- Email: [email protected]
An inside-out theory of attention
/
Medical Hypotheses
Ildd &mthu (1992) 29.295301 0LaqlmnomupuKLtd199.2
An Inside-Out Theory of Attention B. ANDERSON Department
of Neurology,
University of Alabama
at Birmingham,
Birmingham,
AL 352944007.
USA
Abstract-The brain has two attentional systems; one attentional system is devoted to intrapersonal space and the second attentional system to extrapersonal space. The development of these two attentional systems with inward or outward focus arose from the hemispheral neural systems designed for attending to extrapersonal contralateral space. As a necessary consequence of the brain changes to develop language, one of the hemispheral attentional systems was altered for intrapersonal attention while the other hemispheral attentional system became obligated to attend to all of extrapersonal space. This conceptualization of attention provides a coherent explanation for the confusing hemispheral dominance proposed for attention on the basis of clinical hemineglect, i.e. that the left hemisphere attends to right extrapersonal space and that the right hemisphere attends to both left and right extrapersonal space. This conceptualization also provides a meaningful account of denial of limbs (anosognosia) and imagery.
Introduction Left sided neglect is a fascinating clinical phenomenon (1). Patients with this syndrome may deny their illness, deny their left arm and leg, and refuse to acknowledge left sided sensory perceptions (Fig. 1). Current investigations on clinical left hemineglect have focused on the mechanisms of neglect, e.g. directional hypokinesia (2) or sensory inattention (l), or the laterabty of the brain attentional networks. Unexplained by these approaches are such questions as how can a person deny the existence of his own limb, why is left hemineglect seen predominantly, and what can neglect of the world tell us about our own selfawareness and consciousness. The most widely accepted localization for the neural systems subserving attention is that the right hemisphere is responsible for attending to both left and Date received 3 June 1992 Date accepted 10 July 1992
right hemispaces and that the left hemisphere attends predominantly to the right hemispace (Fig. 2)(3). Thus, right hemineglect is infrequently seen because the remaining normal right hemisphere, in a case of left hemisphere injury, has sufficient reserve to attend to both hemispaces and left hemineglect occurs frequently in right hemisphere injury because the residual left hemisphere can only attend to right hemispace. This computational model explains the existing data derived from clinical neglect cases but it is an uncomfortable accommodation since no other brain system has been shown to have such a hybrid pattern of cerebral dominance. The present article examines how such a system might have come into existence and extends the discussion of attention to include imagery and the distinction between extrapersonal space and intrapersonal space.
295
2%
MEDICAL HYPOTHESES
Fig. 1 These are three drawings
by a patient with a right parietal lesion and left extrapersonal neglect. The patient was asked to place the numbers of a clock on a circle drawn by the examiner. Next she was asked to draw a Rower. Lastly, she was asked to copy a cross which had been drawn by the examiner. In all three instances the patient either neglected to place left sided items on the left or failed to include left sided umstructs in her drawings.
Extrapersonal Space Right
(Left Hemisphere)
Fig. 2 lhis
is the accepted model for the lateralization of the brain’s neural networks subserving attention. The right hemisphere can direct attention to either right or left extrapersonal space. The left hemisphere can only direct attention to the right extrapersonal hemispace.
The anomaly of the brain organbation for attention
Given that hemineglect occurs frequently after right hemisphere injury but infrequently after left hemisphere injury the conception of a right hemisphere organized to attend to space bilaterlly and a left hemisphere organized to attend only to contralateral space is logical. However, such a conception is contrary to
our understanding of the organization of other brain
systems. The brain appears to follow two modes of organiza[ion, either responsibility for a function is equally parcelled out to the two hemispheres or one hemisphere is dominant for a function. No prior example exists of one hemisphere being dominant for a function and the other hemisphere retaining its role in contralateral
297
AN INSIDE-OUT THEORY OF AlTElWlON
space, as has been proposed for the brain’s attentional system. An example of shared dominance is motor function. Each hemisphere of the brain controls independently the motor performance of the contralateral body, such that the left hemisphere is responsible for looking to the right with the eyes and head and moving the right arms and legs (4). Some functions may also be ipsilateral, for example the cerebellar contribution to coordinated motor action is considered to be ipsilateral (4). A similar organization underlies the brain’s organization of its sensory systems with each half of the body supplying the contralatera1 hemisphere. While there may remain some small number of fibers which contribute to a bilateral innervation, for example there are some 10% of motor fibers which descend ipsilaterally (5), such ipsilateral projections appear to have a negligible contribution on behavior and the symmetry of this arrangement is maintained for both hemispheres. Language is the prototype of a behavior for which one hemisphere is dominant (6). There is consensus that for most people the left hemisphere is dominant for most aspects of language. There is a small, but important contribution, of the right hemisphere to language. The right hemisphere has been suggested as a dominant hemisphere for the production and comprehension of the affective (emotional), non-semantic aspects of language that contribute to the understanding of an utterance’s meaning (7). Even in providing an exception to the left hemisphere’s dominance of language this example provides further support for the concept of an allocation of functions to one hemisphere or the other, it is the right hemisphere that remains dominant for affective prosody-a component of language-not that language and all its components are divided equally between the two hemispheres. These examples of the brain’s modes of parcelling out functional localization serve to illustrate how atypical is the proposed organization of the brain’s attentional system. Applied to the motor system such an organization would yield people who get weak on the left side after a right hemisphere stroke but do nol show weakness after a left hemisphere stroke or yield stroke victims who can speak out of the left sides of their mouths and write with their left hands after a left hemisphere stroke but show normal language after a right hemisphere stroke. Accepting such a unique organizational system for the brain’s attentional system requires some intellectual underpinning to establish the cogency and authenticity of this unique occurrence.
An evolutionary perspective of attention
To understand the present organization of attentional systems in humans it is useful to consider how attention is organized in the non-human primate brain. The organization of function in the non-human primate brain provides a suggestion of how attentional systems were organized in early homo and immediate progenitor species. From this starting point we can attempt to deduce the factors which prompted development of the current situation. Monkeys and people have functionally different organizations of their attentional systems. Monkeys show a consistent symmetry for lesions which pmduce neglect. Monkeys fail to show a consistent dominance, within a species, for handedness, and do not show the ability to use language in any important form. The contemporaneous development in humans, and possibly other early homo species, of hemispheric dominance for language, praxis, and attention was not coincidental but reflects an interdependence. At a certain stage of animal development the brain was required to engage c0nt.ralatera.l space. Thus, animals, such as monkeys, who existed where the left side of their brains controlled right limbs, directed the head and eyes to regard the space where the right limbs operated, and detected sensation applied to the right limbs, required a neural system for focusing attention within that contralateral space. Consistency suggests that such attentional networks should be located so that the left hemisphere would possess the networks for attending to right hemispace and vice versa. It seems probable though that just as a small number of motor fibers descend ipsilaterally some of the attentional fields of a hemisphere are devoted to ipsilateral attention, although probably, again in analogy to motor and sensory pathways, such a small number that their contribution to overt behavior is not detected. This is the situation found in monkeys studied presently. To permit the development of complex language there is a need to look inward. There is a need to regard internal visual, auditory, and tactile images. To conduct an internal dialogue there is a requirement for both a speaker and listener. Such a development necessitates the ability to attend inward in addition to outward. As a concrete consequence of this requirement one of the hemisphere attentional systems began to specialize in inward attention and the other hemisphere assumed the role of outward attention. Which hemisphere did which was probably arbitrary, since cases of anomalous dominance demonstrate that either hemisphere can subserve either task
298 adequately, but that each hemisphere became dominant for a specialized aspect of attention is manifest. The right hemisphere expanded its purview to include ipsilateral right hemispace though sustaining the initial and probably more strongly linked connection to the left hemispace. The left hemisphere attentional system underwent atrophy of the right hemispace connection, and even more attenuated the nascent left hemispace connection, while assuming a direct role in attention to inward images. Inward attention The objects of an extrapersonal attentional system are the items we directly experience with the five senses. To look at something, regard it, study it, and react to it one needs an extrapersonal attentional system. The objects of an inward attentional system are those images we see (hear, feel, etc) in the mind’s eye. Inward objects also include the system of internally manipulated symbols that relate characteristics of and connections between the events and the items that may bc either internal images or external objects. That system of symbols is language. That we are able to attend inwardly is substantiated by introspection and experience. All people have dreamed. All have conducted an internal dialogue. These are the inward experiences to which we attend and of which we are aware. The nature of these inward objects is not entirely clear. Most consider visual images to be internal representations of visual material in a manner that approximates and replicates the handling of external visual percepts (8) although it is not at the level of primary visual cortex, preserved images in those rendered cortically blind eliminates this localization (9). It has been argued though, that our impressions may deceive us and that all images may actually be propositional information (10). While a little difficult to accept at first since we tend to believe what we see with our own mind’s eye such a formulation is attractive given the development of an inward attentional system at the same time as language skills developed, a simultaneous development that is argued to have been interdependent. Regardless of the precise nature of inner percepts it is likely that they do not follow the same spatial coordinate system as external percepts but tend to be appreciated more holistically. Evidence comes from the few cases of a relatively isolated loss of visual imagery, in none was a lateral neglect of imaged features reported (11, 12). However, the few cases of this nature and the limited systematic investigations of acute stroke patients for lateralized neglect of image features in the presence of preserved
MEDICAL HYRYTHESES
attention to external percepts makes this contention provisional. The inputs to the outward attentional system are the sensory perceptions. The inputs to the inward attentional system are likely stored sensory templates and the system for their symbolic manipulation. Case reports of fractional disturbances of memory after unilateral temporal lobe injury suggest that information may be differentially stored in the two hemispheres depending on its characteristics, such as semantic or visual-spatial, implying that the inward attentional system would require inputs from both temporal lobes (13, 14). Also, cases of left hemineglect which have included neglect of the left sided features of visual images imply that at some level visual information is spatially coded right versus left (15). Again, this would suggest bihemispheric input to an inward attentional system. Deficiencies of inward attention could then derive from either damage to the inward attentional system itself or to afferent connections disconnecting the inward attentional system. Predictions and evaluation of the inside-out attentional model on extrapersonal space Predictions for damage to the extrapersonal, right hemisphere, component of the model are identical to those for the initial computational model which prompted this review. Given the bispacial fields of the right hemisphere attentional system left hemisphere strokes would yield either no neglect or minor neglect in the contralateral hemispace. The latter might occur because the left hemisphere-right hemispace exlrapersonal attentional link is more important for some people. This is consistent with the literature of left hemisphere stroke patients which shows that neglect is less common, less severe, and affects the right hemispace (16, 17). Damage to the right hemisphere would produce neglect in the left hemispace primarily, but often also less severe neglect in the ipsilateral right hemispace (3, 18, 19). The latter would occur because the left hemisphere-right hemispace extrapersonal attentional system is not primarily designed for this task and so does not perform at an optima1 level. Predictions and evaluation of the inside-out attentional model on intrapersonal space Damage to the left hemisphere that affected the inward attentional system primarily would be predicted to yield a degradation in the ability to attend to images and use those images in conscious responses. Complete loss of the ability to attend to inward images would occur only in left hemisphere
AN lNSIDE_Om
299
THEORY OF ATTENTION
lesions.1 Given the interdependent development of this inward attentional system with language it is probable that similar brain regions are important in the comprehension of language and inward attention such that most patients with loss of inward attention will have severe comprehension problems. The severity of the inward attentional disorder will parallel the severity of the comprehension disturbance of language. Such an association has been found in the literature on constructional apraxia after stroke (20). While commonly considered a right hemisphere sign, the occurrence of constructional problems in patients with left hemisphere stroke may bc as frequent as constructional problems in those with right hemisphere stroke (21). Where this variable has been specifically analyzed the severity of the constructional deficit paralleled the severity of the comprehension deficit in the aphasia (20). This suggests that there may be two ways in which a constructional problem may emerge in patients with hemisphere strokes. One way is by interfering with the ability to spatially analyze a perception, right hemisphere, and the second way would be through an interference in the accession of internal visual images, left hemisphere. The emphasis of this prediction is on language, not speech or writing, so that patients with auditory agnosias as the cause of a comprehension defect would not be predicted to have coincident inattention to inward visual images. In further support of this prediction, most cases of a loss of visual imagery, in large part reviewed by Farah (11). have suggested a left hemisphere localization for the lesion. These cases have been interpreted from different theoretical perspectives with emphasis being given either to image generation, disconnection of image inspection and verbal reporting (12). or, as presently, image attention. Cases of inattention of internal images will be rarely reported because of the comprehension language disturbance that makes patient testing problematic. Ideal nonverbal tests of attention to inner images are hard to conceive but would require at least that instructions be able to be given by demonstration and example without verbal input, that correct performance require a simple motor response so that coexistent left-right confusion could not artificially influence performance, the patient would need to demonstrate normal figure copying to imply intact perception of test items, and correct performance would require making use of information that had to be retained as an image, ideally an item that was memorized from before the brain injury so that simple
mnemonic defects would not conflate with attention to internal image abnormalities. A dependence on non-verbal responses makes testing inward attention to images problematic in another respect. The inability to attend to internal images is not identical to the inability to generate images or have such images influence behavior. The neuropsychological literature is becoming rife with reports of covert awareness of stimuli in the presence of overt denial of perception (22). This pertains to extrapersonal attention where neglected targets can influence detection of non-neglected targets (23) and electrodermal responses can be demonstrated to sensory stimuli whose presence is denied (24). Therefore, it is possible to imagine, in a hypothetical situation, that a person could deny the existence of internal images yet be covertly shown to respond when performing tasks the correct solution of which is thought to involve using internal images. A patient could deny the ability to conjure up the images necessary to solve such problems but do much better than chance in a forced choice paradigm showing that such information was at some level accessible. Right hemisphere injuries will also produce affects on attention to internal images. Incomplete loss of attention to inward images will occur if afferent pathways to the inward attentional system are compromised. Such abnormalities will be different depending on the afferent system injured. If the right hemisphere structures important for visual-spatial information or their connections to the left hemisphere intrapersonal space attentional system are damaged it might not be possible to attend to a whole image since the input necessary to construct an integrated image would be impaired but the individual could attend to image parts or features, as has been seen in one patient (cited in (8)). Damage to right hemisphere that interfered with afferents communicating left sided spatial information could yield patients with the inability to report left sided features of imagined scenes (15). Both varieties of clinical deficiencies have been reported. Doubly disassociating inward attention from
outward attention Bisiach and others have reported patients with left sided extrapersonal neglect who have shown a tendency to neglect left sided features of internal or imaged information thereby suggesting a possible functional link between attention to internal and exuapersonal space (15, 25). However, subsequent patients
1. Since l-Z% of all right hemisphere stroke patients show crossed aphasia it is likely that a similar percentage of inward attentional loss would also occur ‘crossed’.
300 have hem discovered that show preserved attention to
intrapersonal space with left hemineglect disassociating extrapersonal neglect from intrapersonal neglect (26). The reported cases of loss of visual imagery reported to date who showed preserved extrapersonal attention complete the double disassociation (11, 12). Another piece of evidence for disassociating the two attentional systems is that people are able to attend to two things at once. For example driving a car on a familiar route may require several actions that are not completely rehearsed such as when to stop, which lane to drive in, determine whether a light is green or red, but it can often be done without conscious recollection while a person is thinking (inwardly attending) to something else. The mechanism of denial Some right hemisphere stroke patients show a denial of illness or a denial of their extremities in addition to the inattention and defective intention. Recognition of loss is postulated to occur when there is a mismatch in internal and external perceptions. We correctly acknowledge our left limbs because we have both an external perception of them and an internal representation of them. Patients with right hemisphere strokes but no neglect will acknowledge a loss of sensation in the left arm because their lack of sensation fails to match their internal representation. However a patient who loses internal attention to the features of his left side by disconnecting the right temporal-parietal lobe from the left hemisphere attentional system and who also loses awareness of the left side through extrapersonal inattention will have no discrepancy. There would be no reason for such a person to even conceive of a left side since he has no external nor internal perception of such. The anosognosia analogue for left hemisphere injury will have to be much restricted since such patients have intact perceptions of their extrapersonal space. In fact the only sphere to which such a syndrome might apply is speech since its external manifestations are not sensical to the right hemisphere, the right hemisphere being only aware of melody and prosody of the speech output. The prediction is that patients with comprehension language disturbances and an internal inattention would not be aware of their language impairment since there would be no mismatch between the internal percepts of language and the external percepts of their verbal output. While one cannot ask a sensory aphasic whether he knows he has a problem their behavior often implies such (27). The classic Wemicke’s aphasic with his word salad jargon carries on with an insouciance that is discomforting.
MEDICAL HYPOTHESES
Note how this behavior contrasts with that of someone made suddenly deaf by peripheral ear damage. Even if the deaf subject fails to recognise his deafness initially he recognises that he cannot understand what is being said to him and acts in his manners and posture appropriate. The Wemicke’s aphasia patient by his behavior does present the predicted anosognosia. Recovery
A founding basis of this model has been that both attentional systems developed from hemispheric neural systems designed to attend to external space and with some degree of connections to both ipsilateral and contralateral space, the latter being dominant. It would follow then that patients with damage to one attentional system could recover by co-opting the other attentional system to perform double duty, The presence of systems with dual capability could explain the relatively good recovery in most cases of left hemineglect (1). In the case of right hemisphere injury the left hemisphere attentional system will be required to subsume its original function of attending to extrapersonal space, the strength of its connections favoring right hemispace and its amount of ipsilateral connections determining the severity of longterm left sided neglect. Such a proposal does not preclude intrahemispheric mechanisms from also contributing to behavioral recovery. In left hemisphere injury the opposite situation will obtain. It follows that if one attentional system is having to subsume two functions the ability to simultaneously perform two behaviors, one requiring extrapersonal attention and the other intrapersonal attention, will be degraded. Predictions
1. The loss of the ability to attend to internal space
is a dominant function of the left hemisphere. 2. Patients with left hemisphere lesions will show
neglect of internal images proportionate to their comprehension language disturbances. 3. Asomatognosia for the right limbs can only occur in people with anomalous dominance for extrapersonal and intrapersonal attention. 4. Asomatognosia can only occur in patients with both severe extrapersonal neglect and interruption of pathways from right hemisphere to left hemisphere yielding left hemineglect for internal images. 5. Patients with left hemineglect from right hemisphere strokes will only show neglect of left sided features of internal images if there is an
301
AN INSIDE-OUT THEORY OF ATTENTION
interruption of pathways from right hemisphere to the left hemisphere attentional system. 6. Patients who have recovered from left hemineglect should have a worsening of their exuapersonal neglect with subsequent posterior left hemisphere dysfunction (e.g. WADA test or stroke). 7. Testing of patients recovered from either intrapersonal or exuapersonal neglect should show a deterioration on their ability to perform tasks requiring simultaneously extrapersonal and intrapersonal attention. An example of a task meeting these criteria would be to have a person answer a series of imagery questions while at the same time performing a well rehearsed detection task like noting the appearance of a certain letter flashed on a computer screen at random with a number of like sized letter foils.
11. 12. 13.
14.
15.
16. 17.
18.
19.
References 1. Heiiman K M. Watson R T, Valenstein E. Neglect and related disorden. ~243 in Clinical Nemopsychology. 2nd ed. (K M Heilman, E Valenstein. eds) Oxford, New York, 1985. 2. Watson R T, Valenstein E, Heilman K M. Normal tactile threshold in monkeys with neglect. Neurology 36: 636. 1986. 3. Weintraub S, Mesulam M-M. Right cerebral dominance in spatial attention. Archives of Neurology 44: 621, 1987. 4. Adams R D. Victor M. Principles of Neurology. 3rd ed. McGraw-Hill. New York. 1985. 5. Tmex R C. Carpenter M B. Strong and Elwyn’s Human Neuroanatomy. 5tb ed. Williams and Wilkins, Baltimore, 1964. 6. Kolb B. Whishaw 1 Q. Fundamentals of Human Neuropsychology. 2nd ed. W H Freeman. New York, 1985. 7. Ross E D. Tbe aprosodias. Archives of Neurology 38: 561, 1981. 8. Kosslyn S M. Aspects of a cognitive neuroscience of mental imagery. Science 240: I62 1, 1988. 9. Sacks 0. Neurological dreams. Medical Doctor 35: 29, 1992. 10. Pylyshyn Z W. What tlte mind’s eye tells the mind’s brain: a critique of mental imagery. Psychological Bulletin 80: 1.1973.
20.
21.
22.
23.
24.
25. 26.
27.
Farah M J. The neurological basis of mental imagery: a componerttial analysis. Cognition 18: 245, 1984. Basso A, Bisiach E, Luzatti C. Loss of mental imagery: a case study. Neuropsychologia 18: 435, 1980. Ross E D. Sensory-specific and fractional disorders of recent memory in man: 1. Isolated loss of visual recent memory. Archives of Neurology 37: 193, 1980. Ross E D. Sensory-specific and fractional disorders of mcent memory in man: II. Unilateral loss of tactile recent memory. Archives of Neurology 37: 267, 1980. Bisiach E, Capitani E, Luzatti C, Perani D. Brain and conscious representation of outside reality. Neuropsychologis 19: 543, 1981. Albert M C. A simple test of visual neglect. Neurology 23: 658, 1973. Gainotti G, Messerli P, Tissot R. Qualitative analysis of tmilsteral spatial neglect in relation to lsterality of cerebral lesicms. Journal of Neurology, Neurosurgery and Psychiatry 35: 545. 1972. Weintrsub S, Ahem G L, Dsffner K R, Price B H, Mesulam M-M. Right sided hemispstial neglect. Neurology (suppl) 42: 223, 1992. Feinberg T E. Haber L D. Stacy C B. lpsilstersl extinction in the hemineglect syndrome. Archives of Neurology 47: 802. 1990. Benton A L. Visuoconstntctive disability in patients with cerebral disease: its relationship to side of lesion and aphasic disorder. Documents Gphthamologica 34: 67, 1973. Benton A. Visuoperceptual. visuospatisl, and visuoconstmctive disorders. ~151 in Clinical Neuropsychology 2nd ed. (K M Heilman, E Valenstein, eds) Oxford, New York, 1985. Cowey A, Stoerig P. Reflections on blindsight pll in the Neuropsychology of Consciousness. (A D Milner, M D Rugg. eds) Academic Press, London, 1992. Chatterjee A, Mennemeier M. Heilmsn K M. Stimulus-response relationship in unilateral neglect. Journal of Clinical and Experimental Neuropsychology 14: 39, 1992. Vallar G, Bottini G. Sterzi R. Pssserini D. Ruscari M L. Hemianesthesia. sensory neglect, and defective access to conscious experience. Neurology 41: 650. 1991. Meador K J. Loring D W. Bowers D. Heilman K M. Remote memory and neglect syndrome. Neurology 37: 522, 1987. Anderson B. Spared awareness for the left side of internal visual images in patients with left sided extrapersonal neglect. Neurology (in press). Benson D F, Geschwind N. Aphasia and related disorders: s clinical approach. ~193 in Principles of Behavioral Neurology. (M-M Mesulam. ed) F A Davis, Philadelphia, 1985.
Medical Hypotheses
Ildd &mthu (1992) 29.295301 0LaqlmnomupuKLtd199.2
An Inside-Out Theory of Attention B. ANDERSON Department
of Neurology,
University of Alabama
at Birmingham,
Birmingham,
AL 352944007.
USA
Abstract-The brain has two attentional systems; one attentional system is devoted to intrapersonal space and the second attentional system to extrapersonal space. The development of these two attentional systems with inward or outward focus arose from the hemispheral neural systems designed for attending to extrapersonal contralateral space. As a necessary consequence of the brain changes to develop language, one of the hemispheral attentional systems was altered for intrapersonal attention while the other hemispheral attentional system became obligated to attend to all of extrapersonal space. This conceptualization of attention provides a coherent explanation for the confusing hemispheral dominance proposed for attention on the basis of clinical hemineglect, i.e. that the left hemisphere attends to right extrapersonal space and that the right hemisphere attends to both left and right extrapersonal space. This conceptualization also provides a meaningful account of denial of limbs (anosognosia) and imagery.
Introduction Left sided neglect is a fascinating clinical phenomenon (1). Patients with this syndrome may deny their illness, deny their left arm and leg, and refuse to acknowledge left sided sensory perceptions (Fig. 1). Current investigations on clinical left hemineglect have focused on the mechanisms of neglect, e.g. directional hypokinesia (2) or sensory inattention (l), or the laterabty of the brain attentional networks. Unexplained by these approaches are such questions as how can a person deny the existence of his own limb, why is left hemineglect seen predominantly, and what can neglect of the world tell us about our own selfawareness and consciousness. The most widely accepted localization for the neural systems subserving attention is that the right hemisphere is responsible for attending to both left and Date received 3 June 1992 Date accepted 10 July 1992
right hemispaces and that the left hemisphere attends predominantly to the right hemispace (Fig. 2)(3). Thus, right hemineglect is infrequently seen because the remaining normal right hemisphere, in a case of left hemisphere injury, has sufficient reserve to attend to both hemispaces and left hemineglect occurs frequently in right hemisphere injury because the residual left hemisphere can only attend to right hemispace. This computational model explains the existing data derived from clinical neglect cases but it is an uncomfortable accommodation since no other brain system has been shown to have such a hybrid pattern of cerebral dominance. The present article examines how such a system might have come into existence and extends the discussion of attention to include imagery and the distinction between extrapersonal space and intrapersonal space.
295
2%
MEDICAL HYPOTHESES
Fig. 1 These are three drawings
by a patient with a right parietal lesion and left extrapersonal neglect. The patient was asked to place the numbers of a clock on a circle drawn by the examiner. Next she was asked to draw a Rower. Lastly, she was asked to copy a cross which had been drawn by the examiner. In all three instances the patient either neglected to place left sided items on the left or failed to include left sided umstructs in her drawings.
Extrapersonal Space Right
(Left Hemisphere)
Fig. 2 lhis
is the accepted model for the lateralization of the brain’s neural networks subserving attention. The right hemisphere can direct attention to either right or left extrapersonal space. The left hemisphere can only direct attention to the right extrapersonal hemispace.
The anomaly of the brain organbation for attention
Given that hemineglect occurs frequently after right hemisphere injury but infrequently after left hemisphere injury the conception of a right hemisphere organized to attend to space bilaterlly and a left hemisphere organized to attend only to contralateral space is logical. However, such a conception is contrary to
our understanding of the organization of other brain
systems. The brain appears to follow two modes of organiza[ion, either responsibility for a function is equally parcelled out to the two hemispheres or one hemisphere is dominant for a function. No prior example exists of one hemisphere being dominant for a function and the other hemisphere retaining its role in contralateral
297
AN INSIDE-OUT THEORY OF AlTElWlON
space, as has been proposed for the brain’s attentional system. An example of shared dominance is motor function. Each hemisphere of the brain controls independently the motor performance of the contralateral body, such that the left hemisphere is responsible for looking to the right with the eyes and head and moving the right arms and legs (4). Some functions may also be ipsilateral, for example the cerebellar contribution to coordinated motor action is considered to be ipsilateral (4). A similar organization underlies the brain’s organization of its sensory systems with each half of the body supplying the contralatera1 hemisphere. While there may remain some small number of fibers which contribute to a bilateral innervation, for example there are some 10% of motor fibers which descend ipsilaterally (5), such ipsilateral projections appear to have a negligible contribution on behavior and the symmetry of this arrangement is maintained for both hemispheres. Language is the prototype of a behavior for which one hemisphere is dominant (6). There is consensus that for most people the left hemisphere is dominant for most aspects of language. There is a small, but important contribution, of the right hemisphere to language. The right hemisphere has been suggested as a dominant hemisphere for the production and comprehension of the affective (emotional), non-semantic aspects of language that contribute to the understanding of an utterance’s meaning (7). Even in providing an exception to the left hemisphere’s dominance of language this example provides further support for the concept of an allocation of functions to one hemisphere or the other, it is the right hemisphere that remains dominant for affective prosody-a component of language-not that language and all its components are divided equally between the two hemispheres. These examples of the brain’s modes of parcelling out functional localization serve to illustrate how atypical is the proposed organization of the brain’s attentional system. Applied to the motor system such an organization would yield people who get weak on the left side after a right hemisphere stroke but do nol show weakness after a left hemisphere stroke or yield stroke victims who can speak out of the left sides of their mouths and write with their left hands after a left hemisphere stroke but show normal language after a right hemisphere stroke. Accepting such a unique organizational system for the brain’s attentional system requires some intellectual underpinning to establish the cogency and authenticity of this unique occurrence.
An evolutionary perspective of attention
To understand the present organization of attentional systems in humans it is useful to consider how attention is organized in the non-human primate brain. The organization of function in the non-human primate brain provides a suggestion of how attentional systems were organized in early homo and immediate progenitor species. From this starting point we can attempt to deduce the factors which prompted development of the current situation. Monkeys and people have functionally different organizations of their attentional systems. Monkeys show a consistent symmetry for lesions which pmduce neglect. Monkeys fail to show a consistent dominance, within a species, for handedness, and do not show the ability to use language in any important form. The contemporaneous development in humans, and possibly other early homo species, of hemispheric dominance for language, praxis, and attention was not coincidental but reflects an interdependence. At a certain stage of animal development the brain was required to engage c0nt.ralatera.l space. Thus, animals, such as monkeys, who existed where the left side of their brains controlled right limbs, directed the head and eyes to regard the space where the right limbs operated, and detected sensation applied to the right limbs, required a neural system for focusing attention within that contralateral space. Consistency suggests that such attentional networks should be located so that the left hemisphere would possess the networks for attending to right hemispace and vice versa. It seems probable though that just as a small number of motor fibers descend ipsilaterally some of the attentional fields of a hemisphere are devoted to ipsilateral attention, although probably, again in analogy to motor and sensory pathways, such a small number that their contribution to overt behavior is not detected. This is the situation found in monkeys studied presently. To permit the development of complex language there is a need to look inward. There is a need to regard internal visual, auditory, and tactile images. To conduct an internal dialogue there is a requirement for both a speaker and listener. Such a development necessitates the ability to attend inward in addition to outward. As a concrete consequence of this requirement one of the hemisphere attentional systems began to specialize in inward attention and the other hemisphere assumed the role of outward attention. Which hemisphere did which was probably arbitrary, since cases of anomalous dominance demonstrate that either hemisphere can subserve either task
298 adequately, but that each hemisphere became dominant for a specialized aspect of attention is manifest. The right hemisphere expanded its purview to include ipsilateral right hemispace though sustaining the initial and probably more strongly linked connection to the left hemispace. The left hemisphere attentional system underwent atrophy of the right hemispace connection, and even more attenuated the nascent left hemispace connection, while assuming a direct role in attention to inward images. Inward attention The objects of an extrapersonal attentional system are the items we directly experience with the five senses. To look at something, regard it, study it, and react to it one needs an extrapersonal attentional system. The objects of an inward attentional system are those images we see (hear, feel, etc) in the mind’s eye. Inward objects also include the system of internally manipulated symbols that relate characteristics of and connections between the events and the items that may bc either internal images or external objects. That system of symbols is language. That we are able to attend inwardly is substantiated by introspection and experience. All people have dreamed. All have conducted an internal dialogue. These are the inward experiences to which we attend and of which we are aware. The nature of these inward objects is not entirely clear. Most consider visual images to be internal representations of visual material in a manner that approximates and replicates the handling of external visual percepts (8) although it is not at the level of primary visual cortex, preserved images in those rendered cortically blind eliminates this localization (9). It has been argued though, that our impressions may deceive us and that all images may actually be propositional information (10). While a little difficult to accept at first since we tend to believe what we see with our own mind’s eye such a formulation is attractive given the development of an inward attentional system at the same time as language skills developed, a simultaneous development that is argued to have been interdependent. Regardless of the precise nature of inner percepts it is likely that they do not follow the same spatial coordinate system as external percepts but tend to be appreciated more holistically. Evidence comes from the few cases of a relatively isolated loss of visual imagery, in none was a lateral neglect of imaged features reported (11, 12). However, the few cases of this nature and the limited systematic investigations of acute stroke patients for lateralized neglect of image features in the presence of preserved
MEDICAL HYRYTHESES
attention to external percepts makes this contention provisional. The inputs to the outward attentional system are the sensory perceptions. The inputs to the inward attentional system are likely stored sensory templates and the system for their symbolic manipulation. Case reports of fractional disturbances of memory after unilateral temporal lobe injury suggest that information may be differentially stored in the two hemispheres depending on its characteristics, such as semantic or visual-spatial, implying that the inward attentional system would require inputs from both temporal lobes (13, 14). Also, cases of left hemineglect which have included neglect of the left sided features of visual images imply that at some level visual information is spatially coded right versus left (15). Again, this would suggest bihemispheric input to an inward attentional system. Deficiencies of inward attention could then derive from either damage to the inward attentional system itself or to afferent connections disconnecting the inward attentional system. Predictions and evaluation of the inside-out attentional model on extrapersonal space Predictions for damage to the extrapersonal, right hemisphere, component of the model are identical to those for the initial computational model which prompted this review. Given the bispacial fields of the right hemisphere attentional system left hemisphere strokes would yield either no neglect or minor neglect in the contralateral hemispace. The latter might occur because the left hemisphere-right hemispace exlrapersonal attentional link is more important for some people. This is consistent with the literature of left hemisphere stroke patients which shows that neglect is less common, less severe, and affects the right hemispace (16, 17). Damage to the right hemisphere would produce neglect in the left hemispace primarily, but often also less severe neglect in the ipsilateral right hemispace (3, 18, 19). The latter would occur because the left hemisphere-right hemispace extrapersonal attentional system is not primarily designed for this task and so does not perform at an optima1 level. Predictions and evaluation of the inside-out attentional model on intrapersonal space Damage to the left hemisphere that affected the inward attentional system primarily would be predicted to yield a degradation in the ability to attend to images and use those images in conscious responses. Complete loss of the ability to attend to inward images would occur only in left hemisphere
AN lNSIDE_Om
299
THEORY OF ATTENTION
lesions.1 Given the interdependent development of this inward attentional system with language it is probable that similar brain regions are important in the comprehension of language and inward attention such that most patients with loss of inward attention will have severe comprehension problems. The severity of the inward attentional disorder will parallel the severity of the comprehension disturbance of language. Such an association has been found in the literature on constructional apraxia after stroke (20). While commonly considered a right hemisphere sign, the occurrence of constructional problems in patients with left hemisphere stroke may bc as frequent as constructional problems in those with right hemisphere stroke (21). Where this variable has been specifically analyzed the severity of the constructional deficit paralleled the severity of the comprehension deficit in the aphasia (20). This suggests that there may be two ways in which a constructional problem may emerge in patients with hemisphere strokes. One way is by interfering with the ability to spatially analyze a perception, right hemisphere, and the second way would be through an interference in the accession of internal visual images, left hemisphere. The emphasis of this prediction is on language, not speech or writing, so that patients with auditory agnosias as the cause of a comprehension defect would not be predicted to have coincident inattention to inward visual images. In further support of this prediction, most cases of a loss of visual imagery, in large part reviewed by Farah (11). have suggested a left hemisphere localization for the lesion. These cases have been interpreted from different theoretical perspectives with emphasis being given either to image generation, disconnection of image inspection and verbal reporting (12). or, as presently, image attention. Cases of inattention of internal images will be rarely reported because of the comprehension language disturbance that makes patient testing problematic. Ideal nonverbal tests of attention to inner images are hard to conceive but would require at least that instructions be able to be given by demonstration and example without verbal input, that correct performance require a simple motor response so that coexistent left-right confusion could not artificially influence performance, the patient would need to demonstrate normal figure copying to imply intact perception of test items, and correct performance would require making use of information that had to be retained as an image, ideally an item that was memorized from before the brain injury so that simple
mnemonic defects would not conflate with attention to internal image abnormalities. A dependence on non-verbal responses makes testing inward attention to images problematic in another respect. The inability to attend to internal images is not identical to the inability to generate images or have such images influence behavior. The neuropsychological literature is becoming rife with reports of covert awareness of stimuli in the presence of overt denial of perception (22). This pertains to extrapersonal attention where neglected targets can influence detection of non-neglected targets (23) and electrodermal responses can be demonstrated to sensory stimuli whose presence is denied (24). Therefore, it is possible to imagine, in a hypothetical situation, that a person could deny the existence of internal images yet be covertly shown to respond when performing tasks the correct solution of which is thought to involve using internal images. A patient could deny the ability to conjure up the images necessary to solve such problems but do much better than chance in a forced choice paradigm showing that such information was at some level accessible. Right hemisphere injuries will also produce affects on attention to internal images. Incomplete loss of attention to inward images will occur if afferent pathways to the inward attentional system are compromised. Such abnormalities will be different depending on the afferent system injured. If the right hemisphere structures important for visual-spatial information or their connections to the left hemisphere intrapersonal space attentional system are damaged it might not be possible to attend to a whole image since the input necessary to construct an integrated image would be impaired but the individual could attend to image parts or features, as has been seen in one patient (cited in (8)). Damage to right hemisphere that interfered with afferents communicating left sided spatial information could yield patients with the inability to report left sided features of imagined scenes (15). Both varieties of clinical deficiencies have been reported. Doubly disassociating inward attention from
outward attention Bisiach and others have reported patients with left sided extrapersonal neglect who have shown a tendency to neglect left sided features of internal or imaged information thereby suggesting a possible functional link between attention to internal and exuapersonal space (15, 25). However, subsequent patients
1. Since l-Z% of all right hemisphere stroke patients show crossed aphasia it is likely that a similar percentage of inward attentional loss would also occur ‘crossed’.
300 have hem discovered that show preserved attention to
intrapersonal space with left hemineglect disassociating extrapersonal neglect from intrapersonal neglect (26). The reported cases of loss of visual imagery reported to date who showed preserved extrapersonal attention complete the double disassociation (11, 12). Another piece of evidence for disassociating the two attentional systems is that people are able to attend to two things at once. For example driving a car on a familiar route may require several actions that are not completely rehearsed such as when to stop, which lane to drive in, determine whether a light is green or red, but it can often be done without conscious recollection while a person is thinking (inwardly attending) to something else. The mechanism of denial Some right hemisphere stroke patients show a denial of illness or a denial of their extremities in addition to the inattention and defective intention. Recognition of loss is postulated to occur when there is a mismatch in internal and external perceptions. We correctly acknowledge our left limbs because we have both an external perception of them and an internal representation of them. Patients with right hemisphere strokes but no neglect will acknowledge a loss of sensation in the left arm because their lack of sensation fails to match their internal representation. However a patient who loses internal attention to the features of his left side by disconnecting the right temporal-parietal lobe from the left hemisphere attentional system and who also loses awareness of the left side through extrapersonal inattention will have no discrepancy. There would be no reason for such a person to even conceive of a left side since he has no external nor internal perception of such. The anosognosia analogue for left hemisphere injury will have to be much restricted since such patients have intact perceptions of their extrapersonal space. In fact the only sphere to which such a syndrome might apply is speech since its external manifestations are not sensical to the right hemisphere, the right hemisphere being only aware of melody and prosody of the speech output. The prediction is that patients with comprehension language disturbances and an internal inattention would not be aware of their language impairment since there would be no mismatch between the internal percepts of language and the external percepts of their verbal output. While one cannot ask a sensory aphasic whether he knows he has a problem their behavior often implies such (27). The classic Wemicke’s aphasic with his word salad jargon carries on with an insouciance that is discomforting.
MEDICAL HYPOTHESES
Note how this behavior contrasts with that of someone made suddenly deaf by peripheral ear damage. Even if the deaf subject fails to recognise his deafness initially he recognises that he cannot understand what is being said to him and acts in his manners and posture appropriate. The Wemicke’s aphasia patient by his behavior does present the predicted anosognosia. Recovery
A founding basis of this model has been that both attentional systems developed from hemispheric neural systems designed to attend to external space and with some degree of connections to both ipsilateral and contralateral space, the latter being dominant. It would follow then that patients with damage to one attentional system could recover by co-opting the other attentional system to perform double duty, The presence of systems with dual capability could explain the relatively good recovery in most cases of left hemineglect (1). In the case of right hemisphere injury the left hemisphere attentional system will be required to subsume its original function of attending to extrapersonal space, the strength of its connections favoring right hemispace and its amount of ipsilateral connections determining the severity of longterm left sided neglect. Such a proposal does not preclude intrahemispheric mechanisms from also contributing to behavioral recovery. In left hemisphere injury the opposite situation will obtain. It follows that if one attentional system is having to subsume two functions the ability to simultaneously perform two behaviors, one requiring extrapersonal attention and the other intrapersonal attention, will be degraded. Predictions
1. The loss of the ability to attend to internal space
is a dominant function of the left hemisphere. 2. Patients with left hemisphere lesions will show
neglect of internal images proportionate to their comprehension language disturbances. 3. Asomatognosia for the right limbs can only occur in people with anomalous dominance for extrapersonal and intrapersonal attention. 4. Asomatognosia can only occur in patients with both severe extrapersonal neglect and interruption of pathways from right hemisphere to left hemisphere yielding left hemineglect for internal images. 5. Patients with left hemineglect from right hemisphere strokes will only show neglect of left sided features of internal images if there is an
301
AN INSIDE-OUT THEORY OF ATTENTION
interruption of pathways from right hemisphere to the left hemisphere attentional system. 6. Patients who have recovered from left hemineglect should have a worsening of their exuapersonal neglect with subsequent posterior left hemisphere dysfunction (e.g. WADA test or stroke). 7. Testing of patients recovered from either intrapersonal or exuapersonal neglect should show a deterioration on their ability to perform tasks requiring simultaneously extrapersonal and intrapersonal attention. An example of a task meeting these criteria would be to have a person answer a series of imagery questions while at the same time performing a well rehearsed detection task like noting the appearance of a certain letter flashed on a computer screen at random with a number of like sized letter foils.
11. 12. 13.
14.
15.
16. 17.
18.
19.
References 1. Heiiman K M. Watson R T, Valenstein E. Neglect and related disorden. ~243 in Clinical Nemopsychology. 2nd ed. (K M Heilman, E Valenstein. eds) Oxford, New York, 1985. 2. Watson R T, Valenstein E, Heilman K M. Normal tactile threshold in monkeys with neglect. Neurology 36: 636. 1986. 3. Weintraub S, Mesulam M-M. Right cerebral dominance in spatial attention. Archives of Neurology 44: 621, 1987. 4. Adams R D. Victor M. Principles of Neurology. 3rd ed. McGraw-Hill. New York. 1985. 5. Tmex R C. Carpenter M B. Strong and Elwyn’s Human Neuroanatomy. 5tb ed. Williams and Wilkins, Baltimore, 1964. 6. Kolb B. Whishaw 1 Q. Fundamentals of Human Neuropsychology. 2nd ed. W H Freeman. New York, 1985. 7. Ross E D. Tbe aprosodias. Archives of Neurology 38: 561, 1981. 8. Kosslyn S M. Aspects of a cognitive neuroscience of mental imagery. Science 240: I62 1, 1988. 9. Sacks 0. Neurological dreams. Medical Doctor 35: 29, 1992. 10. Pylyshyn Z W. What tlte mind’s eye tells the mind’s brain: a critique of mental imagery. Psychological Bulletin 80: 1.1973.
20.
21.
22.
23.
24.
25. 26.
27.
Farah M J. The neurological basis of mental imagery: a componerttial analysis. Cognition 18: 245, 1984. Basso A, Bisiach E, Luzatti C. Loss of mental imagery: a case study. Neuropsychologia 18: 435, 1980. Ross E D. Sensory-specific and fractional disorders of recent memory in man: 1. Isolated loss of visual recent memory. Archives of Neurology 37: 193, 1980. Ross E D. Sensory-specific and fractional disorders of mcent memory in man: II. Unilateral loss of tactile recent memory. Archives of Neurology 37: 267, 1980. Bisiach E, Capitani E, Luzatti C, Perani D. Brain and conscious representation of outside reality. Neuropsychologis 19: 543, 1981. Albert M C. A simple test of visual neglect. Neurology 23: 658, 1973. Gainotti G, Messerli P, Tissot R. Qualitative analysis of tmilsteral spatial neglect in relation to lsterality of cerebral lesicms. Journal of Neurology, Neurosurgery and Psychiatry 35: 545. 1972. Weintrsub S, Ahem G L, Dsffner K R, Price B H, Mesulam M-M. Right sided hemispstial neglect. Neurology (suppl) 42: 223, 1992. Feinberg T E. Haber L D. Stacy C B. lpsilstersl extinction in the hemineglect syndrome. Archives of Neurology 47: 802. 1990. Benton A L. Visuoconstntctive disability in patients with cerebral disease: its relationship to side of lesion and aphasic disorder. Documents Gphthamologica 34: 67, 1973. Benton A. Visuoperceptual. visuospatisl, and visuoconstmctive disorders. ~151 in Clinical Neuropsychology 2nd ed. (K M Heilman, E Valenstein, eds) Oxford, New York, 1985. Cowey A, Stoerig P. Reflections on blindsight pll in the Neuropsychology of Consciousness. (A D Milner, M D Rugg. eds) Academic Press, London, 1992. Chatterjee A, Mennemeier M. Heilmsn K M. Stimulus-response relationship in unilateral neglect. Journal of Clinical and Experimental Neuropsychology 14: 39, 1992. Vallar G, Bottini G. Sterzi R. Pssserini D. Ruscari M L. Hemianesthesia. sensory neglect, and defective access to conscious experience. Neurology 41: 650. 1991. Meador K J. Loring D W. Bowers D. Heilman K M. Remote memory and neglect syndrome. Neurology 37: 522, 1987. Anderson B. Spared awareness for the left side of internal visual images in patients with left sided extrapersonal neglect. Neurology (in press). Benson D F, Geschwind N. Aphasia and related disorders: s clinical approach. ~193 in Principles of Behavioral Neurology. (M-M Mesulam. ed) F A Davis, Philadelphia, 1985.