argument has been
that since such gross impairments are more common with right than with
left-sided lesions, the right parietal region is the dominant one for spatial
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
600 NORMAN GESCHWIND
functions. However, there is a difficulty with this argument. In the first
place the sheer loss of the left half of space in some right parietal lesions
suggests that a dominant left hemisphere is being deprived of information
from the right hemisphere. By contrast, if the right side were usually
dominant then one should frequently see left parietal lesions with neglect
of the right half of space; this picture is, in fact, less common than cases
with neglect of the left half of space. Thus, in particular, drawings with the
right side omitted are less common than those with the left half omitted.
I would like to offer a tentative resolution of this paradox. In the first
place, let us consider the possibility that the main effect of a right posterior
parietal lesion is to disconnect the right visual and somesthetic cortex from
the dominant left hemisphere. (Whether the effects are due entirely to
disconnexion from the speech area, I would hesitate to say.) The left side
of the body and of space is then "lost." The patient will then respond in
many instances by using the technique of confabulatory completion that I
have discussed above and for which I have already cited experimental
evidence.
The confabulatory response may perhaps be responsible for some of the
more severe syndromes with right parietal lesions. If the confabulatory
response is very bizarre, then the patient must make a major readjustment
of all of space in order to deal with his "experiences." For example, an
occasional patient will express the feeling that he has two sets of left arms,
one which he sees and another which he feels. If he tries to order this
bizarre sensation according to his previous experience he must grossly
distort his entire body image. It will be no surprise if he now has trouble
dressing even the right half of his body. Similarly, if in copying a picture
he has a bizarre confabulatory structure in the left half of the field, he may
find it necessary to readjust the right half—albeit unsuccessfully, in order
to make his copy satisfactory. As I have already noted, this is a very
tentative suggestion which probably needs recasting in a clearer form but I
believe it may have the kernel of a new approach to the problems of right
parietal syndromes.1
1
It should be noted here that Denny-Brown (1962) also believes the left hemisphere
to be generally dominant and rejects right parietal dominance but on quite a different
basis from that presented here. Since the initial writing of this paper my attention
has been called to the results of Kimura (cited by Shankweiler, 1964 and confirmed
by him) who found that in normals melodies were better recognized when presented
to the left ear (in contrast to the greater efficiency of the handling of digits by the
right ear). These differences in melody recognition, although significant, were small.
This suggests that although the right hemisphere may be more efficient in certain
perceptual tasks, it is not dominant in the sense that the left hemisphere is dominant
for speech. This view is further supported by Shankweiler's observations that even
the presence of a right epileptogenic lesion shifts the balance in favor of the right
ear in melody recognition. In any case even the confirmation of some greater
efficiency of the right hemisphere in certain tasks would not invalidate the mechanisms
I have suggested for some of the dramatic right-hemisphere syndromes.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 601
Let us consider briefly the vexing problem of why a left parietal lesion
less often produces neglect of half of space than does a right parietal lesion.
In the first place the right hemisphere is probably always at a disadvantage
in relation to the left since the pathway from any part of it to the speech
area is probably at least one neuron longer than the pathway from the
corresponding part of the left hemisphere (there must be at least one extra
neuron to take the impulse across the corpus callosum). One extra neuron
may be of little import when added to a short path but probably the extra
degradation of the stimulus becomes more important the longer the route.
Any "detour" around a lesion used by the right hemisphere always suffers
thisextradisadvantageandtherefore makes it likelythattheright hemisphere
will be more completely disconnected. If in addition the right hemisphere
responds less well than the left to stimuli (see the earlier discussion on
dominance), it is subject to a further disadvantage. Thus disease may
simply aggravate the normal disadvantage of the right hemisphere in being
further away and responding less well to stimuli.
There is another mechanism which may contribute to the special disadvantage
of the right hemisphere. The speech area and other dominant
areas of the left hemisphere conceivably might receive some sensory information
directly from the left thalamus in the situation in which the parietal
association cortex was destroyed. This short cut is however not available
on the right; the thalami have no significant commissural connexions (the
massa intermedia is of minor significance and in any case is very frequently
absent) and therefore impulses from the right thalamus must still reach the
cortex of the right side before crossing to the other side.1
According to this view of right-sided association area lesions, they are
more likely to lead to disability and to confabulatory response than those
of the left side because the normally poorer linkage of the right side to the
speech area, and possibly to other "dominant" areas of the left hemisphere,
is further weakened by the lesions. This approach may contribute to
further understanding of some other phenomena which occur more
frequently in right hemisphere lesions. Thus, Hughlings Jackson (1880)
thought that dija vu was more likely to result from right than from left
cerebral lesions. Bingley (1958) confirmed this result. Mullan and Penfield
(1959) found that visual illusions and dija vu phenomena arose predominantly
from the hemisphere nondominant for speech. Cole and Zangwill
(1963) also found a predominance of dija vu from the minor hemisphere
although the predominance was not quite as great as in Mullan and
Penfield's series. Teuber, Battersby and Bender (1960) found seizures
1
Weinstein, Cole and Mitchell (1963) argue that denial is less common in left-sided
lesions because speech is less available; while this view, in my opinion, explains many
of the phenomena, it does not deal fully with nonverbal aspects. To round out the
theory it must be assumed, as I have already commented, that left dominance is for
more than speech.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
602 NORMAN GESCHWIND
involving visual experiences in 15 patients in all; there was a definite
preponderance of right hemisphere lesions. These authors cite the study of
Hecaen and Badaraco who found that 14 of 16 cases with visual experiences
had right hemisphere lesions.
I would suggest that perhaps this tendency of the minor hemisphere to
give rise to such illusions is not the result of localization of emotion or
memory in the minor hemisphere but another example of the increased
tendency to confabulatory response in cases of lesions of the association
cortex on the right side. The reasons for the right-sided predominance
would be the same as those given above.
It is interesting to note what Hughlings Jackson (1880) said about this
predominance of lesions in the right hemisphere in producing what he
called intellectual aurae: "These 'dreamy states'.. . cannot be owing to an
epileptic discharge. It would be a remarkably well-directed and distributed
epileptic discharge which would give rise to the exceedingly compound
mental state of being somewhere else . . . I submit that (such an elaborate
physical state) is owing to but slightly raised activity of healthy nervous
arrangements consequent on 'loss of control' . . . possibly of some in the
cerebral hemisphere opposite the one,1
which I believe to be nearly always
the right, in which the discharge begins. . . ."
I would like to cite one further recent piece of work to illustrate that it
may be possible to reinterpret certain studies which apparently support the
idea of right hemisphere dominance for certain functions. I have chosen
only one result in one of the many important papers principally by Milner
and Kimura published from the psychological laboratory in Montreal.
Kimura (1963) used a test in which a subject was shown a preliminary
series of picture cards. After this he was presented with a test series of 140
cards; he was to reply, "Yes" if he thought he had seen the picture in the
preliminary series, "No" if he thought that he had not. The total score was
the number of correct recognitions minus the number of false positive
responses. The group with right temporal lobe lesions scored significantly
worse on this task than the group with left temporal lesions. Further
analysis of these data, however, showed the interesting fact that the number
of correct "yes" responses was almost the same for both groups but that
nearly all the difference in scores was due to the large number of false "yes"
responses by the right temporal group.
Could this not be interpreted as another example of confabulatory
response resulting from a lesion of right hemisphere association cortex
rather than as a special deficit of the right hemisphere in the retention of
unfamiliar material ? Kimura gives two reasons why the patients with right
hemisphere lesions cannot be interpreted as having "a general tendency to
make false recognitions." The first argument is that in an auditory
memory task the right-temporal subjects showed no more tendency than
1
The italics are mine (N. G.).
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 603
the left temporal to give false positives. But since I have already given
several examples of confabulatory response confined to a single modality,
this argument would not rule out that the major effect of the lesions was to
produce confabulatory response only to unrecognized visual stimuli. The
other argument is a somewhat stronger one. Kimura points out that in
another task, the overlapping Nonsense Figures test (where immediately
after being shown a nonsense figure tachistoscopically the subject is
required to select it from a group of figures on a card) the right temporal
group performed more poorly than the left, despite the lack of any penalty
for false positives. Inspection of the data shows, however, that the significance
level on this test for differences between the post-operative right and
left temporal groups is between 0-01 and 002 while on the recurring figures
test the significance level for the difference between the groups was less
than -001. Furthermore, the difference between the post-operative groups
on the nonsense figures test was less than 2 items. By any standard this
difference appears far less impressive than the tendency for the right
temporal group to produce false positive responses on the recurring figures
test. Even if one were to accept the result on the nonsense figures test it
would argue for only a slight, if significant, superiority of the right as
against the left hemisphere on this task; it would not support the idea that
the right dominance for this activity was anywhere near as marked as left
dominance for speech. In any case it is obvious that many further careful
studies of the type carried on by the Montreal workers will be needed to
settle this issue as to whether the right hemisphere manifests dominance in
any sphere.
Visual Imagery
Related to the problem of the agnosias is the problem of loss of visual
imagery. Wilbrand's (1887) monograph on psychic blindness actually
concerned itself primarily with patients in whom visual imagery was lost.
I have already commented briefly in discussing Macrae and Trolle's interesting
case on the patient who could not "visualize" a route but could
travel it correctly. The word "visualize" has two connotations. We may,
as it were, see a picture in "our mind's eye," and can then say that we are
visualizing. At a simpler behavioural level we may describe an absent
scene. A classical notion was that visualization was an activity of the visual
association areas. The evidence for this is suggestive but by no means
complete. The peripherally blind can indeed describe absent scenes and
say that they have internal visual imagery. Would a patient with a pure
destruction of the visual cortex proper bilaterally still retain imagery? The
answer to this question is not known.
Loss of the ability to describe an absent scene might reasonably result
from a disconnexion of the visual regions from the speech area, as probably
was the case in MacRae and Trolle's patient. That this failure to give a
40 BRAIN—VOL. LXXXVUI
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
604 NORMAN GESCHWTND
verbal account need not be associated with a loss of visual memories is also
illustrated by the same patient who could in fact correctly drive to work
although he could not verbally describe the route. Was this patient capable
of developing inner pictures? There would be no way of finding this out by
interrogation since if the region in which such pictures were taking place
were disconnected from the speech area, we would be unable to get a
verbal account from the patient. The question is not a purely academic
one since the eventual development of more advanced physiological techniques
may settle some of these questions without recourse to a verbal
account from the patient. This would be analogous to recent work on
dreams for which objective criteria are now being developed.
The question of dreams and their locus is related to that of waking visual
imagery. Are dreams functions of the visual cortex or of the visual association
areas ? These problems of imagery and dreams deserve further
study using both standard clinical testing procedures as well as some of the
new techniques employed so effectively in the study of dreaming.
V. THE MECHANISMS OF THE APRAXIAS
Disconnexion from the Speech Area
The term "apraxia" in its modern usage goes back to Liepmann's (1900,
1906) employment of the term in describing the classic case of the
Regierungsrat. Despite the great complexity of the case, Liepmann devised
such remarkably appropriate methods of testingand analysed hisfindingsso
adroitly that he was able to predict the patient's lesions. At the patient's
death two years after the publication of the paper containing Liepmann's
clinical description, the post-mortem confirmed his predictions; subsequent
study of the material in whole-brain sections established in detail the
correctness of his ideas. The case has often been criticized as having too
many lesions to be useful. It is difficult to understand this point of view
when it is appreciated that the prediction of the loci of ah* the lesions was
made in advance despite the fact that no similar patient had previously
been described. These papers therefore differ importantly from those in
which the correlation of the clinical picture with the lesions is only made
post mortem.
In a group of papers published in the first decade of this century,
Liepmann established the clinical pictures of apraxic disturbances from
lesions in the supramarginal gyrus region and from lesions of the corpus
callosum. Since that time Liepmann's conclusions have been amply confirmed.
Even such workers as von Monakow who purported to be highly
critical of Liepmann actually confirmed his results. It is of interest that
Kurt Goldstein (1908) was one of the first after Liepmann to publish a case
of callosal disconnexion with motor disturbances.
Liepmann's analysis of the Regierungsrat was based on a disconnexion
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 605
approach. Unfortunately the very fact which makes Liepmann's accomplishment
so remarkable, namely, that he was able clinically to separate out
very precisely the patient's many disturbances into a small group of
functional impairments, each correlated to a particular lesion, makes the
case a very difficult one from the heuristic view. I will, therefore, use
instead as a case illustrative of the mechanisms of apraxia the patient
reported by Edith Kaplan and myself (Geschwind and Kaplan, 1962), the
mechanism of whose disturbance is much simpler. The reader is again
referred to the original paper for detailed discussion of the findings.
Our investigation of this case followed on Mrs. Kaplan's discovery that
the patient could write correctly with the right hand (despite the presence
of a severe grasp reflex and mild pyramidal disturbance in this hand) but
not with the left hand which showed no elementary motor disturbance. In
addition to this disturbance of writing with the left hand, we found that the
patient consistently carried out verbal commands with the right hand but
frequently failed to carry them out with the left hand. In classical terms
the patient showed an "apraxia of the minor hand." By further testing we
were able to go beyond this bare statement and found it possible to
delineate the mechanisms of these disturbances.
The disturbance of writing with the left hand could be shown to be
aphasic. Thus, when the patient wrote to dictation with the left hand he
did sometimes produce an illegible scrawl. On many other attempts,
however, he would produce perfectly written but incorrect words (e.g.
"run" for "go," "yonti" for "yesterday"); furthermore he misspelled
words when typing with his left index finger and could not correct the
errors which he was, in fact, able to recognize. By contrast he could copy
correctly with the left hand, but he could not with this hand turn print into
script although he could copy both print and script as such. This series
of studies excluded any elementary "motor" disturbance as being at the
root of the left-sided agraphia.
He often failed, as we have noted, in carrying out verbal commands with,
the left hand. He could, however, copy the movement if it was made
before him by the examiner. In addition, he would, given an object,
handle it correctly with the left hand although he had failed to show how
the object would be used in response to a verbal command alone. Thus,
"Show me how to use a hammer" resulted in a failure of movement or an
incorrect movement; he could, however, imitate the examiner's movements
or could use the actual hammer correctly. With his right hand he showed
none of these difficulties.
We interpreted these disturbances as the effects of disconnecting the
right motor cortex from the speech area. The post-mortem showed, as I
have already noted earlier, an extensive infarction of the corpus callosum
which must be regarded as the cause of these symptoms. The aphasic
writing with the left hand and the failure to carry out verbal commands
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
606 NORMAN GESCHWIND
with the left hand could quite easily be explained as a result of disconnexion
of the right motor cortex from the speech area. The ability to copy writing
correctly but slavishly with the left hand; the ability to imitate seen movements
with the left hand; the capacity to handle objects correctly with the
left hand, all illustrate that the right hemisphere could perform correctly
when no information was required from the opposite side. By contrast the
inability to "copy" print into script is an expected one since such a transformation
implies the participation of the speech areas. Certain movements
in response to verbal command which involved both the right and
left sides simultaneously were preserved as were facial movements to
•command. I will deal subsequently with the analysis of these two special
classes of movements.
The case has several features worthy of comment. It illustrates that the
designation "apraxic" is an inadequate one unless the stimulus conditions
are specified. The left hand in this patient was apraxic to verbal command
but not on imitation or object handling; the right hand failed to perform
correctly when a response was demanded from this hand to somesthetic
stimuli applied to the left hand. Rather than use the term "apraxia" it is
therefore preferable to specify the stimulus-response combinations which
fail.
It is not possible to assert that the failures of this patient resulted from
a general conceptual disturbance or a failure of abstract attitude. The
patient could not, indeed, pretend to perform an act with the left hand, but
he could pretend with the right hand! It would seem highly unlikely that a
disturbance of conceptualization or of the abstract attitude could be confined
to one-half of the body. "Dementia" as an explanation runs up
against the same difficulties. Hysteria or malingering cannot, of course, be
seriously considered at all.
The case also points to the probable incorrectness of the classical distinction
between aphasic and apraxic agraphia. Many writers have assumed
that agraphia confined to the left hand must be "apraxic." According to
this argument an apraxic agraphia is the result of failure to form the
individual letters properly and should result in a pattern of meaningless
loops and scrawls. Further support for the distinction between aphasic
and apraxic agraphia came from the argument that aphasia could not be
confined to one limb. Our case, however, showed both "apraxic" and
•"aphasic" types of writing defect in the left hand. In the early stages his
"writing was in fact usually but not invariably a totally illegible scrawl. But,
particularly after some improvement had occurred, he produced perfectly
-written words which were incorrect, e.g. "yonti" for "yesterday"; his
failures at typing were, of course, also misspellings. It would not be possible
to term these other than aphasic errors. It seems unnecessary to
.assume the presence of two separate writing disturbances in the patient.
Both types of error (i.e. incorrectly formed letters and aphasic words) can
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 607
easily be seen as resulting from a disconnexion between the speech area and
the motor cortex.
I have also made no attempt to deal with this disability in terms of the
classical grouping of apraxias into limb-kinetic, ideomotor, and ideational
which Liepmann (1905a) advanced. It would lead us too far astray to
discuss this classification in any detail in this paper. It is my opinion that
this classification has not been useful and that it has in fact frequently
proved very confusing. It is interesting that the classification plays almost
no role in Liepmann's earlier classic discussions of the Regierungsrat and
even in the later writings is clearly secondary to the intimate discussion of
mechanisms. This formal listing of types of apraxia probably had in
practice the unfortunate effect of overshadowing Liepmann's much more
important detailed analyses of the mechanisms underlying disturbances in
motor performances.
The occurrence of such unilateral agraphias is also of considerable
interest when the hand involved by the agraphia has been the one generally
used by the patient for writing. Thus, Nielsen (1946) described the
remarkable case of a man who had been taught to write with his right hand
but used his left hand for all other skilled movements. He suffered a
transient left hemiplegia; this was accompanied by loss of speech, strongly
suggesting that the right hemisphere was dominant for language. Some
months after recovery from this episode he developed a paralysis of the
right leg and lost the ability to write with his right hand. He could, however,
write with his left hand although he had not used the left hand
previously for this activity. In interpreting this case one must assume that
although the patient had always written with the right hand, the left
hemisphere had, in fact, always been passively under the control of the
speech area in the right hemisphere.
This case is also useful as a comment on an old idea, that cerebral speech
dominance is much weaker in illiterates and that the act of writing, being
unilaterally performed, influences the opposite hemisphere (see Critchley,
1962, for a brief review of the history of this idea). Cases like the one just
cited, however, cast serious doubt on this possibility since it is clear that
Nielsen's patient had not succeeded in establishing speech in the left
hemisphere. The case also makes somewhat unlikely the old suggestion
that training in ambidexterity makes a severe aphasia less probable. This,
of course, is not in conflict with the assertion that those who are naturally
ambidextrous have milder aphasias (Zangwill, 1960).
Extension of the Theory of the Apraxias
Inthepreceding section I have developed the theory of the simplest typeof
apraxia, that due to disconnexion of motor regions from the speecharea and
from other sensory inputs. The interpretation of the case of Geschwind
and Kaplan is clear: callosal disconnexion in man prevents the right
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
608 NORMAN GESCHWIND
hemisphere from carrying out language tasks and also from performing
those tasks for which the stimulus comes from non-language sensory areas
of the left hemisphere. This theory depends on the assumption that the
right hemisphere will deal with non-language tasks correctly as long as
appreciation of the stimulus and production of the response both depend
on the right hemisphere itself. The patient of Geschwind and Kaplan
exhibited independence of the right hemisphere in non-verbal tasks; this
hemisphere functioned successfully in imitation of movements, object
handling and slavish copying.
A deeper search into the literature and further clinical observations soon
convinced me that the independence of the right hemisphere in nonlanguage
functions which was manifested by our patient is not universal
and may indeed be the exception. It was the work of Liepmann which
again contains the earliest and probably the best evidence for this view.
The first description of the clinical picture of extensive disconnexion of
the corpus callosum is contained in the paper of Liepmann and Maas
(1907). Their patient (Ochs) had a right hemiplegia which at post-mortem
was found to be due to a lesion of the pyramidal tract in the left pons. The
patient could carry out many actions very well with his left hand, e.g.
buttoning his coat if the hand was placed on the first button, or carrying a
glass of water (if it was placed in his left hand), to his lips. He failed to
write with the left hand; in fact, he could not even copy. When given
anagram letters he could not form his name with his left hand; instead he
collected all the m's (yet when asked, he could correctly say the letters of
his name). He made many errors in carrying out verbal commands with
the left hand but selected the correct response from a series of movements
made by the examiner. He did not, however, improve on imitation. This
patient frequently mishandled objects placed in the left hand. At postmortem
there were found the already mentioned lesion in the pyramidal
tract in the pons as well as an infarct of the corpus callosum which spared
the splenium. Other patients with callosal lesions have shown a similar
disturbance to that shown by Ochs. Not only can they not carry out verbal
commands with the left hand, write with this hand, or form words with
anagram letters (in which respects they resemble the case of Geschwind and
Kaplan), they also show significant impairment in imitating seen movements
with the left hand and difficulties in object handling with this hand.
The case of Bonhoeffer (1914), for example, showed these deficits. Why
do the cases fall into two such groups? It is interesting that Maas (1907)
himself called attention to this problem. His explanation was that it
depended on the intensity of the callosal lesion, but I do not believe that
this is correct.
Because the first case of "motor" disturbances due to callosal disconnexion
seen by Liepmann and Maas showed the more extensive syndrome
of left-sided apraxia to verbal command, on imitation and in the handling
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 609
of objects, these authors regarded this clinical picture as the "normal"
syndrome of callosal disconnexion. Liepmann explained the data by
assuming that the left hemisphere is dominant not only for speech but also
for movement. This dominance was assumed to manifest itself trivially in
the usual greater dexterity of the right hand; a more important aspect of
this dominance, however, was that the left hemisphere contained the
"memories" of movements.
Evidence from other types of patients support Liepmann's contentions
that the left hemisphere is dominant for more than speech. Thus,
Liepmann (19056) described the syndrome of apraxia of the left hand in
patients with right hemiplegia and aphasia ("sympathetic" dyspraxia as it is
sometimes called). I have reinvestigated this problem (Geschwind, 1963a)
and been able to confirm Liepmann's observations in detail. This syndrome
is of special interest both because of its great frequency (it is, in fact,
much the most common cause of apraxic disturbances) and its theoretical
interest. Liepmann found this syndrome in 14 of a group of 18 patients
with right hemiplegia and severe aphasia. By contrast it occurred in only 6
out of 23 right hemiplegics without severe aphasia. It did not occur at all
in the right hands of a group of left hemiplegics nor in a group of senile
and demented patients. These controls effectively remove the possibility
that the syndrome is the non-specific result of brain damage or dementia.
They also show its close relationship to lesions producing aphasia. The
fact that it did not appear in all the hemiplegic aphasics and that it appears
occasionally without aphasia suggests that the relationship is one of
anatomical propinquity of lesions; the apraxia frequently accompanies the
aphasia but is independent of it.
The syndrome of sympathetic dyspraxia is essentially the same as that
described by Liepmann as occurring in the left hand in callosal lesions. The
patient carries out verbal commands either very poorly or not at all with
the left arm. Demonstration by the examiner helps in some instances but
in most cases little or not at all. Usually the actual handling of objects in
the left hand is significantly better than movements to command or in
imitation of the examiner, but major failures in the handling of objects do
occur in a significant minority of these patients.
Several interpretations of this syndrome can be excluded. I have already
noted the impossibility of attributing this picture nonspecifically to brain
damage or dementia. One obvious interpretation is that since these patients
are generally aphasic, their failures to respond to verbal command are the
result of comprehension deficit. Liepmann showed in several ways that this
interpretation is untenable. In my cases I have also been able to rule this
out. Thus, one of my patients snowed this syndrome in especially clear
form, failing very badly on verbal commands and imitation and also doing
poorly in object handling. He had a marked limitation of speech, being
able to produce only single words and never producing a sentence. He
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
610 NORMAN GESCHWIND
could, however, answer questions very well with single words. He failed to
carry out very simple commands such as to point to the floor but correctly
answered verbally much more complex questions as "What occupation
were you engaged in before you became ill?" and "Can you name some
tools used by carpenters?" Furthermore, the patient could answer specific
questions about the task. When asked, "Do you know how to use a
hammer?" he said, "Nails" but could not show how a hammer was used.
He could, however, indicate when the examiner had made the correct
movement although he could not produce it himself. In other cases a
strong clue that comprehension is intact is given by the fact that the patient
may respond with a movement that is a clear distortion of the correct
response. For example, a patient exhibiting this syndrome when asked to
salute may place his hand in the proper attitude but several inches in front
of his forehead. These facts all contribute to rule out incomprehension as
the cause of the failures. They also highlight a very important aspect of the
examination of aphasia. Many textbooks suggest examining for comprehension
by means of the ability to carry out verbal commands. It is
obvious from the results I have cited that while the ability to carry out
verbal commands indicates retained comprehension, its absence does not
exclude the preservation of excellent comprehension.
The fact that my patient could give good single word verbal responses
while failing to carry out verbal commands probably rules out another
possible interpretation. Some would argue that aphasia is not a disturbance
of speech but one of symbolization and that my patient was incapable
of symbolic manifestations in any form. The patient would under this
interpretation be regarded as having lost gesture as well as spoken
language. However, aside from the difficulty of regarding some of the
movements involved as symbolic, it would be necessary to assume in my
patient that non-verbal symbolization was worse than verbal symbolization!
Once one admits the separability of the disturbance in movement from the
disturbance in language, the concept of a general asymbolia loses its value.
Liepmann felt that this group of patients consisted predominantly of
cases with lesions near Broca's areas as evidenced by their aphasic
symptomatology (paucity of speech with good comprehension). He thought
that the lesions probably would be found in the white matter immediately
beneath Broca's area but above the internal capsule. A lesion in this
location would involve not only descending fibres but would also destroy
fibres going into the corpus callosum and thus to the opposite side. In
those brains which he had had the opportunity to observe at post-mortem
he did find such subcortical lesions. He did not make clear exactly from
what part of the cortex of the left frontal lobe the callosal fibres originated,
a point of importance to which I will return. It is clear that he thought the
lesion resulted in disconnexion of the right motor region from some portion
of the left frontal region; the lesion was regarded by Liepmann as cutting
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 611
off the right hemisphere from the memories of movements stored on the
left side.
The Apraxias of the Supramarginal Gyrus Region
Further details of the above mechanism can be still further elucidated by
considering another group of "apraxic" patients, those with lesions of the
left supramarginal gyrus region. Liepmann (1900,1906) first described this
syndrome in the famous case of the Regierungsrat. The general picture
which emerges from the literature is a consistent one. Unilateral lesions
penetrating deeply in this region may lead to an apraxia which is either
equal bilaterally or worse in the right hand. Except for its bilaterality, it is
similar in its pattern of disabilities to that usually seen in the two preceding
conditions (apraxia from a callosal lesion and "sympathetic dyspraxia").
The patient performs badly to verbal command, imitation of the examiner
is usually poor and object handling is involved, but less often. Liepmann
stressed that the responsible lesion was one which cut off white matter connexions
(lying deep to the supramarginal gyrus) between posterior parts of
the left hemisphere and motor regions. The retained connexions on the
right side would not have been able to substitute, in most cases,
since there had been no learning on the right side. The lesion in the supramarginal
gyrus would cut off connexions running from Wernicke's area to
motor regions via the arcuate (superior longitudinal) fasciculus and also
connexions from the visual region to the motor area.
There is a problem raised by this simple disconnexion hypothesis. This
lesion deep to the supramarginal gyrus should disconnect Wernicke's area
from motor regions and the visual region from motor regions; the failure
of the patient to carry out movements to verbal command or on imitation
is therefore understandable. Why should movements be impaired in the
actual handling of objects ? There are several possible explanations, but we
cannot be sure which, or indeed, whether any of them is the correct
one. Liepmann suggested the following mechanism. He asserted that
even the manual handling of objects was frequently learned "visually."
One can cite simple examples which are suggestive. Many acts, such as
combing the hair, may be routinely carried out under visual control. When
they are done without vision Liepmann's suggestion would be that they are
done with "visual mediation." In other words cross-modal transfer would
be involved in the learning of many tasks, and the performance under
tactual control would depend on transfer from visual learning.
Liepmann brought forward as evidence for this view the pattern of
retained responses to somesthetic stimulation. He pointed out both for the
Regierungsrat and Ochs that while they handled most objects poorly, they
performed well on somesthetic stimulation in those tasks which were
highly overlearned and which were usually done independently of vision.
Thus, the Regierungsrat and Ochs both failed in some very simple object
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
612 NORMAN GESCHWIND
manipulations, yet were able to button their clothing blindfolded so long as
the hand was first passively placed on the garment. Liepmann attributed
this to the ability of the isolated sensory and motor cortex to do this highly
overlearned task without the mediation of vision. Liepmann made use of
the preservation of such movements as important evidence. It enabled him
clinically to rule out any sort of elementary motor or sensory disturbance
since these would be incompatible with the deft buttoning of a coat while
blindfolded. This fact emphasizes again the point that "apraxia" is not a
unitary disturbance since under appropriate conditions these patients could
carry out complex motor tasks.
There is another possible explanation for the difficulty in object handling.
Liepmann had felt it necessary to account for the fact that the patients
could not carry out all tasks of object handling correctly even though the
primary sensory and motor cortex retained their direct connexions. We
know that there are connexions via U-fibres between the postcentral and
precentral gyri. However, are these connexions used for the learning of
complex tasks under somesthetic control ? It is not at all unreasonable to
assume that for the learning of such complex tasks the pathway from
primary sensory to primary motor cortex may be via association cortex,
as in the case of connexions of other modalities to the motor system.
The pathway from sensory to motor cortex would thus run from sensory
cortex via somesthetic association areas and then would travel forward to
the motor region. If this were the path then a lower parietal lesion might
well involve part of this somesthetic-motor pathway. Sparing of this
pathway in its entirety might leave object handling totally unaffected. If
this explanation were correct then one would have to assume that the
preservation of some movements depended on theirhaving been overlearned,
i.e. on their being so redundantly represented in the somesthetic association
cortex that partial lesions left them unimpaired. This problem of
the factors affecting movements made under somesthetic control is one
which obviously deserves further study.
Liepmann repeatedly insisted that the critical lesion in the region of the
supramarginal gyrus involved not the cortex but the white matter running
beneath the gyrus. An alternative explanation would be that the "memories
for movements" were in the supramarginal gyrus region, i.e. that the
cortex was involved. Against this and in favour of the notion of white
matter disconnexion is the fact that lesions anterior to the left Rolandic
fissure produce apraxia on the left side of the body despite an intact left
supramarginal gyrus. One would expect that, if the cortex of the left
supramarginal gyrus were the important structure and were not involved
by the lesion, the impulse could travel across the callosum to the right
supramarginal gyrus and thence forward to the right motor region. In this
instance no apraxia should be seen. The other possible pathway for control
of the left side of the body involves a path running beneath the left
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 613
supramarginal gyms to the left frontal lobe and from here to the right
hemisphere. In this case apraxia of the left side could occur with a
lesion anterior to the left supramarginal gyms.
The sinistral unilaterality of the pathway in most people for the carrying
out of acts by the motor cortex under sensory stimulation is emphasized by
the rarity of well-attested cases of left-sided apraxia occurring as the result
of an isolated right parietal lesion. Foix (1916) emphasized right parietal
lesions as a cause of left-sided apraxia, but his evidence is not convincing.
He even said that "Liepmann's principal case" had a right parietal lesion,
but since this statement applies neither to the brain of the Regiemngsrat
nor that of Ochs, we may dismiss this assertion as simply being incorrect.
Hecaen and Gimeno Alava (1960) have discussed this problem. They
found 16 cases, 3 of their own, the remainder from the literature, in which
patients with apraxia of the left side had clinical evidence of a right
hemispheric lesion. However of these, 11 almost certainly also had lesions
of the left hemisphere. Among the remaining 5,2 were known left-handers.
Of the remaining 3, 1 came to post-mortem and showed no involvement of
the left hemisphere. If one can accept these findings it would appear that a
right parietal lesion causes apraxia of the left hand only exceptionally
unless theie is an associated left hemispheric lesion. This conclusion is
further borne out by the study of Ajuriaguerra, Hecaen, and Angelergues
(1960). If we consider only those cases in their series which are relevant to
our discussion, we find that they had 58 cases of apraxia (divided by them
into 11 cases of ideational apraxia and 47 cases of ideomotor apraxia), of
which 48 occurred in their 206 cases of left-sided post-Rolandic lesions,
and 10 occurred in their 55 cases of bilateral lesions. There was not a
single case of apraxia in their 151 cases of right-sided post-Rolandic
lesions.1
I would in fact expect that there exist occasional cases in which a
right supramarginal gyms lesion might lead to an apraxia of the left arm
on imitation and object handling (but possibly not in response to verbal
command). This unusual situation would exist only in those patients in
whom the right hemisphere was relatively independent.
I have already noted that deep to the supramarginal gyms are fibres
from the visual association areas running into the frontal lobe and fibres
from the speech area (which is, of course, auditory association cortex) also
coursing to the frontal lobe. These are probably the pathways by which
motor responses are carried out in response to complex visual or auditory
stimulation. Where in the left frontal lobe do they terminate and by what
precise pathway does stimulation get across to the opposite motor region?
One's first inclination would be to assume that these pathways terminated
in area 4, the primary motor cortex. But again, the mle of Flechsig
x
Let me stress that I am confining the term apraxia to the sense in which Liepmann
used it. I am not including the cases separately tabulated by Ajuriaguerra et al. as
•"constructional apraxia" and "dressing apraxia."
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
614 NORMAN GESCHWIND
(1901) comes into play. The motor cortex, at least in the primate, receives
no direct fibres from the visual association cortex. Chusid, Sugar and French
(1948) produced a most instructive study of the frontal connexions of the
visual association cortex in the macaque which brought out clearly the fact
that many of these important connexions arise from the tissue buried in the
depths of the lunate sulcus of the occipital lobe. They found with strychnine
neuronography that the anterior bank of the lunate sulcus strongly fired
the cortex immediately anterior to itself. The only distant region where
they found strong firing was the cortex of both banks of the arcuate sulcus
of the frontal lobe and the cortex immediately anterior and posterior to this
sulcus. There was no firing in the motor cortex proper. The posterior bank
of the arcuate sulcus in turn fired strongly into areas lying immediately
posterior to itself including the homolateral area 4 and also strongly fired
the contralateral arcuate sulcus region. Bonin and Bailey (1947) report
that FA (area 4) received fibres from FB (lying above the superior half of
the arcuate sulcus, often called area 6) and from several areas in the parietal
lobe but not from any other part of the frontal lobe. Furthermore, FA had
no callosal connexions in the primate except for some restricted parts of
the trunk and lower face regions. Their results and those of Chusid, Sugar,
and French thus are complementary. Similarly, Bailey, Bonin, and
McCulloch (1950) state that in the chimpanzee the most numerous afferents
to FA come from FB and PC (the postcentral gyrus). The pattern of
callosal projection is the same as in the macaque.
Krieg (1954) studied the efferent connexions of the frontal lobe in
monkeys by the Marchi method. He found that each area 6 has strong
callosal connexions to the opposite area 6. In addition each portion of area
6 sends numerous fibres to the part of area 4 directly behind it. Krieg's
findings are therefore in keeping with the physiological findings of the
workers cited above. Krieg did find callosal fibres from area 4. He
comments, however, that the number and calibre of these fibres is very small
and his illustrations of the actual lesions and the consequent degeneration
support this assertion. Krieg notes that his anatomical findings are in
keeping with the poor callosal responses obtained from area 4 by Bailey,
Bonin and McCulloch (1950).
We might guess from these data that connexions between visual and
motor cortex run as follows: from the visual cortex proper to the visual
association cortex to area FB (roughly area 6) and thence to the motor
cortex. Area FB thus becomes the association cortex of the motor system.
Similarly, the callosal connexions of the motor system would have to be
via area FB.
The view that the premotor cortex acts as the association area of the
motor cortex receives support from the connexions to the motor region of
the auditory system. Thus, Sugar, French, and Chusid's (1948) study
showed that the important frontal projections from the supratemporal
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 615
plane are to the cortex lying within the depths of the arcuate sulcus and to
the cortex immediately posterior to this sulcus, in particular to the regions
behind the lower limb of the sulcus. In other terms the projections from
the supratemporal region are to areas 8, 6b (lower half of area 6), and 44;
there are no direct connexions from the supratemporal plane to area 4. The
earlier study of Ward, Peden, and Sugar (1946) which was the first to
discover afferent connexions to area 6 had also revealed the connexions
from the supratemporal plane to this part of the frontal lobe.
Thus, it would appear that the auditory system like the visual system
projects to this "motor association cortex" lying anterior to area 4, and
that this region in turn projects to area 4. It seems reasonable that this is
the pathway by which motor tasks are carried out to auditory stimulation.
These studies all support the theory that connexions from primary
sensory areas to the motor cortex involve a multisynaptic pathway which
travels by way of the association areas adjacent to the sensory areas in
question and the "motor association areas" lying anterior to the motor
cortex proper. The startling experiments of Welch and Stuteville (1958)
are readily explained by this theory and also constitute further evidence for
it. These authors placed small lesions in the depths of the posterior part of
the superior limb of the arcuate sulcus in monkeys. Confirming earlier
results by Kennard and her co-workers (summarized in Kennard, 1939)
they found that these animals disregarded visual stimuli on the side
opposite the lesion, and indeed even collided with objects on that side. The
animals did not respond to auditory stimuli coming from the side contralateral
to the lesion by turning towards the stimulus (as a normal monkey
would); instead they turned towards the side opposite the stimulus.
Finally, these animals exhibited little response to tactile or painful stimuli
on the side opposite the lesion. These animals showed a marked poverty of
movement on the involved side. Tactile and visual placing reactions were
lost on that side. Recovery from this dramatic syndrome began in five to
seven days and was complete in two weeks.
Welch and Stuteville state, "How vision, hearing, tactile sensibility,
movement and placing reactions are so seriously impaired by a lesion of the
cortex which does not directly interfere with any of the several receiving
areas or the motor area of the cortex is difficult to understand." The
results, however, become easy to understand when one considers that the
area involved by the lesion is in fact part of the pathway from all sensory
modalities in one hemisphere to the motor cortex of that hemisphere. The
lesion therefore effectively disconnects the motor cortex of that side from
all sensory stimuli. Why does the animal not respond with the hand
opposite the normal hemisphere to stimulation on the side opposite the
damaged hemisphere? I would suspect that it is probably because the
lesion has also cut off the origin of callosal fibres between the two areas 6.
This explanation of the Welch and Stuteville data may well apply to other
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
616 NORMAN GESCHWIND
studies in which lateral frontal lesions have been said to lead to difficulties
in sensory discrimination (e.g. Weiskrantz and Mishkin, 1958)—the real
deficit may be in motor response.
Some aspects of the Welch and Stuteville experiments deserve brief comment.
Ettlinger and Kalsbeck (1962) found a loss of tactile placing responses on the side
contralateral to either a lesion of the primary somesthetic cortex or of the parietal
association areas. Welch and Stuteville found that these responses disappeared in
their experiments. The normal pathway for a tactile placing reaction would
appear to travel not via the U-fibres connecting the primary somesthetic and
primary motor cortex but rather by way of the association areas. A striking
finding of the Welch and Stuteville experiments was the transience of this dramatic
disturbance. This is another example of the tendency for small association area
lesions to be compensated. Was the compensation by way of normal cortex
adjacent to the damaged region, or were the animals using totally new pathways ?
This remains to be studied. There was an even more startling result in the Welch
and Stuteville study. An animal showing this syndrome behaves as if he is blind
in thefieldopposite the lesion. Following a removal of the occipital lobe on the
side opposite the lesion there is a return of responsiveness to stimuli in the remaining
half of the visual field although the animal continues to neglect tactile and
auditory stimuli on that side. It would appear that some alternative pathways,
for motor response to visual stimulation of the hemisphere on the side of the
frontal lesion are kept inhibited as long as the opposite visual region is operative.
It would be most interesting to trace the anatomical substrate of this inhibition. If
it is mediated via the corpus callosum it would appear likely that callosal section
might cause the entire syndrome to vanish!
The first case of Hartmann (1907) might appear to be a human example of this
syndrome of marked inattention to all modalities of stimulation coming from one
side as the result of a frontal lesion. There are, however, certain difficulties in the
clinical picture and pathology which prevent the full use of this case. Hartmann,
however, must be credited with considerable prescience for he argues that since
the right arm of his patient showed no movements in response to sensory stimuli,
the frontal lobe (i.e. the region anterior to the motor cortex on the left side) was
necessary for the conduction of sensory stimuli to the central gyri. This would
agree of course with the thesis presented here that the premotor region is a waystation
on the route from sensory cortex to primary motor cortex.
I have postulated a mechanism for the response of the motor to cortex
sensory stimulation. If this thesis is correct we must assume that the
apraxia of the left limbs in the patient with a right hemiplegia is not the
result of the part of the lesion causing the hemiplegia. Rather it is the
result of coincidental damage either to area 6 or to the callosal fibres to
which it gives rise. Alternatively it is the result of damage to the afferents
passing to area 6 from the visual, auditory, and somesthetic regions and
travelling perforce beneath the motor cortex.
There are certain implications of this argument which are worth considering
for the moment in more detail. The argument implies that a
lesion in man of the regions anterior to the motor cortex on the left, but
sparing area 4, should produce bilateral apraxia in our sense, i.e. failure to
carry out movements to command and on imitation of the examiner but
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 617
with variable effects on actual object handling. As many readers will know
it has in fact been asserted for a long time that lesions anterior to the motor
cortex lead to apraxic disturbances {see, for example, Nielsen, 1946; Aring,
1944) and it has often been argued that these are "limb-kinetic" apraxias.
Perhaps the best known of these disturbances from lesions anterior to the
motor cortex are the facial apraxias (whose discussion I leave to the next
section) and the pure agraphias which are said to arise from the posterior
end of the second frontal convolution. I will not discuss the problem of
"frontal apraxia" in detail since my own acquaintance with this large
literature has not been intensive. My feeling has been that "limb-kinetic"
apraxia has not been defined clearly enough to separate it from mild
pyramidal disturbance. Liepmann (1905a) himself must have been somewhat
unsure of limb-kinetic apraxia since he used as an example of this
type of patient no case of his own but rather one published many years
earlier by Westphal. The apraxias of the left side accompanying Broca's
aphasias, which I have already mentioned, are due either to lesions of area
6 itself or of the callosal fibres arising from it. There is no clear-cut
evidence to show that a lesion of the left-sided area 6 itself would produce a
bilateral apraxia in virtue of its cutting off connexions both to the homolateral
area 4 and to the contralateral motor region. The reason that such
cases are lacking is probably the result of the fact that lesions of the left
area 6 are very likely to encroach on the left area 4 and therefore to produce
so much weakness in the right hand as to make assessment of apraxia
difficult.
I believe, however, that this is an area where animal experimentation
may be of greater use than it has been. Earlier experiments on ablation of
area 6 have generally stressed "motor" functions {see, for example, the
discussions in Bucy, 1944), rather than learning activities. The work of
such investigators as Kennard (1939) Welch and Stuteville (1958),
Weiskrantz and Mishkin (1958) should be extended. Of parallel interest
would be further knowledge of way stations to the hippocampal region
from area 6.
Facial Apraxia
This term is used as a shorthand for "apraxia of the cranial musculature,"
i.e. inability to carry out movements of this musculature either to
command or on visual or tactile stimulation. It is the first form of apraxia
to have been recognized (although not under that name) by Hughlings
Jackson (1878) who commented on the inability of certain aphasics to
protrude the tongue. It has continued to arouse interest sporadically.
Many authors would, in fact, like to include Broca's aphasia as a part of
such an apraxia.
Certainly this disturbance is the most common apraxic disturbance in
aphasics. It has probably been recognized much more often than the more
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
618 NORMAN GESCHWIND
extensive apraxia of the left side which I have mentioned earlier. I suspect
that one reason for this is that when the patient fails to carry out
facial commands but does perform limb commands, it is clear that
comprehension is intact. When all commands on the intact left side are
not performed, it is all too easy to attribute the failure to incomprehension.
The clinical picture is essentially similar to that of other apraxic disturbances
but has some distinctive and intriguing characteristics. The
patient usually does most poorly in carrying out facial movements to verbal
command. He may simply fail to perform at all or may make an incorrect
movement, e.g. he mayopen the mouth when asked to protrude the tongue,
or blow instead of suck. He may make movements with one of his limbs to
carry out the demanded task; thus, he may pretend to stub a match in an
ash-tray or stamp on it with his feet when asked how he would blow out a
match. He may remove imaginary crumbs from his lips with his fingers and
not with his tongue even when asked repeatedly to use his tongue; he may
even insist that he has always performed this action with his hand. Most
interesting of all, the patient may echo the command or produce onomatopoietic
responses. Thus, when asked to cough, he may say, "Cough";
when asked to blow out a match he may say, "Blow" or even, "Blow out a
match," or "Puff." This type of response is all the more striking in that
these verbalizations or vocalizations do not appear when the patient
carries out the command correctly; similar disturbances are much rarer
when patients fail to carry out limb movements to command.
The special peculiarity of facial movements is, of course, that they are
generally bilateral. We must therefore ask how this bilaterality is achieved.
A common view is that facial movements are integrated via the corpus
callosum. This seems somewhat reasonable in view of the assertion that by
strychnine neuronography (McCulloch and Garol, 1941) there can be
demonstrated callosal connexions only between very limited parts of area
4, comprising the trunk and neck divisions and part of the face division.
There is some reason, however, to question whether in man even these
callosal connexions are of importance. Thus, the patient of Geschwind and
Kaplan (1962) who performed poorly in carrying out commands with the
left hand or left leg moved his face bilaterally in carrying out facial commands.
This suggests that the pathway by which facial commands are
carried out descends unilaterally to some brain-stem level where bilateral
facial movement is integrated. It may be argued that a unilateral lesion of
the face area of the right cortex produces weakness of the left lower face.
This weakness, however, may be the effect of the withdrawal of tonic
influence resulting from destruction of the right cortical face area and is
probably not good evidence that movement of the left face to verbal command
is mediated via impulses passing over the callosum. The right
cortical face area would thus be regarded as providing a constant tonic
influence but as having a phasic effect only under special conditions.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 619
Alternatively one might speculate that normally these bilateral facial
movements are integrated via the callosum but that when the callosum is
gone there is ready replacement by means of directly descending pathways
providing bilateral innervation. I think that this explanation is less likely
than my earner one, i.e. that each cortical face area separately is normally
capable of causing bilateral activity at a brain-stem level.
We can now consider the pathway for facial movements to command or
on visual stimulation. The pathway for facial movements to verbal command
probably goes from the posterior speech area via the arcuate
(superior longitudinal) fasciculus to the association cortex lying anterior
to the face area. Similarly, the pathway to the face area from the visual
region probably also passes in the inferior parietal region to some area
anterior to the Rolandic face area.
The first implication of this is that lesions of supramarginal gyrus region
should lead to facial apraxia. It has been our experience that they
commonly do so. In particular facial apraxia to command and visual
stimulation is likely, in our experience, to accompany the clinical syndrome
of conduction aphasia, whose importance and whose anatomical basis have
been emphasized lately by Konorski, Kozniewska and Stepien (1961) and
to which I will return in a later section. These patients show no
hemiplegia in the great majority of cases. Their speech may be fluent
with many phonemic paraphasias. Even when speech is limited there are
runs of fluency. While mild dysarthria is occasionally seen, the great
effort and marked dysarthria of the Broca's aphasic are absent. Despite
good to perfect comprehension there is marked difficulty in repetition.
Their lesions probably lie low in the arcuate fasciculus (Konorski et al.,
1961). It is my opinion that the accompanying facial apraxia results from
the same lesion. The fact that limb apraxia may be absent suggests that
the fibres intended for facial "association cortex" run lower down than
those going to the motor association cortex for the limbs.
The common occurrence of facial apraxia with lesions near Broca's area
is readily understood. The lesion may destroy the association cortex
anterior to the face area. Facial movements cannot be carried out to
command or on visual stimulation because the lesion has cut off connexions
to the left face area and cut off the origin of callosal fibres to the right face
area. Alternatively a lesion in this neighbourhood may destroy left
cortical face area or its projection fibres. In this case facial movements to
command or visual stimulation must be carried out via the right face area
receiving stimulation over the corpus callosum. If the lesion extends deep
into the white matter it can involve the callosal fibres and facial apraxia
will result. When a tactile stimulus, e.g. a drinking straw in the mouth, is
used, the act, if overlearned, is often carried out correctly via the right
sensory and motor cortex. In brief, we are regarding the facial apraxia of
the patient with a Broca's aphasia as part of the left-sided apraxia of these
41 BRAIN—VOL. LXXXVUI
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
620 NORMAN GESCHWIND
patients. Facial apraxia is more common than apraxia of the left limbs in
this group of aphasics because a lesion producing aphasia will usually
destroy the association areas and callosal fibres involved in face movements
but will often spare those associative connexions lying more
superiorly which are involved in limb movements.
Is the association cortex involved in facial movements to command or
visual stimulation the same as Broca's area? I am not certain but some
evidence suggests that they are different. For one thing, although some
authors have chosen to regard Broca's aphasia as an apraxia of the speech
organs, it has been clear in my experience that a very severe Broca's
aphasia may be accompanied by little or no apraxia of cranial musculature;
thus, some patients carry out movements with the cranial musculature to
command and may imitate well despite marked restriction of speech. This
would suggest some difference in the anatomical arrangements for speech
and other cranial movements. Furthermore, I have already noted that
patients with conduction aphasia commonly show facial apraxia and yet
may show an aphasia which is quite different in character from a Broca's
aphasia; in particular, dysarthria may be mild or absent. This too suggests
that Broca's aphasia and facial apraxia may vary independently even
though facial apraxia is most often seen with Broca's aphasia. Whether
these differences are related to differences in the association cortex involved
remains an open question. In any case, it is probably unjustifiable to call
Broca's aphasia an apraxia if by this is suggested that it comprises part of a
picture in which facial apraxia is necessarily present. The frequent
tendency of the patient with facial apraxia to repeat the command which
he fails to carry out also suggests a difference in anatomical arrangements
for facial movements and speech. It cannot be ruled out that these
differences are the result only of different degrees of overlearning of
speech and non-speech movements.
Whole Body Movements
I will not discuss here the vexing question of what has sometimes been
called "frontal apraxia" of gait and its possible anatomical substrate. I
have not seen a clinical picture to which I would with assurance apply the
term "apraxia of gait" although I have seen many examples of frontal gait
disturbance. For a gait difficulty to qualify as an "apraxia of gait," it
would have to meet the following criteria: the patient should perform whole
body movements poorly to command or on imitation; the errors should
not merely be failures to perform or clumsiness but should also include
substitution of other well-performed movements for the desired ones. By
contrast, under certain conditions, e.g. appropriate somesthetic stimulation,
the patient should perform whole body movements perfectly well.
This syndrome has not been seen to my knowledge—or at least not
recognized. The patients generally described as manifesting frontal gait
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 621
disturbances show a fixed disturbance in performing integrated whole body
movements, and I would accept Denny-Brown's (1958) view which regards
this disturbance as involving impairment of more elementary motor
mechanisms of some type. Similarly, I have not seen apraxic disturbances
in the limbs in my cases of frontal gait disorder, i.e. my patients have made
leg movements correctly to command and on imitation; I can conceive, of
course, that in some instances apraxias of the legs might accompany a
frontal gait disturbance but it is certainly not an obligatory association.
What I will stress in this section is rather the apparently anomalous
preservation of whole body movements in patients with otherwise widespread
apraxic disturbances and will suggest a mechanism that may underlie
this finding. I have now seen several such patients. The only paper in
which to my knowledge such sparing is noted is Liepmann's (1900). The
patient had failed to carry out limb movements to command and was
thought initially to be suffering from a profound comprehension difficulty.
However, Liepmann wrote, "Against the view that comprehension of
speech was totally lost was the circumstance that the patient promptly
carried out tasks which he could execute with the whole body, such as
standing up, walking to the window or walking to the door." This brief
and incisive observation was thus the first clue that the patient was not so
severely demented as he had been regarded for some time. Liepmann did
not, however, return to the discussion of this point and did not attempt
to analyse the mechanism of the preservation of whole body movements.
It seems clear that the callosum is not necessary for the integration of
this type of bodily movement. The patient of Geschwind and Kaplan
(1962) who often failed to carry out movements to command with the
individual limbs of the left side still walked normally to command. It
seems likely that the act of walking is controlled as an integrated act at
the level of the brain-stem and that a command may descend unilaterally
to this integrating system and set it going.
In another patient whom I have investigated, there was a left parietal
glioblastoma. The patient showed no significant motor findings but
manifested a marked bilateral apraxia of face and limbs. He performed
very poorly to command, showed little or no improvement in attempting
to copy movements made by the examiner and even handled objects very
poorly; e.g. he could not with either hand hold a hammer properly to
extract a nail although this particular manoeuvre was demonstrated to him
several times. In striking contrast was his preserved ability to carry out
whole body movements. He could carry out a command such as, "Stand
up, turn around twice and then sit down again," although he had been
unable to perform such simple commands as, "Make a fist." Even when
he failed to execute a movement, e.g. a bow, correctly to command, he
immediately carried it out after the examiner had demonstrated the movement.
There was no clumsiness at all and indeed it was generally accepted that
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
622 NORMAN GESCHWIND
the patient's bow was considerably more graceful than that of the examiner.
One of the most dramatic manifestations of this discrepancy was seen
when the patient was asked to assume the position of a boxer. He immediately
assumed the boxing stance, leading correctly with the left fist. When
asked to punch he looked perplexedly at his fist. Several different terms
were then used—"punch," "jab," "uppercut," but none of these succeeded
in eliciting a response. This situation set sharply in relief his ability to
perform whole body movements in the face of marked difficulty with
movements of individual limbs. I have seen another patient who showed
a marked bilateral apraxia, worse on the left side, which involved the face
and individual limbs. This case also demonstrated a striking relative
preservation of whole body movements.
These cases have several points of interest. The grace and elegance of
some of the whole body movements in such patients rules out that their
failures in individual limb movements are the result of some general
clumsiness or inco-ordination. The preservation of whole body movements
cannot be interpreted as resulting from the greater simplicity of the
movements. It is difficult to see how, "Stand up, turn around twice and
then sit down again" is simpler than "Make a fist" but our second patient
could do the former and not the latter. Nor is it obvious that "Show me
the position of a boxer" is simpler than "Show me how a soldier salutes"
(all of these patients were ex-Servicemen). Finally these observations also
rule out the notion that apraxia of individual limbs need lead to
inco-ordination of whole body movements.
It seems to me that such a marked difference in performance of whole
body movements to command must depend on the utilization of different
anatomic arrangements than those involved in movements of individual
limbs. The bundle of Ttirck may well represent the efferent pathway for
whole body commands. In the macaque this bundle arises mostly from the
posterior part of the superior temporal gyrus (Brodal, in Jansen and
Brodal, 1954), the region corresponding roughly to Wernicke's area in
man. According to Brodal this tract is present in man and is extensive.
All authors who have described it feel that it arises from the posterior
temporal region although the posterior parts of all three temporal gyri
have been implicated by different observers.1
1
Some authors have, however, denied the existence of this tract. Thus, Whitlock and
Nauta (1956) mention that after placing lesions in the temporal lobe of monkeys, they
found no degeneration running to the pons. These authors themselves admit, however,
that their material did not include lesions in the more caudal temporal regions as is
apparent from inspection of their diagrams. This explanation may apply to other
instances of failure to demonstrate a temporo-pontine tract. Bucy and Kluver (1955)
also strongly deny the existence of this tract but also had no case of a posterior
temporal lesion. By contrast, in support of Brodal's view, Krieg (1963) points out
that in the macaque the only downward connexion of area 22 goes to the lateral cells
of the pons. It is clear that this problem deserves careful reinvestigation.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 623
In the macaque, according to Brodal, Tlirck's bundle descends via the
retro-lenticular portion of the internal capsule and the lateral part of the
cerebral peduncle to the dorsolateral pontine nuclei. These nuclei project
in turn bilaterally to the vermis of the cerebellum. The fronto-pontine
projections descend to the medial pontine nuclei which also project
bilaterally to the cerebellar vermis. Ttirck's bundle would thus appear to
project to a system which on good grounds is regarded as being involved in
the motor control of gait and whole body movements. It is of interest that
the pontine nuclei receiving afferents from the parietal and occipital lobes
do not project to the vermis; their projections go only to the cerebellar
hemisphere of the opposite side, in contrast to the bilateral projections of
the nuclei receiving connexions from the frontal and temporal lobes.
Flechsig (1901) surmised, and I would suspect correctly, that the bundle of
Tilrck subserved "motions of the body and head in consequence of
auditory impressions." Further evidence for Flechsig's supposition is
Foerster's finding (cited by Crosby et al, 1962) of adversive movements on
stimulation of area 22 in man and the occurrence of adversive seizures
with lesions in this location. It appears likely that these effects are the
result of excitation of connexions from this region travelling by Tlirck's
bundle. I would argue that it is highly likely that the carrying out of whole
body postures to command depends on this bundle. The projection from
Wernicke's area via this structure could still be intact even when the
connexions from Wernicke's area to the motor association cortex were cut
off. I would suspect that an appropriate lesion of Tiirck's bundle might
lead to that condition of "apraxia of gait" in the sense in which I used the
term at the start of this section. In the monkey bilateral lesions of this
bundle might lead to failure to perform learned whole body movements to
auditory stimuli.
Other Bilateral Movements
We have so far considered two types of bilateral movements, facial
movements and whole body movements, and analysed the special circumstances
involved in either their special impairment or their preservation in
the face of widespread impairment of individual limb movements. We now
move to another type of bilateral movement, the type in which individual
limbs are involved on both sides. We distinguish two varieties of this type
of movement, the bilaterally symmetrical movement (e.g. making circles
in the air with both hands), and the asymmetrical learned bilateral
movement, e.g. the movements of tying shoelaces.
It is often thought that callosal integration is necessary for the successful
performance of such movements. Most of our evidence so far in fact
suggests that this is not necessarily the case. Thus, the patient of Geschwind
and Kaplan (1962) who often failed to perform actions with the left limbs
to command did tie his shoelaces to command using both hands correctly.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
624 NORMAN GESCHWIND
We were perplexed by this at first but then realized that as long as each
hemisphere had learned its task such bimanual movements could be carried
out. In our case a command to tie the shoelaces would thus have been
conveyed to the left hemisphere; the right hand would then move to begin
the task. But the visual regions of the right hemisphere could then observe
the right hand perform; the right hemisphere would thus receive visual
stimulation and proceed to do its part of the task. Presumably a more
careful analysis of the latencies with which each hand began to do its task
would have helped prove this mechanism. The fact that our patient could
respond to visual stimulation correctly with either hemisphere supports
this interpretation. Our patient carried out other bimanual tasks equally
well.
Ettlinger and Morton (1963) have shown preservation of a bimanual
task in a monkey after callosal section, and I presume that a similar
mechanism is involved.
If this mechanism seems a bit unlikely, it should be recalled that intensely
complex co-ordinations of movements can be carried out by totally
separate individuals. The members of a corps de ballet succeed in carrying
on such well-integrated movements on the basis of visual stimulation and
the use of well-learned sequences. Presumably if two separate individuals
can carry out such actions, then the two halves of a cerebral hemisphere
separated from each other should be able to do so also, as they apparently
did in our patient. Dr. Ira Sherwin (personal communication) has
suggested to me that in animals a similar mechanism may be operative, i.e.
one hemisphere may learn what has previously been learned by the other,
not via the callosum or any other direct nervous connexion but rather by
visual observation of the behaviour of the other hemisphere. This notion
deserves further experimental investigation.
Presumably such a mechanism for the carrying out of complex bimanual
movements in the presence of callosal disconnexion can come into play
only when each hemisphere has separately learned its part of the task;
therefore, this type of movement should be best preserved in those individuals
in whom the right hemisphere has some significant degree of
independence in the learning of motor acts, as was the case for the patient of
Geschwind and Kaplan. In the older writings, such as those of Liepmann,
poor performance of bimanual acts is mentioned as a consequence
of disconnecting lesions. My guess would be that in many patients
bimanual acts would be poorly performed since the right hemisphere does
not in general learn motor patterns as well as the left, as I have noted in
my earlier discussion. There will, however, be cases like that of Geschwind
and Kaplan where the right hemisphere has learned its task independently.
An alternative possibility might present itself—that bimanual tasks are
learned via pathways descending from only one hemisphere and are therefore
likely to be preserved, just as whole body movements are. One patient
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 625
we have observed presents evidence against this supposition. He showed
bilateral apraxia, worse in the left hand, to verbal command, on visual
stimulation and on object handling. When asked to perform a bimanual
act, the right hand could eventually be got with difficulty to perform its
role but the left hand could not. In short each hand was as apraxic as it had
been in unimanual tasks. Yet, this patient showed a striking preservation
of whole body movements. Clearly, the mechanisms of whole body
movements and bimanual movements are different.
I would suspect that a similar mechanism explains the preservation of
symmetrical but not highly overlearned bimanual movements in patients
with callosal disconnexion. Thus when the patient is given the command
to make circles in the air, the right hand can begin and the left can
then join in. The patient of Geschwind and Kaplan showed this
phenomenon. When his eyes were closed, however, the performance
appeared to change somewhat. He then tended to move his arms at
different rates, a performance unusual in a normal and then only attained
by deliberate effort. We are not certain that we can attribute this lack of
synchrony to the absence of callosal connexions since the mild pyramidal
signs in the right arm might have affected its functioning in this task.
Further observations will be needed in future cases to confirm this result.
The Problem of "Motor" Versus "Cognitive" Learning
I would like briefly to refer here to a problem that I have avoided in the
above discussion, i.e. the classical argument in learning theory as to
whether motor learning is the formation of "stimulus-response" connexions
or whether it is the learning of a "cognitive map." Thus, has the
animal who has learned a maze acquired a "picture" of it or has he learned
a sequence of motor responses to a sequence of stimuli? The distinction is
potentially an important one for the investigation of disconnexion phenomena.
We might consider a hypothetical illustration. Suppose an animal
was taught to respond by pressing a lever when a particular visual stimulus
appeared and always used, by preference, his right forepaw in making this
response during the period of training. If the visual cortex to motor cortex
connexions were cut would the performance be lost? If the animal had
acquired a cognitive "map" then the disconnexion between visual and
motor association cortex might be unimportant. He might then be able to
use pathways descending to the brain-stem and perform the task with a
whole body movement, e.g. by sitting on the lever or by making a crude
unilateral movement which was subserved by pathways descending directly
to the brain-stem. Suppose by contrast the animal is taught to make a
series of distinct complex motor movements in response to the visual
stimulus and in fact is taught that these and only these will be rewarded.
It is much more likely that this learning will depend on the motor cortex
and that a lesion causing a disconnexion of the cortical visual and motor
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
626 NORMAN GESCHWIND
areas will permanently impair the performance of this type of task.
Another example might be the contrast between learning the pattern of a
walking maze and learning to type. The pattern of a maze once learned, a
human could get through it readily, walking, crawling, or driving. But
knowing the pattern of a typewriter keyboard alone does not solve the
problem of typing. Similarly, it will take great effort for an untrained
person to play the piano even when he has learned the simple pattern of the
arrangement of the notes.
It might appear at first that my analysis here has been exclusively a
stimulus-response analysis and hence antagonistic to the idea of cognitive
"maps." I have actually stressed the simpler stimulus-response situations
and have not considered such maps. There is no reason, however, for
assuming that these two types of "motor" learning are mutually exclusive.
It would be of great interest to study the pattern of lesions in the nervous
system of animals which would abolish or prevent the formation of such
apparently purely cognitive learning.
Similarly, I have not discussed the problems implicit in the carrying out
of such complex commands as, "Draw a star in the pink square." This
request is quite different from such commands as, "Show me how you
would use a hammer." In the first command we have three separate
elements, the auditory command, and the motor response which, however,
must be carried out under the control of vision, thus bringing a third
system into play. The anatomical pathways for such commands which
involve more than two elements represent an intriguing problem for
further study.
VI. OTHER APHASIC DISTURBANCES
I have so far not discussed that condition whose name in classical
neurology specifically reflected the theoretical assumption that it was the
result of disconnexion, i.e. conduction aphasia or Leitungsaphasie. The
term, of course, meant aphasia due to failure of conduction, i.e. aphasia
due to disconnexion. In this section there will be presented a discussion of
this aphasic disturbance as well as of certain other syndromes which are
probably the result of disconnexions.
Conduction Aphasia
Interest in this condition has been revived by the recent work of
Konorski and his co-workers (1961). Konorski's presentation at the
International Neurological Congress in Rome in 1961 alerted my
colleagues and myself to this condition and we have now seen several very
characteristic examples. In our experience the clinical picture is the
following. The patient usually shows little or no hemiplegia. His spontaneous
speech is often, but not always, copious. Dysarthria tends to be
absent or mild, and whatever phrases are produced tend to be fluent.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 627
While articulation may be normal, the speech is obviously and often
severely aphasic, usually highly circumlocutory and often grossly
paraphasic with a tendency particularly to literal paraphasias. There is
marked difficulty in naming. Writing suffers along with spontaneous
speech. As we have already noted earlier, facial apraxia to command is
often marked and may also be present on imitation.
The notable feature of these cases is the marked discrepancy between
comprehension and repetition. In the most striking of these cases comprehension
is excellent as manifested by the ability to pick out correct and
incorrect sentences, and also in many cases by the capacity to carry out
verbal commands with the limbs. In striking contrast is the difficulty in
repetition. In some instances even the simplest words fail to be repeated.
The patient often says, "Say it again" which may give the impression of his
not having heard or comprehended; yet even when there is positive
evidence of comprehension, repetition does not improve. Thus, a patient
may, on being given the word "president," say, "I know who that is—
Kennedy," but still fail to repeat. The failures may be manifested by total
inability to repeat, paraphasic repetitions or in many instances the production
of an association to the correct word.
A remarkable feature of many of these cases is the frequently preserved
ability of the patient to repeat polysyllabic numbers, e.g. "seventy-eight,"
while he fails to repeat even shorter words or repeats them paraphasically.
The contrast is brought out sharply by such phrases as, "fifty-five per cent"
where our patient said, "fifty-five progum" and "eleven plus eight" where
the patient would say, "eleven, eight . . . nineteen" but failed to produce
the word "plus" on repeated trials. "Three-quarters" was repeated as
"three-four" for several weeks. Although the patient could say the series,
"Penny, nickel, dime, quarter," he could not use the word quarter in
repeating "three-quarters." Another patient said "fifty" for "one-half"
and "seventy-five" for "three-quarters." Even when these patients fail to
repeat numbers correctly, their errors are different from those with words.
Thus, the patient tends to substitute other numbers but rarely to produce
grossly paraphasic responses and even more rarely not to repeat at all.
Thus, the patient may say "six, eight" for "sixty-eight" or "nineteen
seven" for "ninety-seven." This advantage of numbers is not confined to
the spoken modality since our first patient read printed words paraphasically
but read numbers aloud correctly whether printed as numerals
or as words. Thus, he read "train" for "travel," but correctly read "twentyeight"
just as easily as "28." It was very striking in this patient that whenever
he was presented with numbers, he manifested an immediate and
obvious relaxation of effort.
The mechanism of conduction aphasia was discussed by Wernicke
(1874). He assumed a lesion in the connexions between what later came to
be called Wernicke's area in the first temporal gyrus and Broca's area in
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
628 NORMAN GESCHWIND
the third frontal gyms. He assumed that these connexions ran in the insula
and called this "aphasia of the insular region." Some thirty years later
Wernicke (1908) discussed this problem again. He stated here that the
autopsy findings had not confirmed his view that conduction aphasia was
the result of lesions of the insula. He points out later on in this paper von
Monakow's emphasis on the importance of the fasciculus arcuatus. This
tract runs from the posterior superior temporal region, arches around the
posterior end of the Sylvian fissure and then runs forward in the lower
parietal lobe, eventually to reach the frontal lobe, and in particular Broca's
area.
Kurt Goldstein (1927, 1948) discussed this condition in great detail. He
called it not "conduction aphasia" but "central aphasia" and defined it as
"a speech disturbance, which, in the presence of relatively intact
comprehension of speech, is characterized by a disturbance in repetition,
paraphasic manifestations in spontaneous speech, reading and writing,
disturbance in spelling. . . ." He goes on to say that it appears with
lesions between the sensory and motor speech zones.1
Goldstein mentions several theories of this condition. Liepmann and
Pappenheim. (cited by Goldstein, 1927) thought that this syndrome was
only a milder variant of ordinary sensory aphasia. Goldstein points out,
however, that most of Wernicke's area was intact in their case. Goldstein
himself supported Wernicke's original view that it was a lesion of the
insula which caused this syndrome and that it was actually the insular
cortex and short association fibres which were involved. He did not,
however, stress the disconnexion between Wernicke's and Broca's areas
that Wernicke had postulated.
Potzl and Stengel (1936) pointed out in a highly detailed study the
interesting combination of conduction aphasia and pain asymboly (I have
discussed the latter condition earlier in this paper). They point out the
sparing of the insula in several cases of this condition such as the classic
case of Liepmann and Pappenheim. In addition, in their own case the
lesion did not directly destroy the arcuate fasciculus. They point out that
the lesion common to their case and other earlier cases was the involvement
of the left Heschl's gyms and the planum temporale. Their own case had
in addition a lesion of posterior insula and supramarginal gyms; these as
well as the other lesions were purely cortical. They theorized that their
1
It should be noted that Goldstein (1927) explicitly equates his central aphasia to
conduction aphasia. Thus he writes, "The symptom-picture which is usually designated
as conduction aphasia. . . . / called it central aphasia. .. ." The descriptions by Brain
(1961) and Russell and Espir (1961) appear to me not to make this distinction and to
intermix elements of other aphasias. Neither of these authors stresses the component
of disturbed repetition with relatively preserved comprehension which is the essence of
this disorder. In his later writings Goldstein (1948) continues to use this earlier
definition; he writes, "Understanding is usually preserved best, repetition and
spontaneous speech are always severely damaged. . .."
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 629
lesions probably destroyed the cells of origin of the pathway connecting
the sensory and motor speech regions which they assumed to run both via
the arcuate fasciculus and the insula. In order to test this theory a more
detailed knowledge of the precise cells of origin of the connexions from
upper temporal lobe to Broca's area and other opercular regions is needed.
One interesting feature of this case was the cortical destruction of the
supramarginal gyrus. Since there is considerable evidence that the arcuate
fasciculus consists in great part of short fibres, it is not at all unlikely that
this lesion might in effect destroy part of the pathway. They attributed, in
keeping with the discussion presented earlier, the pain asymboly to the
supramarginal gyrus lesion.
Konorski et al. (1961) have, as I have noted, revived interest in the
problem of conduction aphasia and championed the idea that conduction
aphasia results from a lesion of the arcuate fasciculus. I would think that
in the light of the preceding discussion their view is probably correct.
It has been suggested that the development of language in man depends
on his possession of an arcuate fasciculus, while a monkey does not have
this pathway. This seems at first a reasonable conclusion from the findings
of Bailey, Bonin, Garol and McCulloch (1943a) who show no arcuate
fasciculus in their diagram of the long association fibres in the chimpanzee.
There are, however, several reasons for rejecting this conclusion: (1) They
were studying only long association fibres. Since it is possible, as we have
noted earlier that much of the arcuate fasciculus consists of short fibres
(Crosby et ah, 1962), these would not have been demonstrated in the study
of Bailey et al. (2) There are, however, also long arcuate fibres revealed in
later studies by some of Bailey's pupils (Sugar, French and Chusid, 1948)
which demonstrated unquestioned evidence of fibres from the temporal
operculum to the parietal and frontal opercula. (Their diagram does not
show these fibres arching around the back end of the Sylvian fissure. Since
they were using strychnine neuronography, there would have been no way
for them to know the course of the fibres involved but only the origins and
terminations.) These had been missed in the earlier study because Bailey
et al. had not explored the depths of the sulci. The study of Sugar et al.
certainly proves the origin of such fibres in the supratemporal plane.
Whether these definitely arise from primary auditory cortex or from
association cortex is not completely clear because of the great crowding of
these structures in the small primate brain. (3) Bailey and Bonin (1951)
themselves state, "Physiological neuronography presents evidence for both
uncinate and arcuate bundles.. . . Whether the firing of the inferior frontal
gyrus and the parietal operculum by strychninization of the first temporal
convolution, that of the inferior frontal gyrus from the second and of the
middle frontal gyrus from the inferior temporal gyrus (Petr, Holden, and
Jirout, 1949) is due to conductions in the arcuate or uncinate fascicle
cannot be decided at present; that the last, if not the latter two, observa-
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
630 NORMAN GESCHWIND
tions should be ascribed to the arcuate fascicle seems more plausible. The
firing of the middle temporal gyms from the inferior parietal lobule as well
as the firings of the inferior parietal lobule from the inferior frontal gyrus
can be taken as further evidence for 'arcuate' fibres."
There is one further theory of the mechanism of conduction aphasia, one
first mentioned by Liepmann and Pappenheim but defended most recently
by Kleist (1962). This author argued that conduction aphasia occurs in
patients in whom there has been actual destruction of the left temporal
speech area itself. This lesion would in most instances lead to profound
comprehension defect but does not do so in certain people because their
right temporal region can take over this function. He assumes that these
patients, however, still must rely on the left Broca's area for speech. The
pathway from the right temporal speech area to the left Broca's area is,
however, interrupted in these patients by the destruction of the left temporal
lobe and therefore repetition is poor. Kleist presents 4 cases in support of
this thesis. The first case, Spratt, is particularly striking since the destruction
of the left superior temporal region was indeed so extensive as to be
expected to give rise to profound incomprehension. This case would
appear to be strong evidence for Kleist's contention. It is further of
interest that this patient was ambidextrous, using his left hand for many
complex functions. By contrast the lesion of his fourth case, Treusch,
involved the left temporal speech area very little and appeared to involve
deep white matter, particularly in the inferior parietal lobule. This case
would appear to be much better evidence for the theory of involvement of
arcuate fasciculus than for Kleist's own theory. In fact, Kleist's own
diagram implies that even if the mechanism he suggests is correct in some
cases, involvement of arcuate fasciculus should still produce the same
syndrome. In summary, it appears that Kleist has good evidence for his
mechanism in at least some cases of conduction aphasia. I would think
that if there are two forms of the disorder, they might well be distinguishable
clinically. We are attempting in our own cases of conduction aphasia
to ascertain whether the right hemisphere really plays a role in any of them,
as asserted by Kleist.
A brief note is appropriate on the already mentioned observation that
number repetition is better than word repetition in many of these cases. It
is conceivable that this is the result of sparing of a different anatomical
pathway for numbers. It would seem unlikely, however, that numbers are
normally spoken over a different pathway from words. I believe that a
more likely explanation comes from a consideration of how any repetition
is carried out in these patients. Thus, although the patient fails to repeat
the correct word he is likely to give an association of this word. In fact,
in some cases when the correct word itself is repeated, it is likely that this
has been accomplished only by way of associations. One patient on being
given his own last name first replied with the given names of his brothers
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 631
and himself and then finally gave the family name. Had the intervening
names not been spoken aloud the fact that the eventual correct repetition
was by way of associations and not "direct" could easily have been overlooked.
This type of associative repetition probably does depend on
finding a "detour" around the damaged normal direct pathway from
Wernicke's to Broca's areas. It is natural that this should lead to many
errors in repetition of words which in general arouse a large reservoir of
associations. The associations to numbers are likely to be more limited,
and in fact are not at all unlikely to be confined to the number itself. The
errors in number repetition support this theory; thus, "six, eight" for
"sixty-eight" and even more dramatically "three-four" for "three-quarters,"
"fifty" for "one-half" and "seventy-five" for "three-quarters." The
marked difficulty in nonsense syllable repetition stressed by Konorski
would result from the paucity of associations to such material. In brief, I
think that these patients probably do not repeat either numbers or words
via the normally used pathways but that in response to a heard number or
word, the patient gives an associative response; the associative response is
more likely to resemble the original if this is a number but not a word. One
might say that there is "pseudo-repetition" of numbers. It is interesting
that, as pointed out by Dr. Susan Ervin (personal communication),
"parrot-like repetition" (i.e. precise repetition of what has been said) is a
late development in children acquiring language. Contrary to ordinary
views such precise repetition may well be an advanced activity dependent
on specialized use of the pathways between the posterior and anterior
speech regions.
The Case of Bonhoeffer
Bonhoeffer (1914) described a remarkable case which, although it
presents certain difficulties in interpretation, I believe is worth presenting
here. Although it is apparently unique, I suspect that other such cases have
been overlooked by other observers.1
Bonhoeffer's patient sustained a
transient right hemiplegia, the arm subsequently improving faster than the
leg. There was, however, permanent reduction of speech to one or two
words; verbal comprehension was much better preserved but reading and
writing were both very poor. The patient showed difficulties in carrying
out verbal commands, in imitation and in object handling, all these disturbances
being much more marked on the left than on the right. At postmortem
there was an infarct involving the posterior parts of the left
superior and middle frontal gyri and the anterior four-fifths of the left side
of the corpus callosum. There was an infarct of the anterior limb of the
1
In fact, after having read Bonhoeffer's observation, I have come to suspect that a
case I saw several years ago probably was an example of the same disturbance but that I
completely missed its significance at the time. Chance favours, as is well known, the
prepared mind; Bonhoeffer's was obviously very well prepared for this patient.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
632 NORMAN GESCHWIND
left internal capsule; finally a small lesion was present under the left
parieto-occipital region. The cause of death was a fresh haemorrhage in
the lenticular nucleus on tbe right but the right hemisphere was otherwise
intact. For obvious reasons Bonhoeffer attributed the severe apraxia of
the left arm to the callosal lesion. To explain the aphasic disturbance of
verbal expression, he pointed out that the infarct of the anterior part of
the internal capsule had cut the descending pathways from the speech area.
The transient hemiplegia was, of course, probably due to transitory neighbourhood
effects of this lesion. This capsular lesion alone would, of course,
produce no aphasia; nor would the frontal lesion which clearly spared
Broca's area. Bonhoeffer pointed out that normally no aphasia is seen
with a capsular lesion. In the presence of such a lesion an outflow path
still exists from Broca's area running across the callosum to the corresponding
cortical area on the right, and from there eventually reaching the
right internal capsule. The callosal infarct had, however, cut off this
alternative pathway and Broca's area was isolated.
My own inclination would be to agree on the whole with this
interpretation of Bonhoeffer's. Thereremain to be explained the mild rightsided
apraxia as well as the alexia and agraphia. Bonhoeffer himself was
uncertain as to their cause, but I would regard them as having been caused
by the lesion of left motor association cortex as well as by the lesion
under the left angular gyrus. It is important to note that the latter
was a small lesion which Bonhoeffer felt would not be very significant
clinically. Bonhoeffer, however, then goes on to express his agreement with
von Monakow's view that combinations of lesions may produce clinical
pictures that none of the lesions could cause in isolation. (I have, of course,
given several examples of such combinations.) The left angular gyrus
lesion may have been much more disabling in this patient because of the
co-existing extensive callosal lesion, which must have cut off most of the
callosal connexions of the parietal association areas.
Echolalia
Kurt Goldstein (1917, 1948) has shown great interest in this symptom;
Stengel (1947) has written an excellent paper on the functional aspects of
echolalia. A characteristic clinical picture is that of marked preservation of
repetition, indeed automatic repetition, in the face of gross difficulty in
comprehension. I would stress that what is being spoken of here is a
retention of repetition which is in sharp contrast to the disturbance in
comprehension; other explanations probably are forthcoming for the
common situations in which repetition is only moderately better than
comprehension. To explain the sparing of repetition in such cases on the
basis that it is "simpler" runs counter to the fact that repetition can be
especially impaired, as in conduction aphasia.
Goldstein stressed that for this syndrome to be present it was necessary
that Wernicke's area, Broca's area and the connexions between them be
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 633
intact. This intactness of the speech area and of its afferent auditory connexions
guarantees repetition. Coupled with this intact speech area there
must be a lesion which isolates the speech area from much of the rest of the
cerebral hemispheres, i.e. a large parietal lesion. It is the isolation of the
speech area which tends to produce this picture. There is no comprehension
because language arouses no associations; there is gross disturbance
of spontaneous speech since the speech area receives no information from
elsewhere in the brain.
Segarra and Quadfasel (1961) have recorded the post-mortem findings of
a most interesting patient who had been studied extensively by
Quadfasel. This patient, a woman of 32 years of age, suffered severe
illuminating gas poisoning, following which she survived for almost ten
years. During this period she never showed any evidence of comprehension
of language, nor ever spoke spontaneously a propositional phrase, but
uttered only expletives. She could do nothing for herself. By contrast, she
echoed what was spoken to her and even more strikingly she echoed songs.
In fact, she was even able to learn new songs not in existence before the onset
of her illness. Following her death her brain was cut in whole-brain
sections and stained for cells andmyelin. Preliminary surveyofthese sections
has shown intactness of the speech area (Wemicke's and Broca's areas) and
of the arcuate fasciculus and insula. In addition the hippocampal region is
intact as well as connexions to it within the temporal lobe. The
corresponding structures on the right side are symmetrically preserved. By
contrast there is extensive infarction extending in mantle form around these
intact regions. This patient had essentially an intact speech area which was
able to carry on repetition. It would even manifest verbal learning because
of the preservation of the connexions to the hippocampal region.
Comprehension and propositional speech were, however, lacking for the
reasons which I have already presented.
Cases of isolation of the speech area such as these are most valuable in
elucidating the functions of this region. It is most remarkable that this
patient could learn verbal material. This task is intensely difficult for
patients with partial lesions within the speech area who may show
syndromes far less devastating than that revealed by this patient. It also
suggests that language is not "comprehended" by Wernicke's area but
rather that this region serves to arouse associations elsewhere probably by
way of the inferior parietal region. It illustrates also that speech is not
created in Wernicke's area; rather it serves to transform what has come
from the remainder of the brain into language.
VII. POSSIBLE OBJECTIONS AND PITFALLS
The Results of Akelaitis and His Co-workers
I would like to deal briefly here with the objections that might be raised
against this theory of the importance of lesions of association cortex and
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
634 NORMAN GESCHWIND
association fibres in producing disturbances of the higher functions of the
nervous system. These objections arise chiefly from the results of the
generally negative investigations made by Akelaitis and his co-workers
(Akelaitis, 1941a, 19416, 1941c, 1942a, 19426, 1943, 1944, 1945; Akelaitis
et aI., 1942) in cases with surgical section of the corpus callosum for the
treatment of epilepsy. I have discussed this problem briefly elsewhere
(Geschwind, 1962). I believe that the patients were well examined;
inadequate examination is therefore probably not the explanation for the
discrepancies between Akelaitis's results and many of those which I have
cited.
The majority of Akelaitis's patients had had cerebral lesions dating from
early childhood. Such early lesions might well explain the absence of
cerebral dominance for language. Thus, many of the functions which are
normally unilateral could probably be carried on by both hemispheres. A
similar explanation probably applies to the lack of such syndromes in cases
of agenesis of the corpus callosum. The early lesion is more likely to lead
to functional re-organization than lesions in adult life (as examples of this
one may cite the work of Kennard, 1942 and of Scharlock, Tucker, and
Strominger, 1963), i.e. it is more likely to lead to use of other
pathways.
A second factor is the presence of epilepsy in nearly every case in
Akelaitis's series. It is certainly highly possible that seizures may lead to
the "learning" of new pathways (I have commented earlier in the paper on
Morrell's work on the parallels between epilepsy and other learning).
Similarly, Penfield and Boldrey (1939) wrote, "An habitual seizure, by
virtue of its frequent repetition, may eventually establish a complicated
neurone pattern . . . (The spread of a seizure) is not diffuse but. . . along a
definite neurone system which may be preformed or acquired. By 'preformed'
is meant a system of neuronal connexions ordinarily recognized as
physiologically functional in normal brains. By 'acquired' is meant a
pattern of neuronal connexions established by the conditioning influence
of previous experience. ... In complicated seizures the advance of
discharge along this neurone pattern may be so slow and episodic as to
suggest that isolated ganglionic collections are fired in an advancing
series. . . ." The general failure of callosal section as a therapeutic agent
in Akelaitis's cases is compatible with the seizures having long since
"learned" some complex or unusual pathways or having "learned" such
new pathways after the transection. Erickson (1940) pointed out that
section of the corpus callosum in monkeys prevented electrographic spread
of the seizure to the opposite hemisphere and altered the pattern of the
seizure. Whether electrographic spread to the opposite hemisphere might
have continued to take place despite transection of the corpus callosum if
the experiments had been done on chronically epileptic monkeys rather
than in acute experiments is not known. Such an experiment would,
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 635
however, be of great interest in determining whether seizures "open" less
commonly used pathways to the other hemisphere. The second part of
Erickson's experiment seems to confirm strongly that with a lesion
confined within the bounds of a hemisphere seizures rapidly "find" new
paths to circumvent the obstruction. Thus, in one monkey a seizure initiated
in the right cortical arm area spread in the intact brain to the right leg area,
left leg area, and left arm area. After a wide section at right angles to the
Rolandic cortex superior to the right arm area, the seizure travelled from
the right arm area to the left arm area then to the left leg area and finally
to the right leg area. Section of the callosum abolished this new mode of
spread.1
As an example of the joint effects of seizures and early lesions, we may
compare the cases reported by Akelaitis in whom there had been section of
the splenium, and those reported by other authors. There were 6 such cases
in Akelaitis's series, all of them severely epileptic, 5 with lesions from early
childhood. None developed alexia in the left visual field following section
of the splenium. By contrast the case of Trescher and Ford (1937) and the
2 cases of Maspes (1948) all had colloid cysts of the third ventricle and none
had had seizures. All three developed alexia in the left visual field as a
consequence of having had the splenium sectioned.
The notion that repeated seizures favour the opening of less used
pathways leads to several interesting implications. One should be able in
experimental animals to overcome the effects of disconnecting lesions by
repeated seizures. Thus, an animal with the callosum sectioned who has
had repeated seizures for a long period before callosal section might show
interhemispheric transfer after surgery while a control without seizures
might not. It goes without saying that the long-term effects of section of
the callosum should be less disturbing if carried out in infant animals than
in adults.
It is remotely possible that the effects of repeated seizures in favouring
the opening of new pathways, if verified, might be useful therapeutically.
Repeated induced seizures might favour recovery from disturbances
following lesions of association cortex or fibres. Obviously, such a
1
The term "arm area" does not necessarily mean area 4; the electrical disturbance
was probably, in my opinion, going via area 6. This discussion is not meant to suggest
that seizures spread only between cortical regions. Certainly some of Erickson's
observations in the above paper confirmed that bilateral involvement of the body can
occur with the callosum sectioned—but this is different in type from the kind of bilateral
spread seen when the callosum is present. Obviously bilateral spread with the callosum
sectioned depends on involvement of subcortical structures. Gastaut and FischerWilliams
(1959) have discussed such subcortical spread extensively in their review. One
of the major reasons for the failure of surgery in Akelaitis's patients may have been
such spread to subcortical motor systems rather than spread to the opposite hemisphere
via "new" subcortical pathways. We do not have enough information to decide which
is the explanation.
42 BRAIN—VOL. Lxxxvm
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
636 NORMAN GESCHWIND
therapeutic investigation could be justified only after experimentation
had confirmed the effectiveness of such procedures in animals.1
There are two other but less likely explanations for Akelaitis's results.
One is that alternative pathways to the opposite hemisphere via
subcortical routes are readily available in most people and that these
syndromes therefore appear only exceptionally. The other is that the
syndromes had cleared by the time the patients were tested. Although the
patient of Geschwind and Kaplan (1962) did show definite improvement in
his callosal disconnexion symptomatology over several months, it was still
evident after this period. Furthermore, acquired lesions of many other
association pathways may show no improvement with time. Restitution
after white matter lesions might conceivably be less damaging permanently
than lesions of the corresponding cells of origin; as long as the cells are
intact, there is a possibility that collateral pathways might be brought into
play. We know, however, almost nothing of the pathways of secondary
importance.
This leads us to the greatest advantage but at the same time the greatest
danger of thinking in terms of disconnexions. Theories of this type are
rich in readily testable theoretical implications. They can readily
degenerate into naively anatomical systems. This is, however, not inherent
in the structure of this approach and there is no conflict between this kind
of reasoning and sophisticated psychological analysis. In fact, this approach
frequently demands a reinvestigation of standard psychological categories
and opens new ways of looking at psychological phenomena. Perhaps the
greatest danger is that of "working backwards" and of inventing pathways
to correspond to every difference in behaviour. I have tried to show here
that while some differences in behaviour are probably based on anatomically
different pathways (e.g. the discussion on isolated limb movements
versus whole body movements) others are dependent on non-anatomical
factors (e.g. the discussion on the relative preservation of number repetition
in conduction aphasia). I can only agree with the hope expressed by Adolf
Meyer (1905) for "convincing observations of patients . . . with such
anatomical examination as will put an end to the regrettable tendency of so
many clinicians to consider the white matter of the hemispheres the
cornucopia of all the desirable conduction paths."
VIII. PHILOSOPHICAL IMPLICATIONS
It is not my intention to embark on an extensive philosophical discussion
in this section. It seems reasonable, however, to sketch rapidly the
implications of many of the results discussed here for the philosophical
foundations of study of the higher functions of the nervous system.
1
To forestall any questions in advance I am in no way suggesting that the effects of
electric shock therapy are related to such a mechanismI
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 637
The Whole Man
For the past forty years there have been schools of thought which have
stressed the importance of thinking of the patient as a whole, of seeing his
responses as those of an integrated unitary structure, even in the face of
damage. The ramifications of this thinking in neurology, psychiatry,
psychology and other fields must be well known to most readers. It should
be clear from much of our discussion that this principle, while it may be
useful in some cases as a stimulus, may be actively misleading when it is
regarded as a philosophical law. When Edith Kaplan and I were studying
our patient, we constantly found that many confusions about the patient in
our own minds as well as those of others resulted from failure to do the exact
opposite of what the rule to look at the patient as a whole demanded, i.e.
from our failure to regard the patient as made of connected parts rather
than as an indissoluble whole. We were constantly dealing with questions
such as "If he can speak normally and he knows what he's holding in his
left hand why can't he tell you?" We had to point out that we couldn't
say that "the patient knew what was in his left hand" and that "the patient
could speak normally," since that part of the patient which could speak
normally was not the same part of the patient which "knew" (nonverbally)
what was in the left hand. This is at first blush an odd way to speak
—it is hard not to say "the patient" and yet it is clear that this terminology
is misleading.
We have little difficulty with the concept of disconnexion at lower levels.
If the spinal cord is transected we are usually capable of treating it as
separate from the higher centres. We may say of such a person, "The
patient urinated" since we know that in this case this means "The patient
urinated involuntarily." Although "the patient urinated" is at first glance
ambiguous, in practice we understand and do not find it necessary to use
clumsy locutions such as "The patient's spinal cord urinated." We get into
difficulties, however, with disconnexions at higher levels since we do not
expect highly organized activities dependent on learning to be carried on in
disconnected parts of the brain. We must become accustomed to thinking
in this way in order to understand some of the more complex disturbances
consequent on lesions of the brain. I am not advancing "the atomistic
approach" as a basic philosophical postulate to replace "the holistic
approach," but am rather suggesting that failure to consider the applicability
of either type of analysis will in one situation or another lead to
errors. It should be pointed out that the usefulness of sometimes considering
animals or humans not as a unit but as a union of loosely joined wholes
need not apply only to disease states although probably it will find its
greatest use in that situation. Probably even in the normal person parts of
the brain are so weakly connected as to make their interaction difficult. I
have suggested that some connexions are normally present in adult man
which are absent or of less extent in lower animals; these connexions may
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
€38 NORMAN GESCHWIND
take years to develop in some children. Perhaps an adult man is more
unitary than a chimpanzee, but perhaps total unity is never obtainable
kecause of the necessary separation of some structures.
The Unity of Consciousness
A corollary to the above discussion is that it forces us to be somewhat
more precise than we have been about "the unity of consciousness." It
would no doubt be startling to suggest that the patient of Geschwind and
Kaplan had separate consciousness in each hemisphere; it would on the
other hand be a little difficult to understand just what would be meant by
saying that his consciousness was unitary. If the ability to give a verbal
account is a prerequisite of consciousness then only the left hemisphere was
conscious; if the ability to respond in a highly organized manner and to use
the results of past experience constitutes consciousness then he had
multiple consciousness. Perhaps there are better criteria of consciousness
than these; this case and similar ones only re-emphasize the necessity of
re-evaluating the idea of the unity of consciousness if it is to be at all useful.
I believe that Kurt Goldstein was perhaps the first to stress the nonunity
of the personality in patients with callosal section and its possible
psychiatric effects. Thus he wrote (Goldstein, 1927), "The separation of so
large a part of the brain and the resulting impossibility of evaluating
stimuli perceived with the right hemisphere . . . surely cannot be without
effect on the total personality. . . . I have pointed out the presence in my
patient of a feeling of strangeness in relation to movements of the left hand,
which she described with such curious expressions (she would.say that someone
was moving her hand and that she wasn't doing it herself) that she was
regarded at first as a paranoiac. It appears to me not to be excluded that
on this basis and under certain conditions there may develop paranoid
states, perhaps also the experience of doubled personality and above all the
experience of being influenced from without. . . . "
The Value of Introspection
Still another corollary to the observation that parts of the brain may be
disconnected is the conclusion that introspection may be an extremely
ineffective way of obtaining information about many of the patient's
•experiences. I have already made this point in the presentation of some of
the syndromes of disconnexion, e.g. in respect of the colour-naming
•disturbance of pure alexia without agraphia. If a part of the brain is fully
disconnected from the speech area it will not be possible for the speech area
to give an account of what goes on in that part of the brain. The patient
with a colour-naming disturbance can give only a poor account of his
•colour-experience or indeed none at all since his speech area has little or
no access to information about the colour-experiences of the visual cortex.
•Other examples were cited illustrating the same principle. I have also
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 639
discussed in some detail the tendency for the patient to show confabulatory
response to demands for introspective observations. It is certainly well
known that even in normals introspection may be misleading or incorrect.
Let me point out carefully that what is presented here is not the gross
behaviouristic assertion that introspective information is in all situations
uesless nor that one should never listen to what the patient is saying. The
conclusion being presented is the less extensive one that there are certain
situations, particularly in the presence of lesions of the brain, in which the
patient for simple anatomical reasons is incapable of recounting verbally
the experiences of parts of his nervous system which are functioning at a
complex discriminative level.
Language and Thought
An old problem is the one of the extent to which language controls one's
"perception" of the world. Whorf (1956) suggested an extreme form of the
view that language influences what is perceived. The view at the opposite
extreme is the naive "natural" one that language is an infinitely flexible
tool for describing without prejudice the impact of the world on one's
nervous system. The behaviour of our patient with colour-naming defect is
of interest here. Despite his inability to name colours correctly, he had no
difficulty in sorting, and did so by colour. This, of course, does not
necessarily reflect the behaviour of a "naive" nervous system since the
visual region may have been trained by earlier verbal experience. It is
clear at least that words need not be available for the sorting process to
take place.1
On the other hand, as I noted in my discussion of secondary
sorting errors, the use of the incorrect word may lead to errors in sorting
under the conditions of certain types of instruction. This type of "secondary"
error might be more marked in tests where the subject is required to
remember colours after an interval since it seems reasonable that such
memory is frequently mediated verbally (Brown and Lenneberg, 1954).
The study of patients with such specialized disturbances may aid in further
exploring the relationship of language to cognition.
SUMMARY
A complete summary of all the material presented would be much too
extensive and, indeed, much too repetitious. I will therefore try to outline
here the major points presented in Part I (Geschwind, 1965) as well as
Part II. I have attempted to show that many disturbances of the higher
functions of the nervous system, such as the aphasias, apraxias, and
1
Some informal experiments on children suggest that perceptual differentiations
precede naming. My 4-year-old son misnamed colours but could sort correctly and
could trace out correct numbers on the Ishihara test; although I could name the colours
of dots, I did much poorly than he did on these tests. His difficulty was one of colournaming,
mine was partial colour-blindness.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
640 NORMAN GESCHWIND
agnosias may be most fruitfully studied as disturbances produced by
anatomical disconnexion of primary receptive and motor areas from one
another. For a detailed discussion the reader is referred to the appropriate
sections of the paper.
In the lower mammals connexions between regions of the cortex may
arise directly from the primary receptive or motor areas. As one moves up
the phylogenetic scale, these connexions come to be made between newly
developed regions of cortex interspersed between the older zones. These
regions are called "association cortex." As Flechsig pointed out for the
human brain, all intercortical long connexions (whether in or between
hemispheres) are made by way of these association areas and not between
the primary motor or receptive areas. It follows from this that lesions of
association cortex, if extensive enough, act to disconnect primary receptive
or motor areas from other regions of the cortex in the same or in the
opposite hemisphere.
The connexions of the visual association regions were discussed in some
detail, and it was pointed out that the major outflow of these regions is to
the lateral and basal neocortex of the temporal lobe which in turn
connects to limbic structures. Lesions of the lateral and basal temporal lobe
therefore tend to disconnect the visual region from the limbic system. This
leads to a failure of visual stimulation to activate limbic responses, such as
fight, flight, and sexual approach. It also leads to difficulties in visual
learning. These can be thought of as resulting from the failure of the
animal to form visual-limbic associations (such as learning that a visual
stimulus equals the food reward given for correct choice) because of the
lack of appropriate connexions. They can also be regarded as disturbances
in visual recent memory resulting from a disconnexion between the visual
region and the hippocampal region. The discussion was then applied to
the tactile and auditory systems. Learning difficulties in primates involving
these systems also were thought to result from disconnexions from the
limbic system. "Agnosia" in the sense of failure to respond to stimuli
within a single modality appropriately in the face of intact perception in
that modality is regarded as being a part of the syndrome of disconnexion
of primary sensory modalities from the limbic system. Since callosal fibres
arise from association cortex, failures of interhemispheric transfer
may result from lesions of association cortex. The problem of whether
disconnexions of single modalities from the limbic system in man occur
was briefly discussed.
While connexions between primary receptive regions and limbic structures
are powerful in subhuman forms, intermodal connexions between
vision, audition, and somesthesis are probably weak in these animals, a
view for which evidence is available both on the basis of experimental
behavioural investigations (e.g. studies on intermodal transfer of learning
and on higher-order conditioning) and on the basis of anatomical evidence.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 641
In man the situation changes with the development of the association
areas of the human inferior parietal lobule, situated at the junction of the
older association areas attached to the visual, somesthetic, and auditory
regions. It is speculated that this new "association area of association
areas" now frees man from the dominant pattern of sensory-limbic
associations and permits cross-modal associations involving non-limbic
modalities. It is particularly the visual-auditory and tactile-auditory
associations which constitute the basis of the development of speech in
most humans. In man the speech area (which constitutes the auditory
association cortex, particularly that part of it on the convexity of the
temporal lobe, also Broca's area and the connexions between these
regions) becomes a structure of major importance in the analysis of all the
higher functions.
Pure word-blindness without agraphia was then discussed as an excellent,
classical example of disconnexion from the speech area; this syndrome
results from a combination of lesions, the usual one being destruction of
the left visual cortex and of the splenium of the corpus callosum. The
association with this syndrome of colour-naming difficulties and inability
to read music is noted, along with the relatively strong preservation of the
reading of numbers and the naming of objects. Reasons are advanced for
these discrepancies. The problem of childhood dyslexia and its associated
disturbances and its possible relation to the acquired dyslexia of adults was
briefly presented.
Other disorders with similar pathogenesis (isolation of a particular
sensory modality from the speech area), i.e. pure word-deafness and tactile
aphasia were then briefly discussed.
The problem of the "agnosias" was then presented. Evidence was
presented against the idea that there exist disturbances of "recognition"
regarded as a unitary faculty. It was argued that most of the "agnosias"
are in fact modality-specific naming defects resulting from isolation of the
primary sensory cortex from the speech area and associated with marked
confabulatory response. A critique was presented of the classical "aphasicagnosic"
distinction. There was presented some further discussion on the
determinants of confabulatory response. The problem of right parietal
syndromes was presented in the light of the preceding discussion of the
"agnosias."
"Apraxic" disturbances were analysed in detail and were regarded as
resulting from disconnexions of the posterior speech area from association
areas which lie anterior to the primary motor cortex, and from
disconnexions of visual association areas from these "motor association"
areas. The problem of left-sided predominance was discussed. In particular
apraxic disturbances resulting from callosal lesions, from lesions of "motor
association" cortex and from damage deep to the supramarginal gyrus
were discussed. The apraxia of the left side ("sympathetic dyspraxia") of
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
642 NORMAN GESCHWIND
aphasic right hemiplegics was discussed as well as facial apraxia. The
sparing of certain types of movement in the apraxias was discussed,
particularly whole body movements and was related to the probable
preservation of Tlirck's bundle (whose connexions were presented in
some detail) running from the posterior temporal region to the pontine
nuclei and then via synapses to the cerebellar vermis.
Finally syndromes resulting from disconnexions within the speech area
(conduction aphasia) and the pattern resulting from the isolation of the
speech area were presented.
Some classical objections to the disconnexion approachwere presented, in
particular the results of Akelaitis and reasons for his negative results were
discussed. This section closed by pointing out that this type of theory
suggests many experiments and anatomical investigations. The dangers of
ad hoc postulation of connexions were mentioned.
In a short section attention was called to some philosophical
implications of these findings, particularly for the notions of "regarding
the patient as a whole man," the unity of consciousness, the uses of
introspection and the relations between language and one's view of the
world.
I would like to express my deep gratitude to my secretary, Mrs. Ceoria M. Coates,
who has so effectively dealt with the burdensome task of dealing with the preliminary
versions of this paper.
BIBLIOGRAPHY
AJURIAGUERRA, J. DE, HECAEN, H., and ANGELERGUES, R. (1960) Rev. neurol., 102, 566*
AKELATTIS, A. J. (1941a) Amer. J. Psychiat., 97, 1147.
(19416) Arch. Neurol. Psychiat., Chicago, 45, 788.
(1941c) Amer. J. Psychiat., 98, 409.
(1942a) Arch. Neurol. Psychiat., Chicago, 48, 108.
(19426) Arch. Neurol. Psychiat., Chicago, 48, 914.
(1943) /. Neuropath., 2, 226.
(1944) /. Neurosurg., 1, 94.
(1945) Amer. J. Psychiat., 101, 594.
, RISTEEN, W. A., HERREN, R. Y., and VAN WAGENEN, W. P. (1942) Arch.
Neurol. Psychiat., Chicago, 47, 971.
ARING, C. D. (1944) In "The Precentral Motor Cortex." Edited by P. C. Bucy,
Urbana, p. 409.
BAILEY, P., and BONTN, G. (1951) "The Isocortex of Man." Urbana.
, , GAROL, H. W., and MCCULLOCH, W. S. (1943a)/. Neurophysiol., 6,
121.
, , , (19436) /. Neurophysiol., 6, 129.
, , and MCCULLOCH, W. S. (1950) "The Isocortex of the Chimpanzee."
Urbana.
BINOLEY, T. (1958) "Mental Symptoms in Temporal Lobe Epilepsy and Temporal Lobe
Gliomas." Copenhagen. (Ada psychiat., Kbh. Supp: 120.)
BONHOEFFER, K. (1914) Mschr. Psychiat. Neurol, 35, 113.
BONIN, G., and BAILEY, P. (1947) "The Neocortex of Macaca Mulatta." Urbana.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
DISCONNEXION SYNDROMES IN ANIMALS AND MAN 643
BRAIN, W. R. (1941) Brain, 64, 43.
(1961) "Speech Disorders." London.
BROWN, R. W., and LENNEBERG, E. H. (1954) /. abnorm. soc. Psychol., 49, 454.
BUCY, P. C. (1944) "The Precentral Motor Cortex." Urbana.
, and KXOVER, H. (1955) /. comp. Neurol., 103, 151.
CHUSID, J. G., SUGAR, O., and FRENCH, J. D. (1948) /. Neuropath., 7, 439.
COLE, M., and ZANGWILL, O. L. (1963) /. Neurol. Neurosurg. Psychiat., 26, 37.
CRTTCHLEY, M. (1962) In "Interhemispheric Relations and Cerebral Dominance."
Edited by V. B. Mountcastle, Baltimore, p. 208.
CROSBY, E. C, HUMPHREY, T., and LAUER, E. W. (1962) "Correlative Anatomy of the
Nervous System." New York.
DENNY-BROWN, D. (1958) /. nerv. ment. Dis., 126, 9.
(1962) In "Interhemispheric Relations and Cerebral Dominance." Edited by
V. B. Mountcastle, Baltimore, p. 244.
ERICKSON, T. C. (1940) Arch. Neurol. Psychiat., Chicago, 43, 429.
ETTLINGER, G., and KALSBECK, J. E. (1962) /. Neurol. Neurosurg. Psychiat., 25, 256.
, and MORTON, H. B. (1963) Science, 139, 485.
, and WYKE, M. (1961) /. Neurol. Neurosurg. Psychiat., 7A, 254.
FLECHSIG, P. (1901) Lancet, 2, 1027.
Forx, C. (1916) Rev. neurol., 29 (0, 283.
GASTAUT, H., and FISCHER-WILLIAMS, M. (1959) In (American Physiological Society)
"Handbook of Physiology." Washington. Section I, VoL 1, p. 329.
GESCHWTND, N. (1962) In "Reading Disability." Edited by J. Money, Baltimore.
(1963a) Trans. Amer. neurol. Ass., 88, 219.
(19636) Trans. Amer. neurol. Ass., 88, 174.
(1965) Brain, 88, 237.
, and Fusnxo, M. (1964) Trans. Amer. neurol. Ass., 89, 172.
, and KAPLAN, E. (1962) Neurology, 12, 675.
GOLDSTEIN, K. (1908) /. Psychol. Neurol., Lpz., 11, 169 and 270.
(1917) "Die transkortikalen Aphasien." Jena.
(1927) In Bethe, A., and v. Bergmann, G. "Handbuch der Normalen und
Pathologischen Physiologic" VoL 10, Berlin.
(1948) "Language and Language Disturbances." New York.
HARTMANN, F. (1907) Mschr. Psychiat. Neurol., 21, 97 and 248.
HECAEN, H., and ANGELERGUES, R. (1963) "La Ce"cit6 Psychique." Paris (Masson).
, and GIMENO ALAVA, A. (1960) Rev. neurol., 102, 648.
JACKSON, J. HUGHLINGS (1878) Lancet, 1, 716. Reprinted in "Selected Writings of
John Hughlings Jackson." Edited by J. Taylor, Vol. 1,1958. New York (Basic
Books).
(1880) Brain, 3, 192. Reprinted in "Selected Writings of John Hughlings
Jackson. Edited by J. Taylor, Vol. 1, 1958. New York (Basic Books).
JANSEN, J., and BRODAL, A. (1954) "Aspects of Cerebellar Anatomy." Oslo.
KENNARD, M. A. (1939) Arch. Neurol. Psychiat., Chicago, 41, 1153.
(1942) Arch. Neurol. Psychiat., Chicago, 48, 227.
KIMURA, D. (1963) Arch. Neurol., 8, 264.
KJNSBOURNE, M., and WARRTNGTON, E. K. (1962) /. Neurol. Neurosurg. Psychiat., 25,
339.
KLEIST, K. (1962) "Sensory Aphasia and Amusia." New York.
KONORSKL. J., KOZNIEWSKA, H., and STEPIEN, L. (1961), Proceedings of the VII
International Congress of Neurology, Vol. II, p. 234, Rome.
KRIEG, W. J. S. (1954) "Connections of the Frontal Cortex of the Monkey." Spring-
field.
(1963) "Connections of the Cerebral Cortex." Evanston.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom
644 NORMAN GESCHWIND
LANGE, J. (1936) In Bumke, O., and Foerster, O. "Handbuch der Neurologic"
Berlin. Vol. 6, p. 807.
LIEPMANN, H. (1900) "Das Krankheitsbild der Apraxie ('motorischen Asymbolie')."
Berlin.
(1905a) "OberStorungen des Handelns bei Gehirnkranken." Berlin.
(19056) Munch, med. Wschr., 2, 2322 and 2375; also reprinted in Liepmann,
H. (1908) "Drei Aufsatze aus dem Apraxiegebiet." Berlin.
(1906) "Der weitere Krankheitsverlauf bei dem einseitig Apraktischen und
der Gehirnbefund auf Grund von Serienschnitten." Berlin.
, and MAAS, O. (1907) /. Psychol. NeuroL, Lpz., 10, 214.
LISSAUER, H. (1889) Arch. Psychiat. Nervenkr., 21, 222.
MAAS, O. (1907) Neurol. Zbl, 26, 789.
MCCULLOCH, W. S., and GAROL, H. W. (1941) /. Neurophysiol, 4, 555.
MACRAE, D., and TROLLE, E. (1956) Brain, 79, 94.
MASPES, P. E. (1948) Rev. neurol., 80, 100.
MEYER, A. (1895-1896) Amer. J. Insan., 52, 118. Reprinted in "The Collected Papers
of Adolf Meyer." Vol. I, 1950, Baltimore, p. 641.
(1905) Psychol. Bull., 2, 261. Reprinted in "The Collected Papers of Adolf
Meyer," Vol. I, 1950, Baltimore, p. 334.
MULLAN, S., and PENFIELD, W. (1959) Arch. Neurol. Psychiat., Chicago, 81, 269.
NIELSEN, J. M. (1937) Arch. Neurol. Psychiat., Chicago, 38, 108.
(1946) "Agnosia, Apraxia, Aphasia." 2nd edition, p. 186. New York and
London.
PENHELX), W., and BOLDREY, E. (1939) Amer. J. Psychiat., 96, 255.
PETR, R., HOLDEN, L. B., and JIROUT, J. (1949) /. Neuropath, exp. Neurol, 8, 100.
PICK, A. (1931) In Bethe, A., and v. Bergmann, G. "Handbuch der Normalen und
Pathologischen Physiologic" Vol. 15 (ii), Berlin, p. 1416.
POTZL, O., and STENGEL, E. (1936) Jb. Psychiat. Neurol, 53, 174.
RUSSELL, W. R., and ESPIR, M. L. E. (1961) "Traumatic Aphasia." London.
SHANKWEILER, D. (1964) Paper presented at Eastern Psychological Association,
Philadelphia.
SCHARLOCK, D. P., TUCKER, T. J., and STROMINGER, N. L. (1963) Science, 141, 1197.
SEOARRA, J. M., and QUADFASEL, F. A. (1961) Proceedings of the VII International
Congress of Neurology, Vol. II, p. 377.
STENGEL, E. (1947)/. ment. Set., 93, 598.
SUGAR, O., FRENCH, J. D., and CHUSID, J. G. (1948) /. Neurophysiol, 11, 175.
TEUBER, H. L., BATTERSBY, W. S., and BENDER, M. B. (1960) "Visual Field Defects
after Penetrating Missile Wounds of the Brain." Cambridge, Mass.
TRESCHER, J. H., and FORD, F. R. (1937) Arch. Neurol Psychiat., Chicago, 37, 959.
WARD, A. A., Jr., PEDEN, J. K., and SUGAR, O. (1946) /. Neurophysiol, 9, 453.
WARRTNGTON, E. K., (1962) /. Neurol Neurosurg. Psychiat., 25, 208.
WEINSTEIN, E. A., COLE, M., and MITCHELL, M. S. (1963) Trans. Amer. neurol Ass.,
88, 172.
WEISKRANTZ, L., and MISHKIN, M. (1958) Brain, 81, 406.
WELCH, K., and STUTEVILLE, P. (1958) Brain, 81, 341.
WERNICKE, C. (1874) "Der Aphasische Symptomencomplex." Breslau.
(1908) In "Diseases of the Nervous System" (Modern Clinical Medicine).
Edited by A. Church, New York and London.
WHITLOCK, D. G., and NAUTA, W. J. H. (1956) /. comp. Neurol, 106, 183.
WHORF, B. L. (1956) "Language, Thought, and Reality." New York.
WILBRAND, H. (1887) "Die Seelenblindheit als Herderscheinung." Wiesbaden.
ZANGWILL, O. (1960) "Cerebral Dominance and its Relation to Psychological
Function." Edinburgh.
byguestonOctober5,2011brain.oxfordjournals.orgDownloadedfrom