- Email: [email protected]
A distortion of perceived space in patients withright-hemisphere lesions and visual hemineglect
\ PERGAMON
Neuropsychologia 26 "0888# 808Ð814
A distortion of perceived space in patients with right!hemisphere lesions and visual hemineglect L[ Irving!Bella\ \ M[ Smallb\ A[ Coweya a
b
University of Oxford\ Department of Experiment Psycholo`y\ South Parks Road\ Oxford OX0 2UD\ UK University of Oxford\ Department of Clinical Neurolo`y\ Radcliffe In_rmary\ Woodstock Road\ Oxford OX1 5HE\ UK Received 7 June 0887^ accepted 8 November 0887
Abstract In twelve patients with left visuospatial hemineglect following acute right hemisphere cerebrovascular accident "CVA#\ and in twelve normal controls\ we used a matching task to investigate judgement of the length of lines[ Their task was to choose which of three lines of di}erent length matched a separate single line[ On each trial\ the lines were all vertical or all horizontal[ The set of three lines was presented either to the left or the right of the single line[ When the lines were vertical\ the choices made both by normal subjects and patients were nearly always correct[ However\ when the lines were horizontal and the set of comparison lines was on the left\ patients mostly selected a line that was longer than the sample on the right\ and never chose shorter lines\ suggesting that the length of horizontal lines on the left was underestimated[ When the set of three lines was on the right\ patients chose the correct line more often\ but made errors in both directions\ i[e[ longer or shorter\ although shorter lines were chosen more often than longer lines[ The rare errors made by normal subjects were in the same directions as those observed in the patients[ The results provide further evidence for a non!Euclidean distortion of perceptual space in patients with left sided visuospatial neglect\ and suggest that this distortion could be a gross exaggeration of normal performance or a product of diminished spatial attention\ or both[ Þ 0888 Elsevier Science Ltd[ All rights reserved[ Keywords] Parietal^ Visuo!spatial neglect^ Spatial compression
0[ Introduction Some patients with unilateral visuo!spatial neglect per! ceive shapes in right hemispace as larger than physically identical shapes in left hemispace ð4Ł[ Recent reports have shown that such patients with left visuo!spatial neglect show a consistent bias when making size judgements about stimuli to the left and right of the straight!ahead position ð7\ 02Ł[ This bias consists in underestimating the left stimulus when right and left stimuli are identical[ Moreover\ it seemed that the horizontal extent is more a}ected than the vertical[ The above conclusions stemmed from the dem! onstration of rightward shifts of the perceived {centre| of horizontal lines in such patients ð7\ 02Ł[ In the study by Harvey et al[ ð7Ł the possible e}ects of the manual response were dissociated from the perceptual judgement by asking neglect patients to point to the end of the line that was nearest to a transection mark in a series of pre! transected lines[ Patients made predominantly leftwards Corresponding author[ Tel[] ¦33!90754!160303^ fax] ¦33!90754! 209333^ e!mail] irvingbÝpsy[ox[ac[uk
judgements with bisected lines\ and some leftwards errors were even made for lines on which the transection mark was shifted to the right of the true centre[ Both types of error suggest a {distortion| in the perceived length of the horizontal line[ The same authors tested whether there was similar distortion in the perceived size of items in computer generated images that varied across a number of dimensions\ i[e[\ height\ width\ brightness and spatial frequency[ Compared to control subjects\ who performed almost perfectly\ their patient MJ consistently under! estimated the size of the stimulus on the left of the display compared to the one on the right[ This experiment was subsequently carried out with three further patients\ using pairs of horizontal or vertical rectangles\ and nonsense shapes as stimuli ð01Ł[ Although the patients performed within the normal range when presented with vertical rectangles\ they consistently underestimated the left!sided stimulus when horizontal lines or nonsense _gures were presented[ Collectively\ the preceding experiments sup! port the explanation given by Milner ð8Ł for the right! wards transections made by patients in the line bisection task\ namely that patients with visuo!spatial neglect experience a predominantly horizontal "lateral# com!
9917Ð2821:88:, ! see front matter Þ 0888 Elsevier Science Ltd[ All rights reserved PII] S 9 9 1 7 ! 2 8 2 1 " 8 7 # 9 9 0 3 6 ! X
819
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
pression of the left side of space\ although this can not explain other features of contralateral neglect such as ignoring the left side of an object that is closer to the mid! line than the perceived right side of a more eccentric object "for review see ð09Ł#[ The size distortion was initially interpreted as resulting from a disproportionate allocation of spatial selective attention ð00Ł\ because even normal subjects\ when cued to one end of a centrally presented pre!bisected line\ overestimate that half of the line\ and also transect lines towards the cued end when attempting to mark the geo! metric midpoint ð5\ 7\ 00\ 03Ð05Ł[ In addition\ cueing to the left end of the line in the bisection task in patients can ameliorate the neglect\ i[e[\ the patient places the transection mark more leftwards than in the uncued con! dition[ However\ when pre!bisected lines were cued at one end\ patients pointed more often to the cued end of the line when asked which end the bisection mark was nearest to ð7Ł[ It was concluded\ therefore\ that the cue overtly or covertly orients the eyes towards it\ and that this in~uenced the motor response\ possibly by activating the visuomotor systems in the hemisphere contralateral to the cue ð09Ł[ The following two experiments were designed to inves! tigate this apparent spatial distortion[ In the standard clinical line bisection task\ the patient is asked to bisect a horizontal line that is more or less directly in front and aligned horizontally within the visual _eld[ To try to reduce the tendency to look purely along the horizontal meridian within the visual _eld\ both of the following experiments introduce a choice of three horizontal lines of di}erent lengths on either the right or left of the vertical mid!line\ from which the patient is asked to choose one line[ The three lines are placed one above another so that the patient is encouraged to search vertically as well as horizontally[ The patient must choose the line that mat! ches the single horizontal line positioned on the other lateral half of the sheet[ The experimental results of Harvey et al[ ð7Ł show that if the patient has an option of responding towards the left or right in the {cued| condition\ he:she will respond signi_cantly more often towards the cued side[ The pre! sent experiments were therefore designed to provide a prominent {cue|\ i[e[\ the set of three lines\ but to eliminate the option of a left or right!sided response[ The reason for considering the set of three lines to be a {cue| is that they result in a heavier {weighting| of the display on that side compared to the single line[ In addition\ at the point of making a decision about the matching line\ the patient of necessity is attending to and comparing the three lines\ so this in itself should act like a {cue|[ The response could only be to the left if the con_guration of three lines was on the left\ or to the right if it was on the right\ thus eliminating the confounding e}ect of the response choice present in the experiments of Harvey et al[ ð7Ł\ and dis! cussed further by Milner ð09Ł[ If patients overestimate the
length of an attended stimulus compared to an unattended stimulus\ then the three lines representing the choice may be overestimated\ and the smaller line would be chosen as a match for the single line\ regardless of whether the choice was on the left\ or on the right[ Milner and Harvey ð01Ł found left!stimulus unde! restimations of 04\ 19 and 14) in their three neglect patients[ The actual di}erences in the lengths of the lines in the triad must not\ therefore\ be so great as to exceed the range of any subjective distortions but\ equally\ they need to be su.cient to enable them to be detected[ The 00) di}erence chosen for the present experiments should therefore be su.cient to reveal any subjective distortions\ whilst not being so great as to exceed their e}ects[ Alter! natively\ when the choice of lines was on the left\ patients with an attention bias towards the right should choose the correct line if there was a trade o} between the cueing e}ect towards the left\ and the bias patients have towards the right[
1[ Method Twelve control subjects and six patients were tested in each of two experiments[ Table 0 gives clinical details of the twelve patients tested[ The _rst six patients in the table "DB\ PM\ RA\ DH\ MB and JD# were tested in Experiment 0\ and the remaining six were tested in Experiment 1[ The patients ranged in age from 35 to 73 years "mean 56[4\ S[D[ 03[14#\ and had sustained an acute cerebrovascular accident "CVA# leading to prominent left visual neglect[ They had no known previous CVA\ other on!going neurological disease\ psychiatric disorder\ dementia or confusional state[ All had adequate visual acuity\ for example they could read\ and could sit up and cope easily with a 19 min testing session[ Assessment for visuo!spatial neglect was made using the star cancellation task from the Behavioural Inattention Test ð06Ł[ Informed consent was obtained from all patients in accordance with local ethical committee approval "COREC\ No[ 1300#[ Reconstructions of CT scans for each patient are pro! vided in Fig[ 0 and show the axial section at the maximal extent of the lesion in the plane of the anterior and pos! terior commissures[ Control subjects\ seven male\ _ve female\ were recruited by personal approach[ Their ages ranged from 32 to 65 years "mean 50[14\ S[D[ 09[42#[ Experiments 0 and 1 were carried out with all twelve control subjects[ Separate stimulus displays were prepared for vertical and horizontal lines[ The stimulus lines were drawn onto sheets of A3 paper\ with the longest edges of the paper at the top and bottom[ All trials consisted of a single line on one lateral half of the sheet\ and three lines of di}erent lengths on the other half[ The single line and the middle line in the group of three lines were aligned centrally and\ when horizontal\ colinearly on their respective half of the
810
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814 Table 0 Details of patients taking part in Experiments 0 and 1 Patient
DB PM RA DH MB JD BG GC OE OH AC QF
Sex
M M M M M F F M F F M F
Age
47 35 41 71 42 71 73 64 51 67 45 71
Weeks since stroke
0 0 0 0 0a 0 0 0 0 0 0 0b
Lesion locus
Hemiplegia
R!parieto!temporal R!int[ cap + B[G[ R!MCA area Right int[ capsule and B[G[ R!fronto!parietal Right frontal R!int[ capsule R!parieto!occip[ R occip:R[ coron radiata R!sept[ ovale R!temp!pariet[ R!MCA areab
¦ − ¦ ¦ ¦ ¦ ¦ − ¦ ¦ ¦ ¦
Left visual _eld de_cit
− − − ¦ ¦ ¦ − − − ¦ − −
Star cancel score "B[I[T[#c L
R
08:16 9:16 10:16 9:16 9:16 2:16 9:16 9:16 9:16 9:16 9:16 9:16
16:16 06:16 16:16 19:16 02:16 14:16 04:16 7:16 19:16 09:16 14:16 02:16
a
Patient MB also has an old infarct high in R!parietal lobe "date not known#[ Patient QF also has an old infarct in the righ basal ganglia "date not known#[ c The scores on the star cancellation test refer to the number of cancellations in the left and right halves of the display[ b
Fig[ 0[ CT scans showing maximum extent of the lesion superimposed on the horizontal templates taken from _gure A[03 of Damasio and Damasio ð2Ł[ This axis corresponds most closely to the axis of the scans[
stimulus sheet[ The single line was always 89 mm in length[ The three lines forming the {choice| on the other part of the sheet were 099\ 89\ and 79 mm in length\ and were 19 mm apart[ The presentation of the lines was balanced so that the same number were presented with the {choice| on the left and right of the display[ In addition\ the three lines of di}erent length were assigned in a counterbalanced design to all six possible relative positions[ This created 13 possible displays\ 01 hori! zontal\ 01 vertical\ each of which was presented twice\ giving each subject 37 trials in all\ presented in random order[ As the geometrically correct choice was always the 89 mm line the subject could gain a perfect score by always selecting the line of intermediate length[ However\ all subjects were told that line lengths varied slightly\ all looked back and forth between sample and matches\ and no feedback about performance was given until the end of testing[ Using greater variation in line length was not possible in the time available with each patient[ When referring to {hemispace presentation|\ this refers to the set of three lines representing the {choice|[ Where possible\ patients were seated with the exper! imenter facing them[ If the patient was in bed\ the exper! imenter sat to the patient|s right[ In all cases\ the displays were aligned with the patient|s mid!sagittal plane[ Stimu! lus sheets were always presented to control subjects with the experimenter sitting opposite at a small table[ The experimenter explained that the task involved looking at a single line\ and then choosing which of three other lines most closely matched it in length[ The experimenter placed stimulus sheets in front of the patient one at a time\ and indicated the single line verbally and by pointing\ and then indicated the three lines from which the {choice| had
811
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
to be made[ If a patient took more than about 6 s to respond\ the experimenter reminded the patient what he:she had to do[ The patient usually pointed to the chosen line\ but occasionally indicated it verbally "e[g[ when the lines were horizontal\ by saying\ {top\ middle|\ or {bottom|#[ The same procedure was carried out with the control subjects\ although the initial explanation of the task requirements was usually all that was required[ 2[ Results The results of Experiment 0 are shown in Fig[ 1\ which shows the mean number of choices of each line length for horizontal trials when the choice was on the left or on the right[ On trials with vertical lines\ performance was almost errorless[ In contrast\ with horizontal lines per! formance was strikingly imperfect[ The absolute numbers of responses for each line length for patients and control subjects were analysed to see if the choice of line length depended on hemispace presentation[ Patients "DF 1\ x1 59[79\ P ³ 9[9990#\ and control subjects "DF 1\ x1 01[33\ P ³ 9[90#\ showed a signi_cant e}ect of hem!
ispace presentation on line length choice[ A further analy! sis was carried out between the two groups of subjects for presentation of the choice!lines in left and right hem! ispace[ The patient and control groups were signi_cantly di}erent for both left "DF 1\ x1 51[74\ P ³ 9[9990#\ and right hemispace "DF 1\ x1 21[46\ P ³ 9[9990#[ For left hemispace\ errors were almost entirely towards choosing the longer line length in both subject groups\ but for right hemispace\ errors by both groups were in both directions[ However\ patients chose a shorter line length more often than a longer line whereas controls chose a longer line length more often than a shorter one\ although the latter di}erence was slight[ In summary\ patients usually chose a longer line with left hemispace presentation\ whereas with right hemispace presentation\ responses were shifted towards the shorter line length choice[ 3[ Experiment 1 The second experiment examined the e}ect of aligning of the set of three lines at one end[ In Experiment 0\ the longest line always protruded further rightwards than the other two lines\ and this may have in~uenced the choice[ When the choice is on the left\ and the lines are aligned at their left ends\ the longest line would protrude even further rightwards than the other two lines\ whereas if they are aligned at the right\ none of the lines would protrude rightwards[ The same applies to right hemispace presentation[ If this was in~uencing line choice\ it is likely that there would be a signi_cant di}erence between the two di}erent types of alignment in left hemispace\ and the same two in right hemispace[ As no e}ect of line length had been observed in the vertical line presentations in Experiment 0\ only horizontal lines were used for this experiment[ Stimulus displays were essentially as before\ with the di}erence that the three lines representing the {choice| were either aligned at their left ends\ or their right ends[ This provided four conditions for the experiment\ namely] choice left\ aligned left or aligned right\ and choice right\ aligned left or aligned right[ 3[0[ Subjects and method Twelve control subjects and the last six patients shown in Table 0 "BG\ GC\ OE\ OH\ AC\ QF#\ were tested[ The methods were the same as in Experiment 0[ 3[1[ Results
Fig[ 1[ E}ect of hemispace on line length perception[
Figure 2 shows the mean number of choices for trials with horizontal lines when the choice was on the left "aligned left or right#\ and when the choice was on the right "aligned left or right#[ The absolute numbers of responses for each condition and each line length for
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
812
ni_cantly di}erent for both left "DF 1\ x1 79[87\ P ³ 9[9990#\ and right hemispace "DF 1\ x1 010[26\ P ³ 9[9990#[
4[ Discussion
Fig[ 2[ E}ect of hemispace and alignment of stimulus lines on length perception[
patients and control subjects were analysed to see if the length of the chosen line depended on hemispace pres! entation and:or the side at which the stimulus lines were aligned[ For patients\ the di}erence between results for left or right alignment just failed to reach the 9[94 sig! ni_cance level in left hemispace "DF 1\ x1 4[7\ P 9[95# and was non!signi_cant for right hemispace "DF 1\ x1 9[94\ P 9[87#[ In the control group\ the di}erence between left and right alignment when lines were presented on the right just failed to reach sig! ni_cance "DF 1\ x1 4[97\ P 9[97#[ However\ there was a signi_cant di}erence for control subjects between left and right alignment when the three choice lines were in left hemispace "DF 1\ x1 03[35\ P × 9[990#[ The absolute numbers of responses for all left hemispace and right hemispace presentations were compared within each subject group[ Both patients "DF 1\ x1 67[3\ P ³ 9[9990#\ and control subjects "DF 1\ x1 29[95\ P ³ 9[9990#\ showed a signi_cant e}ect of hemispace presentation on line length choice[ The absolute number of responses for all left hemispace presentations and all right hemispace presentations were compared between patient and control groups[ The two groups were sig!
Although there was a signi_cant and large di}erence between the performance of the patient and control groups\ the results of both experiments are similar in that for both groups of subjects the results of selecting a line that matched the sample line depended on whether the single line was on the left and the array of three lines on the right\ or vice versa[ This asymmetry was not predicted from the presumed {cueing e}ect| of the set of three lines[ Cueing to one end of a bisected line in normal subjects leads to an overestimation of the size of the cued half of the line ð7Ł\ and can apparently ameliorate neglect in patients with visuo!spatial neglect ð5\ 03Ð05Ł[ The pre! diction for both groups was therefore that the shorter line would be chosen if the only factor in~uencing the choice was the allocation of selective attention[ The pat! ients did not choose the shorter line with right hemispace presentation as often as they chose the longer line with left hemispace presentation[ Most errors in the control group were towards choosing the longer line with both left and right hemispace presentation\ but they chose the shorter line more often with right hemispace presentation than left[ In fact\ both groups made the correct choice more often with right rather than left hemispace pres! entation[ This suggests that the asymmetry present in patients may be a gross exaggeration of a normal phenomenon[ What the two experiments show very clearly is that patients with left!sided visuo!spatial neglect experience a distortion of subjective space on the left that exceeds any normal asymmetry\ and that may not be solely the result of a disproportionate allocation of selective attention\ or a bias in motor response[ Although the results from patients are asymmetrical\ they do suggest that the dis! tortion of space that exists is non!Euclidean ð4\ 8Ł\ as there was a substantial tendency towards choosing a longer line to match the single line on the right when the line choice is in left hemispace\ and a shift towards choosing a shorter line to match the single line on the left when the choice of lines is in right hemispace[ This supports previous _ndings using di}erent methods of evaluating linear dis! tortion ð7\ 01\ 02Ł[ Such an asymmetry in how the brain forms or attends to representations should not be surpris! ing\ for some representational de_cits\ e[g[ apperceptive object and face agnosia\ are more likely to occur after right hemisphere than left hemisphere lesions[ A further way of assessing the contribution\ if any\ of the perceptual distortion to unilateral neglect is to exam! ine the correlation between them[ When the severity of the neglect\ as determined by the star cancellation test\
813
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
was compared with the magnitude of the perceptual dis! tortion there was no signi_cant correlation "Spearman Test\ rs 9[11\ t 9[60^ P × 9[1#[ The perceptual dis! tortion is therefore more likely to be an associated symp! tom than a cause of the neglect[ As the distortion of subjective space in the patient group is not entirely explicable by attentional factors or a motor bias\ this element of neglect suggests that the so! called ventral visual stream could be involved in the lesion\ in addition to the parietal damage that undoubt! edly gives rise to the more prominent hemineglect[ Although CT scans do not always show that the temporal lobe is also damaged in patients with visuo!spatial neglect\ the most e}ective lesion occupies the parieto! occipito!temporal junction and the blood supply to the temporal lobe may be compromised if the stroke involves the middle cerebral artery[ Cerebro!vascular accidents involving the middle cerebral artery may a}ect one or more of the artery|s three branches which respectively supply the cortex and white matter of the lateral and inferior parts of the frontal lobe\ of the parietal lobe including the somatosensory cortex\ angular and sup! ramarginal convolutions\ and of the superior parts of the temporal lobe and insula[ The penetrating branches supply the putamen\ part of the head and body of the caudate nucleus\ the outer globus pallidus\ the posterior limb of the internal capsule and corona radiata[ Although each of the two divisions of the MCA in the sylvian sulcus and their major branches can be occluded separately\ an embolus entering the MCA often lodges in one of its two main divisions\ either supplying the rolandic and pre! rolandic area or\ inferiorly\ supplying the inferior parietal and lateral temporal lobe ð0Ł[ It is likely therefore that damage involving the most inferior part of the lateral parietal lobe will also a}ect the dorso lateral temporal lobe[ The perceptual distortion present in patients with visuo!spatial neglect may be similar to that observed in hemi!micropsia\ a visual disorder that is rare in pure form and characterised by an apparent reduction of the size of objects in one hemi_eld[ Micropsia after focal brain lesions was reported in two patients by Cohen et al[ ð1Ł[ Patient 0 experienced a sudden hemianopia[ Although the _eld defect rapidly regressed\ one week afterwards he complained that objects in his left visual _eld appeared to have shrunk[ This was conspicuous in his drawings\ where the left!sided elements were drawn physically larger than those on the right\ when they should have been symmetrical[ When presented with symmetrical pat! terns\ he perceived the elements on the left as smaller than those on the right[ The patient died 16 months later\ and a post!mortem revealed destruction of the inferior occipital gyrus and the inferior part of the middle occipi! tal gyrus\ regions corresponding to the lower part of the lateral aspect of areas 07 and 08[ Laterally\ the lesion also a}ected the depths of the inferior temporal sulcus\
part of the middle temporal gyrus\ and the depths of the superior temporal sulcus[ The second patient perceived objects on the right as smaller than those on the left[ An MRI revealed a circumscribed left hemisphere lesion a}ecting the lower lateral aspect of areas 07 and 08 and the underlying white matter[ Further investigations are required to see whether the perceptual distortion observed in some patients with visuo!spatial neglect cor! responds to micropsia\ but as the areas involved in both conditions are adjacent\ this is highly feasible\ par! ticularly if subcortical damage could compromise both areas 07 and 08 of Brodmann[ All the patients in the current study had a right hemisphere lesion and left! sided neglect and\ perhaps crucially\ all had sub!cortical damage\ extensive in some cases\ in the immediate vicinity of the focal cortical damage that leads to micropsia[ The perceptual distortion described here appears to be a gross exaggeration of normal performance[ Our suggestion is that it could have either or both of two bases\ an attentional one and a genuine perceptual dis! tortion[ However\ in decribing the distortion we have throughout referred to it as a shrinkage on the left even though\ theoretically\ it could be an expansion on the right[ We have no direct evidence that bears on this possibility but in hemimicropsia the patient actually claims that things are unusually small contralaterally\ not large ipsilaterally[ Furthermore\ a right!sided expansion would have to be more prominent in the horizontal axis to explain the results\ and although such a change in horizontal extent might remain unnoticed by the patient within the neglected hemi_eld\ it should be noticed if it occurs in the non!neglected hemi_eld[ This did not occur[ A curious aspect of unilateral visuo!spatial neglect is that when studied by line bisection\ the characteristic rightward errors can be greater\ in angular terms\ in near space "within reaching distance# ð6Ł or in far space ð2Ł[ The pathological basis for this dissociation is unclear[ If the mis!bisection occurs because the patient experiences a subjective contraction of visual space on the left\ as suggested here and by others ð02Ł\ then the so!called dis! tance e}ect should be present in a matching task like that used here\ as well as for line bisection\ unless the distance e}ect arises for other reasons such as di}erent motor response programmes for near and far space[ We are now pursuing this possibility[ Acknowledgements This research was supported by the MRC IRC in Brain and Behaviour\ Oxford\ and by an Oxford McDonnell! Pew Studentship to L[ Irving!Bell[ References ð0Ł Adams RD\ Victor M[ Principles of neurology[ 3th ed[ New York] McGraw!Hill\ 0889]516Ð7[
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814 ð1Ł Cohen L\ Gray F\ Meyrignac C\ Dehaene S\ Degos JD[ Selective de_cit of size perception] two cases of hemimicropsia[ Journal of Neurology\ Neurosurgery and Psychiatry 0883^46"0#]62Ð7[ ð2Ł Cowey A\ Small M\ Ellis S[ Left visuo!spatial neglect can be worse in far than in near space[ Neuropsychologia 0883^21"8#]0948Ð55[ ð3Ł Damasio H\ Damasio HR[ Lesion analysis in neuropsychology[ New York] Oxford University Press\ 0878[ ð4Ł Gainotti G\ Tiacci C[ The relationship between disorders of visual perception and unilateral spatial neglect[ Neuropsychologia 0860^8]340Ð7[ ð5Ł Halligan PW\ Marshall JC[ Perceptual cueing and perceptuo! motor compatability in visuo!spatial neglect] a single case study[ Cognitive Neuropsychology 0878^5"3#]312Ð24[ ð6Ł Halligan PW\ Marshall JC[ Left neglect for near but not far space in man[ Nature 0880^249]387Ð499[ ð7Ł Harvey M\ Milner AD\ Roberts RC[ An investigation of hemi! spatial neglect using the landmark task[ Brain and Cognition 0884^16]48Ð67[ ð8Ł Milner AD[ Animal models for the syndrome of spatial neglect[ In] Jeannerod M\ editor[ Neurophysiological and neuropsychological aspects of spatial neglect[ Amsterdam] Elsevier\ 0876]148Ð77[ ð09Ł Milner AD[ Neglect\ extinction\ and the cortical streams of visual
ð00Ł
ð01Ł ð02Ł
ð03Ł
ð04Ł
ð05Ł ð06Ł
814
processing[ In] Thier P\ Karnath H!O\ editors[ Parietal lobe con! tributions to orientation in 2D space[ Heidelberg] Springer!Verlag\ 0886]2Ð11[ Milner AD\ Brechmann M\ Pagliarini L[ To halve and to halve not] an analysis of line bisection judgements in normal subjects[ Neuropsychologia 0881^29]404Ð15[ Milner AD\ Harvey M[ Distortion of size perception in vis! uospatial neglect[ Current Biology 0884^4"0#]74Ð8[ Milner AD\ Harvey M\ Roberts RC\ Forster SV[ Line bisection errors in visual neglect] misguided action or size distortion< Neuro! psychologia 0882^20"0#]28Ð38[ Nichelli P\ Rinaldi M\ Cubelli R[ Selective spatial attention and length representation in normal subjects and in patients with uni! lateral spatial neglect[ Brain and Cognition 0878^8]46Ð69[ Reuter!Lorenz PA\ Posner MI[ Components of neglect from right hemisphere damage] an analysis of line bisection[ Neu! ropsychologia 0889^17"3#]216Ð22[ Riddoch MJ\ Humphries G[ The e}ect of cueing on unilateral neglect[ Neuropsychologia 0872^10]478Ð88[ Wilson B\ Cockburn J\ Halligan PW[ Behavioural Inattention Test[ Titch_eld\ Hampshire] Thames Valley Test Company[
Neuropsychologia 26 "0888# 808Ð814
A distortion of perceived space in patients with right!hemisphere lesions and visual hemineglect L[ Irving!Bella\ \ M[ Smallb\ A[ Coweya a
b
University of Oxford\ Department of Experiment Psycholo`y\ South Parks Road\ Oxford OX0 2UD\ UK University of Oxford\ Department of Clinical Neurolo`y\ Radcliffe In_rmary\ Woodstock Road\ Oxford OX1 5HE\ UK Received 7 June 0887^ accepted 8 November 0887
Abstract In twelve patients with left visuospatial hemineglect following acute right hemisphere cerebrovascular accident "CVA#\ and in twelve normal controls\ we used a matching task to investigate judgement of the length of lines[ Their task was to choose which of three lines of di}erent length matched a separate single line[ On each trial\ the lines were all vertical or all horizontal[ The set of three lines was presented either to the left or the right of the single line[ When the lines were vertical\ the choices made both by normal subjects and patients were nearly always correct[ However\ when the lines were horizontal and the set of comparison lines was on the left\ patients mostly selected a line that was longer than the sample on the right\ and never chose shorter lines\ suggesting that the length of horizontal lines on the left was underestimated[ When the set of three lines was on the right\ patients chose the correct line more often\ but made errors in both directions\ i[e[ longer or shorter\ although shorter lines were chosen more often than longer lines[ The rare errors made by normal subjects were in the same directions as those observed in the patients[ The results provide further evidence for a non!Euclidean distortion of perceptual space in patients with left sided visuospatial neglect\ and suggest that this distortion could be a gross exaggeration of normal performance or a product of diminished spatial attention\ or both[ Þ 0888 Elsevier Science Ltd[ All rights reserved[ Keywords] Parietal^ Visuo!spatial neglect^ Spatial compression
0[ Introduction Some patients with unilateral visuo!spatial neglect per! ceive shapes in right hemispace as larger than physically identical shapes in left hemispace ð4Ł[ Recent reports have shown that such patients with left visuo!spatial neglect show a consistent bias when making size judgements about stimuli to the left and right of the straight!ahead position ð7\ 02Ł[ This bias consists in underestimating the left stimulus when right and left stimuli are identical[ Moreover\ it seemed that the horizontal extent is more a}ected than the vertical[ The above conclusions stemmed from the dem! onstration of rightward shifts of the perceived {centre| of horizontal lines in such patients ð7\ 02Ł[ In the study by Harvey et al[ ð7Ł the possible e}ects of the manual response were dissociated from the perceptual judgement by asking neglect patients to point to the end of the line that was nearest to a transection mark in a series of pre! transected lines[ Patients made predominantly leftwards Corresponding author[ Tel[] ¦33!90754!160303^ fax] ¦33!90754! 209333^ e!mail] irvingbÝpsy[ox[ac[uk
judgements with bisected lines\ and some leftwards errors were even made for lines on which the transection mark was shifted to the right of the true centre[ Both types of error suggest a {distortion| in the perceived length of the horizontal line[ The same authors tested whether there was similar distortion in the perceived size of items in computer generated images that varied across a number of dimensions\ i[e[\ height\ width\ brightness and spatial frequency[ Compared to control subjects\ who performed almost perfectly\ their patient MJ consistently under! estimated the size of the stimulus on the left of the display compared to the one on the right[ This experiment was subsequently carried out with three further patients\ using pairs of horizontal or vertical rectangles\ and nonsense shapes as stimuli ð01Ł[ Although the patients performed within the normal range when presented with vertical rectangles\ they consistently underestimated the left!sided stimulus when horizontal lines or nonsense _gures were presented[ Collectively\ the preceding experiments sup! port the explanation given by Milner ð8Ł for the right! wards transections made by patients in the line bisection task\ namely that patients with visuo!spatial neglect experience a predominantly horizontal "lateral# com!
9917Ð2821:88:, ! see front matter Þ 0888 Elsevier Science Ltd[ All rights reserved PII] S 9 9 1 7 ! 2 8 2 1 " 8 7 # 9 9 0 3 6 ! X
819
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
pression of the left side of space\ although this can not explain other features of contralateral neglect such as ignoring the left side of an object that is closer to the mid! line than the perceived right side of a more eccentric object "for review see ð09Ł#[ The size distortion was initially interpreted as resulting from a disproportionate allocation of spatial selective attention ð00Ł\ because even normal subjects\ when cued to one end of a centrally presented pre!bisected line\ overestimate that half of the line\ and also transect lines towards the cued end when attempting to mark the geo! metric midpoint ð5\ 7\ 00\ 03Ð05Ł[ In addition\ cueing to the left end of the line in the bisection task in patients can ameliorate the neglect\ i[e[\ the patient places the transection mark more leftwards than in the uncued con! dition[ However\ when pre!bisected lines were cued at one end\ patients pointed more often to the cued end of the line when asked which end the bisection mark was nearest to ð7Ł[ It was concluded\ therefore\ that the cue overtly or covertly orients the eyes towards it\ and that this in~uenced the motor response\ possibly by activating the visuomotor systems in the hemisphere contralateral to the cue ð09Ł[ The following two experiments were designed to inves! tigate this apparent spatial distortion[ In the standard clinical line bisection task\ the patient is asked to bisect a horizontal line that is more or less directly in front and aligned horizontally within the visual _eld[ To try to reduce the tendency to look purely along the horizontal meridian within the visual _eld\ both of the following experiments introduce a choice of three horizontal lines of di}erent lengths on either the right or left of the vertical mid!line\ from which the patient is asked to choose one line[ The three lines are placed one above another so that the patient is encouraged to search vertically as well as horizontally[ The patient must choose the line that mat! ches the single horizontal line positioned on the other lateral half of the sheet[ The experimental results of Harvey et al[ ð7Ł show that if the patient has an option of responding towards the left or right in the {cued| condition\ he:she will respond signi_cantly more often towards the cued side[ The pre! sent experiments were therefore designed to provide a prominent {cue|\ i[e[\ the set of three lines\ but to eliminate the option of a left or right!sided response[ The reason for considering the set of three lines to be a {cue| is that they result in a heavier {weighting| of the display on that side compared to the single line[ In addition\ at the point of making a decision about the matching line\ the patient of necessity is attending to and comparing the three lines\ so this in itself should act like a {cue|[ The response could only be to the left if the con_guration of three lines was on the left\ or to the right if it was on the right\ thus eliminating the confounding e}ect of the response choice present in the experiments of Harvey et al[ ð7Ł\ and dis! cussed further by Milner ð09Ł[ If patients overestimate the
length of an attended stimulus compared to an unattended stimulus\ then the three lines representing the choice may be overestimated\ and the smaller line would be chosen as a match for the single line\ regardless of whether the choice was on the left\ or on the right[ Milner and Harvey ð01Ł found left!stimulus unde! restimations of 04\ 19 and 14) in their three neglect patients[ The actual di}erences in the lengths of the lines in the triad must not\ therefore\ be so great as to exceed the range of any subjective distortions but\ equally\ they need to be su.cient to enable them to be detected[ The 00) di}erence chosen for the present experiments should therefore be su.cient to reveal any subjective distortions\ whilst not being so great as to exceed their e}ects[ Alter! natively\ when the choice of lines was on the left\ patients with an attention bias towards the right should choose the correct line if there was a trade o} between the cueing e}ect towards the left\ and the bias patients have towards the right[
1[ Method Twelve control subjects and six patients were tested in each of two experiments[ Table 0 gives clinical details of the twelve patients tested[ The _rst six patients in the table "DB\ PM\ RA\ DH\ MB and JD# were tested in Experiment 0\ and the remaining six were tested in Experiment 1[ The patients ranged in age from 35 to 73 years "mean 56[4\ S[D[ 03[14#\ and had sustained an acute cerebrovascular accident "CVA# leading to prominent left visual neglect[ They had no known previous CVA\ other on!going neurological disease\ psychiatric disorder\ dementia or confusional state[ All had adequate visual acuity\ for example they could read\ and could sit up and cope easily with a 19 min testing session[ Assessment for visuo!spatial neglect was made using the star cancellation task from the Behavioural Inattention Test ð06Ł[ Informed consent was obtained from all patients in accordance with local ethical committee approval "COREC\ No[ 1300#[ Reconstructions of CT scans for each patient are pro! vided in Fig[ 0 and show the axial section at the maximal extent of the lesion in the plane of the anterior and pos! terior commissures[ Control subjects\ seven male\ _ve female\ were recruited by personal approach[ Their ages ranged from 32 to 65 years "mean 50[14\ S[D[ 09[42#[ Experiments 0 and 1 were carried out with all twelve control subjects[ Separate stimulus displays were prepared for vertical and horizontal lines[ The stimulus lines were drawn onto sheets of A3 paper\ with the longest edges of the paper at the top and bottom[ All trials consisted of a single line on one lateral half of the sheet\ and three lines of di}erent lengths on the other half[ The single line and the middle line in the group of three lines were aligned centrally and\ when horizontal\ colinearly on their respective half of the
810
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814 Table 0 Details of patients taking part in Experiments 0 and 1 Patient
DB PM RA DH MB JD BG GC OE OH AC QF
Sex
M M M M M F F M F F M F
Age
47 35 41 71 42 71 73 64 51 67 45 71
Weeks since stroke
0 0 0 0 0a 0 0 0 0 0 0 0b
Lesion locus
Hemiplegia
R!parieto!temporal R!int[ cap + B[G[ R!MCA area Right int[ capsule and B[G[ R!fronto!parietal Right frontal R!int[ capsule R!parieto!occip[ R occip:R[ coron radiata R!sept[ ovale R!temp!pariet[ R!MCA areab
¦ − ¦ ¦ ¦ ¦ ¦ − ¦ ¦ ¦ ¦
Left visual _eld de_cit
− − − ¦ ¦ ¦ − − − ¦ − −
Star cancel score "B[I[T[#c L
R
08:16 9:16 10:16 9:16 9:16 2:16 9:16 9:16 9:16 9:16 9:16 9:16
16:16 06:16 16:16 19:16 02:16 14:16 04:16 7:16 19:16 09:16 14:16 02:16
a
Patient MB also has an old infarct high in R!parietal lobe "date not known#[ Patient QF also has an old infarct in the righ basal ganglia "date not known#[ c The scores on the star cancellation test refer to the number of cancellations in the left and right halves of the display[ b
Fig[ 0[ CT scans showing maximum extent of the lesion superimposed on the horizontal templates taken from _gure A[03 of Damasio and Damasio ð2Ł[ This axis corresponds most closely to the axis of the scans[
stimulus sheet[ The single line was always 89 mm in length[ The three lines forming the {choice| on the other part of the sheet were 099\ 89\ and 79 mm in length\ and were 19 mm apart[ The presentation of the lines was balanced so that the same number were presented with the {choice| on the left and right of the display[ In addition\ the three lines of di}erent length were assigned in a counterbalanced design to all six possible relative positions[ This created 13 possible displays\ 01 hori! zontal\ 01 vertical\ each of which was presented twice\ giving each subject 37 trials in all\ presented in random order[ As the geometrically correct choice was always the 89 mm line the subject could gain a perfect score by always selecting the line of intermediate length[ However\ all subjects were told that line lengths varied slightly\ all looked back and forth between sample and matches\ and no feedback about performance was given until the end of testing[ Using greater variation in line length was not possible in the time available with each patient[ When referring to {hemispace presentation|\ this refers to the set of three lines representing the {choice|[ Where possible\ patients were seated with the exper! imenter facing them[ If the patient was in bed\ the exper! imenter sat to the patient|s right[ In all cases\ the displays were aligned with the patient|s mid!sagittal plane[ Stimu! lus sheets were always presented to control subjects with the experimenter sitting opposite at a small table[ The experimenter explained that the task involved looking at a single line\ and then choosing which of three other lines most closely matched it in length[ The experimenter placed stimulus sheets in front of the patient one at a time\ and indicated the single line verbally and by pointing\ and then indicated the three lines from which the {choice| had
811
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
to be made[ If a patient took more than about 6 s to respond\ the experimenter reminded the patient what he:she had to do[ The patient usually pointed to the chosen line\ but occasionally indicated it verbally "e[g[ when the lines were horizontal\ by saying\ {top\ middle|\ or {bottom|#[ The same procedure was carried out with the control subjects\ although the initial explanation of the task requirements was usually all that was required[ 2[ Results The results of Experiment 0 are shown in Fig[ 1\ which shows the mean number of choices of each line length for horizontal trials when the choice was on the left or on the right[ On trials with vertical lines\ performance was almost errorless[ In contrast\ with horizontal lines per! formance was strikingly imperfect[ The absolute numbers of responses for each line length for patients and control subjects were analysed to see if the choice of line length depended on hemispace presentation[ Patients "DF 1\ x1 59[79\ P ³ 9[9990#\ and control subjects "DF 1\ x1 01[33\ P ³ 9[90#\ showed a signi_cant e}ect of hem!
ispace presentation on line length choice[ A further analy! sis was carried out between the two groups of subjects for presentation of the choice!lines in left and right hem! ispace[ The patient and control groups were signi_cantly di}erent for both left "DF 1\ x1 51[74\ P ³ 9[9990#\ and right hemispace "DF 1\ x1 21[46\ P ³ 9[9990#[ For left hemispace\ errors were almost entirely towards choosing the longer line length in both subject groups\ but for right hemispace\ errors by both groups were in both directions[ However\ patients chose a shorter line length more often than a longer line whereas controls chose a longer line length more often than a shorter one\ although the latter di}erence was slight[ In summary\ patients usually chose a longer line with left hemispace presentation\ whereas with right hemispace presentation\ responses were shifted towards the shorter line length choice[ 3[ Experiment 1 The second experiment examined the e}ect of aligning of the set of three lines at one end[ In Experiment 0\ the longest line always protruded further rightwards than the other two lines\ and this may have in~uenced the choice[ When the choice is on the left\ and the lines are aligned at their left ends\ the longest line would protrude even further rightwards than the other two lines\ whereas if they are aligned at the right\ none of the lines would protrude rightwards[ The same applies to right hemispace presentation[ If this was in~uencing line choice\ it is likely that there would be a signi_cant di}erence between the two di}erent types of alignment in left hemispace\ and the same two in right hemispace[ As no e}ect of line length had been observed in the vertical line presentations in Experiment 0\ only horizontal lines were used for this experiment[ Stimulus displays were essentially as before\ with the di}erence that the three lines representing the {choice| were either aligned at their left ends\ or their right ends[ This provided four conditions for the experiment\ namely] choice left\ aligned left or aligned right\ and choice right\ aligned left or aligned right[ 3[0[ Subjects and method Twelve control subjects and the last six patients shown in Table 0 "BG\ GC\ OE\ OH\ AC\ QF#\ were tested[ The methods were the same as in Experiment 0[ 3[1[ Results
Fig[ 1[ E}ect of hemispace on line length perception[
Figure 2 shows the mean number of choices for trials with horizontal lines when the choice was on the left "aligned left or right#\ and when the choice was on the right "aligned left or right#[ The absolute numbers of responses for each condition and each line length for
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
812
ni_cantly di}erent for both left "DF 1\ x1 79[87\ P ³ 9[9990#\ and right hemispace "DF 1\ x1 010[26\ P ³ 9[9990#[
4[ Discussion
Fig[ 2[ E}ect of hemispace and alignment of stimulus lines on length perception[
patients and control subjects were analysed to see if the length of the chosen line depended on hemispace pres! entation and:or the side at which the stimulus lines were aligned[ For patients\ the di}erence between results for left or right alignment just failed to reach the 9[94 sig! ni_cance level in left hemispace "DF 1\ x1 4[7\ P 9[95# and was non!signi_cant for right hemispace "DF 1\ x1 9[94\ P 9[87#[ In the control group\ the di}erence between left and right alignment when lines were presented on the right just failed to reach sig! ni_cance "DF 1\ x1 4[97\ P 9[97#[ However\ there was a signi_cant di}erence for control subjects between left and right alignment when the three choice lines were in left hemispace "DF 1\ x1 03[35\ P × 9[990#[ The absolute numbers of responses for all left hemispace and right hemispace presentations were compared within each subject group[ Both patients "DF 1\ x1 67[3\ P ³ 9[9990#\ and control subjects "DF 1\ x1 29[95\ P ³ 9[9990#\ showed a signi_cant e}ect of hemispace presentation on line length choice[ The absolute number of responses for all left hemispace presentations and all right hemispace presentations were compared between patient and control groups[ The two groups were sig!
Although there was a signi_cant and large di}erence between the performance of the patient and control groups\ the results of both experiments are similar in that for both groups of subjects the results of selecting a line that matched the sample line depended on whether the single line was on the left and the array of three lines on the right\ or vice versa[ This asymmetry was not predicted from the presumed {cueing e}ect| of the set of three lines[ Cueing to one end of a bisected line in normal subjects leads to an overestimation of the size of the cued half of the line ð7Ł\ and can apparently ameliorate neglect in patients with visuo!spatial neglect ð5\ 03Ð05Ł[ The pre! diction for both groups was therefore that the shorter line would be chosen if the only factor in~uencing the choice was the allocation of selective attention[ The pat! ients did not choose the shorter line with right hemispace presentation as often as they chose the longer line with left hemispace presentation[ Most errors in the control group were towards choosing the longer line with both left and right hemispace presentation\ but they chose the shorter line more often with right hemispace presentation than left[ In fact\ both groups made the correct choice more often with right rather than left hemispace pres! entation[ This suggests that the asymmetry present in patients may be a gross exaggeration of a normal phenomenon[ What the two experiments show very clearly is that patients with left!sided visuo!spatial neglect experience a distortion of subjective space on the left that exceeds any normal asymmetry\ and that may not be solely the result of a disproportionate allocation of selective attention\ or a bias in motor response[ Although the results from patients are asymmetrical\ they do suggest that the dis! tortion of space that exists is non!Euclidean ð4\ 8Ł\ as there was a substantial tendency towards choosing a longer line to match the single line on the right when the line choice is in left hemispace\ and a shift towards choosing a shorter line to match the single line on the left when the choice of lines is in right hemispace[ This supports previous _ndings using di}erent methods of evaluating linear dis! tortion ð7\ 01\ 02Ł[ Such an asymmetry in how the brain forms or attends to representations should not be surpris! ing\ for some representational de_cits\ e[g[ apperceptive object and face agnosia\ are more likely to occur after right hemisphere than left hemisphere lesions[ A further way of assessing the contribution\ if any\ of the perceptual distortion to unilateral neglect is to exam! ine the correlation between them[ When the severity of the neglect\ as determined by the star cancellation test\
813
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814
was compared with the magnitude of the perceptual dis! tortion there was no signi_cant correlation "Spearman Test\ rs 9[11\ t 9[60^ P × 9[1#[ The perceptual dis! tortion is therefore more likely to be an associated symp! tom than a cause of the neglect[ As the distortion of subjective space in the patient group is not entirely explicable by attentional factors or a motor bias\ this element of neglect suggests that the so! called ventral visual stream could be involved in the lesion\ in addition to the parietal damage that undoubt! edly gives rise to the more prominent hemineglect[ Although CT scans do not always show that the temporal lobe is also damaged in patients with visuo!spatial neglect\ the most e}ective lesion occupies the parieto! occipito!temporal junction and the blood supply to the temporal lobe may be compromised if the stroke involves the middle cerebral artery[ Cerebro!vascular accidents involving the middle cerebral artery may a}ect one or more of the artery|s three branches which respectively supply the cortex and white matter of the lateral and inferior parts of the frontal lobe\ of the parietal lobe including the somatosensory cortex\ angular and sup! ramarginal convolutions\ and of the superior parts of the temporal lobe and insula[ The penetrating branches supply the putamen\ part of the head and body of the caudate nucleus\ the outer globus pallidus\ the posterior limb of the internal capsule and corona radiata[ Although each of the two divisions of the MCA in the sylvian sulcus and their major branches can be occluded separately\ an embolus entering the MCA often lodges in one of its two main divisions\ either supplying the rolandic and pre! rolandic area or\ inferiorly\ supplying the inferior parietal and lateral temporal lobe ð0Ł[ It is likely therefore that damage involving the most inferior part of the lateral parietal lobe will also a}ect the dorso lateral temporal lobe[ The perceptual distortion present in patients with visuo!spatial neglect may be similar to that observed in hemi!micropsia\ a visual disorder that is rare in pure form and characterised by an apparent reduction of the size of objects in one hemi_eld[ Micropsia after focal brain lesions was reported in two patients by Cohen et al[ ð1Ł[ Patient 0 experienced a sudden hemianopia[ Although the _eld defect rapidly regressed\ one week afterwards he complained that objects in his left visual _eld appeared to have shrunk[ This was conspicuous in his drawings\ where the left!sided elements were drawn physically larger than those on the right\ when they should have been symmetrical[ When presented with symmetrical pat! terns\ he perceived the elements on the left as smaller than those on the right[ The patient died 16 months later\ and a post!mortem revealed destruction of the inferior occipital gyrus and the inferior part of the middle occipi! tal gyrus\ regions corresponding to the lower part of the lateral aspect of areas 07 and 08[ Laterally\ the lesion also a}ected the depths of the inferior temporal sulcus\
part of the middle temporal gyrus\ and the depths of the superior temporal sulcus[ The second patient perceived objects on the right as smaller than those on the left[ An MRI revealed a circumscribed left hemisphere lesion a}ecting the lower lateral aspect of areas 07 and 08 and the underlying white matter[ Further investigations are required to see whether the perceptual distortion observed in some patients with visuo!spatial neglect cor! responds to micropsia\ but as the areas involved in both conditions are adjacent\ this is highly feasible\ par! ticularly if subcortical damage could compromise both areas 07 and 08 of Brodmann[ All the patients in the current study had a right hemisphere lesion and left! sided neglect and\ perhaps crucially\ all had sub!cortical damage\ extensive in some cases\ in the immediate vicinity of the focal cortical damage that leads to micropsia[ The perceptual distortion described here appears to be a gross exaggeration of normal performance[ Our suggestion is that it could have either or both of two bases\ an attentional one and a genuine perceptual dis! tortion[ However\ in decribing the distortion we have throughout referred to it as a shrinkage on the left even though\ theoretically\ it could be an expansion on the right[ We have no direct evidence that bears on this possibility but in hemimicropsia the patient actually claims that things are unusually small contralaterally\ not large ipsilaterally[ Furthermore\ a right!sided expansion would have to be more prominent in the horizontal axis to explain the results\ and although such a change in horizontal extent might remain unnoticed by the patient within the neglected hemi_eld\ it should be noticed if it occurs in the non!neglected hemi_eld[ This did not occur[ A curious aspect of unilateral visuo!spatial neglect is that when studied by line bisection\ the characteristic rightward errors can be greater\ in angular terms\ in near space "within reaching distance# ð6Ł or in far space ð2Ł[ The pathological basis for this dissociation is unclear[ If the mis!bisection occurs because the patient experiences a subjective contraction of visual space on the left\ as suggested here and by others ð02Ł\ then the so!called dis! tance e}ect should be present in a matching task like that used here\ as well as for line bisection\ unless the distance e}ect arises for other reasons such as di}erent motor response programmes for near and far space[ We are now pursuing this possibility[ Acknowledgements This research was supported by the MRC IRC in Brain and Behaviour\ Oxford\ and by an Oxford McDonnell! Pew Studentship to L[ Irving!Bell[ References ð0Ł Adams RD\ Victor M[ Principles of neurology[ 3th ed[ New York] McGraw!Hill\ 0889]516Ð7[
L[ Irvin`!Bell et al[ : Neuropsycholo`ia 26 "0888# 808Ð814 ð1Ł Cohen L\ Gray F\ Meyrignac C\ Dehaene S\ Degos JD[ Selective de_cit of size perception] two cases of hemimicropsia[ Journal of Neurology\ Neurosurgery and Psychiatry 0883^46"0#]62Ð7[ ð2Ł Cowey A\ Small M\ Ellis S[ Left visuo!spatial neglect can be worse in far than in near space[ Neuropsychologia 0883^21"8#]0948Ð55[ ð3Ł Damasio H\ Damasio HR[ Lesion analysis in neuropsychology[ New York] Oxford University Press\ 0878[ ð4Ł Gainotti G\ Tiacci C[ The relationship between disorders of visual perception and unilateral spatial neglect[ Neuropsychologia 0860^8]340Ð7[ ð5Ł Halligan PW\ Marshall JC[ Perceptual cueing and perceptuo! motor compatability in visuo!spatial neglect] a single case study[ Cognitive Neuropsychology 0878^5"3#]312Ð24[ ð6Ł Halligan PW\ Marshall JC[ Left neglect for near but not far space in man[ Nature 0880^249]387Ð499[ ð7Ł Harvey M\ Milner AD\ Roberts RC[ An investigation of hemi! spatial neglect using the landmark task[ Brain and Cognition 0884^16]48Ð67[ ð8Ł Milner AD[ Animal models for the syndrome of spatial neglect[ In] Jeannerod M\ editor[ Neurophysiological and neuropsychological aspects of spatial neglect[ Amsterdam] Elsevier\ 0876]148Ð77[ ð09Ł Milner AD[ Neglect\ extinction\ and the cortical streams of visual
ð00Ł
ð01Ł ð02Ł
ð03Ł
ð04Ł
ð05Ł ð06Ł
814
processing[ In] Thier P\ Karnath H!O\ editors[ Parietal lobe con! tributions to orientation in 2D space[ Heidelberg] Springer!Verlag\ 0886]2Ð11[ Milner AD\ Brechmann M\ Pagliarini L[ To halve and to halve not] an analysis of line bisection judgements in normal subjects[ Neuropsychologia 0881^29]404Ð15[ Milner AD\ Harvey M[ Distortion of size perception in vis! uospatial neglect[ Current Biology 0884^4"0#]74Ð8[ Milner AD\ Harvey M\ Roberts RC\ Forster SV[ Line bisection errors in visual neglect] misguided action or size distortion< Neuro! psychologia 0882^20"0#]28Ð38[ Nichelli P\ Rinaldi M\ Cubelli R[ Selective spatial attention and length representation in normal subjects and in patients with uni! lateral spatial neglect[ Brain and Cognition 0878^8]46Ð69[ Reuter!Lorenz PA\ Posner MI[ Components of neglect from right hemisphere damage] an analysis of line bisection[ Neu! ropsychologia 0889^17"3#]216Ð22[ Riddoch MJ\ Humphries G[ The e}ect of cueing on unilateral neglect[ Neuropsychologia 0872^10]478Ð88[ Wilson B\ Cockburn J\ Halligan PW[ Behavioural Inattention Test[ Titch_eld\ Hampshire] Thames Valley Test Company[