- Email: [email protected]
Line Bisection in Visuo-Spatial Neglect: Disproof of a Conjecture
NOTE
LINE BISECTION IN VISUO-SPATIAL NEGLECT: DISPROOF OF A CONJECTURE Peter W. Halligan and John C. Marshall
(Rivermead Rehabilitation Centre, Oxford, and Neuropsychology Unit, University Department of Clinical Neurology, The Radcliffe Infirmary, Oxford)
INTRODUCTION
Patients with lesions to right parietal cortex, or to subcortical regions that are interconnected thereto, often manifest a striking tendency to 'neglect' stimuli in left hemispace (De Renzi, 1982; Heilman, Watson and Valenstein, 1985). One commonly employed clinical test for neglect is line bisection (Axenfeld, 1915). Requested to bisect a line in free vision, patients with left neglect will systematically displace their bisection to the right of true centre. This contrasts with the performance of control subjects who show a tendency to bisect horizontal lines slightly to the left of centre. The magnitude of the rightwards displacement in neglect increases linearly with the length of the stimulus line (Riddoch and Humphreys, 1983), and the proportional rate of increase may differ from patient to patient (Bisiach, Bulgarelli, Sterzi and Vallar, 1983). We have shown, in a single-case study, that the rate of increase can be such that, for lines smaller than those employed in any previous investigation, the effect crosses over objective centre (Halligan and Marshall, 1988). Consistent with a linear regression across all line lengths, there was thus a length (2") at which the patient (P.B.) showed a mean displacement of zero, and a smaller length (I") at which his bisections were consistently left displaced to a far greater extent than is characteristic of normal subjects. Over the range of line lengths employed (I" to 11"), a good approximation to the patient's mean performance is given by the formula D =
~-
C, where Dis the observed displacement from true centre, Lis the (objective)
length of the stimulus line, and Cis a constant with the empirical value (in this case) of 11mm. Using the formula given, the linear correlaticm between predicted and observed 'bisections' is r = + 0.99 (p<0.001). A simple processing model would give rise to the excellent fi~so obtained. In the standard procedure for the assessment of bisection performance, the stimulus line is positioned on the desk so that the objective midpoint lies in the sagittal midplane of the subject's trunk. Although the task is performed in free vision, suppose that the patient we have previously reported (Halligan and Marshall, 1988) cannot consistently direct his attention further to the left than the visual angle corresponding to a point on the stimulus sheet that is 22mm left of true centre. If visual fixation was determined by that point, the extent of the stimulus line to the right of fixation would project to the (intact) left hemisphere. Points to the left of fixation would either fall within the hemianopic field (many, but not all, patients with left neglect have a left homonymous hemianopia); or, if sensation was intact, would fail to be further processed by the (damaged) right hemisphere. In order to explain the crossover from right displacements to left displacements at short line lengths, one must further postulate that when the objective line does not extend as far as the attentional boundary, it is representationally completed to that boundary; such Cortex, (1989) 25, 517-521
sus
Peter W Halligan and John C. Marshall
'completion' must, ex hypothesi, over-ride the veridical detection of the left end of the stimulus line. There is independent, albeit controversial, evidence for such completion across hemianopic field deficits (Warrington, 1962; Sergent, 1988), and, more generally, for image generation by left hemisphere mechanisms (Farah, 1988). Whether the concept of a fixed boundary should be interpreted in strictly 'physical' terms, i.e. direction of gaze (Ishiai, Furukawa and Tsukagoshi, 1987), is debatable in the light of demonstrations of 'imaginal neglect' (Bisiach and Luzzatti, 1978). It is certainly possible that some more central representation determines the extent that can be 'seen' by the left hemisphere. Be that as it may, our model proposes that the bisection task in left neglect is undertaken by intact regions of the left hemisphere. The relevant mechanisms perform the task correctly, given the representations of the stimulus that are available to them. When these mechanisms bisect the represented line accurately, the result will be a quartering of the actual (or 'completed') line minus half the distance from true centre to the attentional boundary. Within this model, differences between patients must be accounted for by variability in the placement of the attentional boundary (that is, the empirical value that the constant in D =
~-
C takes in each patient). If, within an individual patient,
the requisite value of C differs between the different line lengths, our conjecture is falsified; the account would be trivialized by multiple post hoc adjustments for each length. For the model to hold (in the simple form given here), the equation for the regression of displacement on line length must include a multiplier that is (not significantly different from) + 0.25. The best-fitting regression equation for P.B.'s performance (Halligan and Marshall, 1988) is - 14.52 + (0.276 X line length). All measurements are in millimetres. Clearly, this calculation is consistent with the 'quartering' and 'attentional boundary' model (and accounts for 98% of the observed variance). Accordingly, we now inquire as to whether the model is indeed fully general with respect to the variability in extent of neglect from patient to patient.
MATERIALS AND METHODS
Subjects
Twenty control subjects, ten male and ten female, without known neurological disease were recruited. Their mean age was 43.9 (range 25-74), with a standard deviation of 13.2. All were right-handed. J.H. is a fifty-six year-old, right-handed man who sustained a right hemisphere stroke. CT scan showed an area of low density with some slight mass effect affecting the right caudate nucleus and putamen, including the anterior limb of the internal capsule. These changes are consistent with a small infarct. There was no hemiparesis or hemianopia. J.H. showed mild left neglect. On letter-cancellation from the Behavioural Inattention Test (Wilson, Cockburn and Halligan, 1987), he obtained a score of 29140; cut-off for normality is 32140. On star-cancellation, from the same test battery, a score of 49154 was gained; normal cut-off is 51 I 54. P.S. is a forty-nine year-old, right-handed woman who sustained a subarachnoid haemorrhage. CT scan showed subarachnoid blood with mild hydrocephalus. Angiograpy showed that she had an aneurysm at the bifurcation of the basilar artery. Fronto-temporal craniotomy and clipping of the basilar aneurysm was undertaken, but she afterwards developed left hemiplegia with dense left homonymous hemianopia. The only neuropsychological feature of note was florid left neglect. Assessed on the Behavioural Inattention Test (Wilson, Cockburn and Halligan, 1987), P.S. obtained a score of 37 I 146; the cut-off for normality is 1301146. This aggregate is derived from performance on: Line cancellation; letter-and star-cancellation; figure copying; line bisection; and representational drawing.
Line bisection in visuo-spatial neglect
519
Stimuli and Procedure Eleven horizontal black lines were individually drawn on sheets of white paper (280 x 298 mm). Each line was approximately 1mm wide. The lines varied in length from 1" (25 mm) to 11" (279 mm) in steps of 1". Each line was presented on a separate sheet in pseudo-random order, and placed on the desk so that the objective midpoint was positioned in the sagittal midplane of the subject's trunk. The lines themselves were likewise centred on the page, both horizontally and vertically. The subjects were instructed to mark the midpoint of each line with a fine pen or pencil, using the preferred (right) hand. Viewing distance was approximately 450 mm. Control subjects performed the task (11 lines) once each; the two patients performed the task ten times, each producing 110 transections, ten at each line length. Stimulus lengths were pseudo-randomized across all 110 presentations.
RESULTS
The data (mean displacements, and standard deviations, at all line lengths) are shown in Table I. It can be seen that mean performance of the normal controls is extremely accurate at all line lengths, although the previously reported tendency towards leftward displacement is confirmed. The best-fitting regression equation is + 0.06577 - (0.00434 X line length), and accounts for 43% of the variance. Variability increases as a function of length; there is a linear correlation of r = + 0.88 (p<0.001) between standard deviation and length of the line presented. By contrast, J.H. shows a progressively increasing rightwards displacement of his bisections from a (leftward) value of - 0.2 mm at length 25 mm to a (rightward) value of + 13.9 mm at 279 mm. The linear correlation between (objective line length and signed displacement is r = + 0.96 (p<0.001). The correlation between line length and standard deviation is r = + 0.84 (p
TABLE I
Magnitudes of Signed Displacements in mm ( + Standard Deviations) Line length 279 mrn/11" 254 mm/10" 228 mrn/9" 203 mm/8" 178 mrn/7" 152 mm/6" 127 mrn/5" 102 mrn/4" 77 mm/3" 51 mm/2" 25 mrn/1"
Control Subjects (N - 1.35 (4.36) - 1.30 (3.66) - 1.10 (2. 73) -0.40(3.10) -0.15 (2.85) -0.10 (3.85) - 1.35 (2.11) -0.05 (2.70) -0.35 (1.76) -0.40 (1.14) 0.00 (0.56)
= 20)
J.H.
P.S.
13.9 (5.8) 15.7 (3.3) 12.1 (3.2) 13.4 (5.9) 11.9 (3.5) 6.7 (3.4) + 5.8 (2.3) + 2.2 (1.7) + 2.6 (1.8) + 1.0 (1.1) 0.2(1.1)
+ 94.9 (11.8) +93.3 (12.1) +80.2 ( 6.1) + 70.7 ( 7.7) +58.5 ( 2.8) +48.8 ( 4.7) +40.3 ( 4.1) +26.4( 5.1) + 12.6 ( 2.9) + 0.7 ( 4.2) - 1.9 ( 4.5)
+ + + + + +
Peter W Halligan and John C. Marshall
520
11"(279mm)
-----------+--,----,-------,---
10" (254mm)
I
9" (228mm)
8" (203mm)
/
I II I
6" (152mm)
4" (102mm)
\
/
7" (178mm)
5" (127mm)
/
I
til I /
/
I
/
I
I
I
I
I
iL,· / /
!
3" {77mm)
'I
2" (51mm)
1" (25mm)
/
_j_
----PB -------- J1
------ PS
Fig. 1 -Mean displacements obtained from three neurological patients (P B, J H, and PS) requested to bisect horizontal lines presented in free vision.
from patient P.B. (Halligan and Marshall, 1988) included for ease of reference and comparison. DISCUSSION
With respect to our original conjecture, the datum at stake is the slope of the regression line in patients J.H. and P.S. The best-fitting regression equation for I .H.'s performance is - 2.45 + (0.067 X line length), a calculation that accounts for 93% of the observed variance. For P.S., the equation is - 15.92 + (0.42 per line length), accounting for 99% of the variance. In both cases, the slope of the regression is dramatically different from that predicted by the model. The multiplier is, in one case, far too large ( + 0.42) and, in the other, far too small ( + 0.067). We thus confirm the observation of Bisiach, Bulgarelli, Sterzi and Vallar (1983) that there are differences between neglect patients in the proportional rate of bisection displacements as a function of line length; and we have now shown that these differences are too great to be accounted for by a simple model in which a line's extension to the left of an attentional boundary is 'neglected' whilst the left hemisphere (accurately) bisects the rightward extent of the line. Are there, then (at least) two types of neglect, one manifested by P.B. (whose performance does give a close fit to the model) and another shown by J.H. and P.S. for which the underlying disorder is quite distinct? Alternatively, is there a model (as yet unknown) within which variation of a single parameter could suffice to interpret the full range of empirical variation reported here? ABSTRACT
We describe two patients with left neglect who show a linear relationship between the magnitude of their line bisection displacements and the veridical length of the line
Line bisection in visuo-spatial neglect
521
presented. For lines smaller than those typically employed in previous investigations, the patients' mean performance is fairly accurate; with even smaller lines, the direction of displacement crosses over true centre to give a leftward displacement ('right neglect'). In both cases, the slope of the regression line falsifies the conjecture that (intact) left hemisphere mechanisms (correctly) bisect the (subtotal or supertotal) line length that is available in right hemispace. Acknowledgements. This research was supported by the Chest, Heart, and Stroke Association (P.W.H.) and the Medical Research Council (J.C.M.). We are also grateful to Dr. Elaine Funnell for a number of helpful discussions. REFERENCES
AxENFELD, Th. Hemianopische Gesichtsfeldstorungen nach Schlidelschiissen. Klinische Monatsblatter fur Augenheilkunde, 55: 126-143, 1915. BISIACH, E., and LUZZATTI, C. Unilateral neglect of representational space. Cortex, 14: 129-133, 1978. BISIACH, E., BULGARELLI, C., STERZI, R., and VALLAR, G. Line bisection and cognitive plasticity of unilateral neglect of space. Brain and Cognition, 2: 32-38, 1983. DE RENZI, E. Disorders of Space Exploration and Cognition. New York: Wiley, 1982. FARAH, M.J. Is visual imagery really visual?. Overlooked evidence from neuropsychology. Psychological Review, 95: 307-317, 1988. HALLIGAN, P.W., and MARSHALL, J.C. How long is a piece of string? A study of line bisection in a case of visual neglect. Cortex, 24: 321-328, 1988. HEILMAN, K.M., WATSON, R.T., and VALENSTEIN, E. Neglect and related disorders. In K.M. Heilman and E. Valenstein (Eds.), Clinical Neuropsychology (2nd ed.). New York: Oxford University Press, 1985. ISHIAI, S., FuRUKAWA, T., and TsuKAGOSHI, H. Eye fixation patterns in homonymous hemianopia and unilateral visual neglect. Neuropsychologia, 25: 675-679, 1987. RronocH, M.J., and HUMPHREYS. G.W. The effect of cueing on unilateral neglect. Neuropsychologia, 21: 589-599, 1983. SERGENT, J. An investigation into perceptual completion in blind areas of the visual field. Brain, Ill: 347-373, 1988. WARRINGTON, E.K. The completion of visual forms across hemianopic field defects. Journal of Neurology, Neurosurgery and Psychiatry, 25: 208-218, 1962. WILSON, B., CoCKBURN, J., and HALLIGAN, P.W. Behavioural Inattention Test. Titchfield, Hants.: Thames Valley Test Company, 1987. Peter W. Halligan, Rivermead Rehabilitation Centre, Abingdon Road, Oxford OX! 4XD, U.K. John C. Marshall, Neuropsychology Unit, The Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, U.K.
LINE BISECTION IN VISUO-SPATIAL NEGLECT: DISPROOF OF A CONJECTURE Peter W. Halligan and John C. Marshall
(Rivermead Rehabilitation Centre, Oxford, and Neuropsychology Unit, University Department of Clinical Neurology, The Radcliffe Infirmary, Oxford)
INTRODUCTION
Patients with lesions to right parietal cortex, or to subcortical regions that are interconnected thereto, often manifest a striking tendency to 'neglect' stimuli in left hemispace (De Renzi, 1982; Heilman, Watson and Valenstein, 1985). One commonly employed clinical test for neglect is line bisection (Axenfeld, 1915). Requested to bisect a line in free vision, patients with left neglect will systematically displace their bisection to the right of true centre. This contrasts with the performance of control subjects who show a tendency to bisect horizontal lines slightly to the left of centre. The magnitude of the rightwards displacement in neglect increases linearly with the length of the stimulus line (Riddoch and Humphreys, 1983), and the proportional rate of increase may differ from patient to patient (Bisiach, Bulgarelli, Sterzi and Vallar, 1983). We have shown, in a single-case study, that the rate of increase can be such that, for lines smaller than those employed in any previous investigation, the effect crosses over objective centre (Halligan and Marshall, 1988). Consistent with a linear regression across all line lengths, there was thus a length (2") at which the patient (P.B.) showed a mean displacement of zero, and a smaller length (I") at which his bisections were consistently left displaced to a far greater extent than is characteristic of normal subjects. Over the range of line lengths employed (I" to 11"), a good approximation to the patient's mean performance is given by the formula D =
~-
C, where Dis the observed displacement from true centre, Lis the (objective)
length of the stimulus line, and Cis a constant with the empirical value (in this case) of 11mm. Using the formula given, the linear correlaticm between predicted and observed 'bisections' is r = + 0.99 (p<0.001). A simple processing model would give rise to the excellent fi~so obtained. In the standard procedure for the assessment of bisection performance, the stimulus line is positioned on the desk so that the objective midpoint lies in the sagittal midplane of the subject's trunk. Although the task is performed in free vision, suppose that the patient we have previously reported (Halligan and Marshall, 1988) cannot consistently direct his attention further to the left than the visual angle corresponding to a point on the stimulus sheet that is 22mm left of true centre. If visual fixation was determined by that point, the extent of the stimulus line to the right of fixation would project to the (intact) left hemisphere. Points to the left of fixation would either fall within the hemianopic field (many, but not all, patients with left neglect have a left homonymous hemianopia); or, if sensation was intact, would fail to be further processed by the (damaged) right hemisphere. In order to explain the crossover from right displacements to left displacements at short line lengths, one must further postulate that when the objective line does not extend as far as the attentional boundary, it is representationally completed to that boundary; such Cortex, (1989) 25, 517-521
sus
Peter W Halligan and John C. Marshall
'completion' must, ex hypothesi, over-ride the veridical detection of the left end of the stimulus line. There is independent, albeit controversial, evidence for such completion across hemianopic field deficits (Warrington, 1962; Sergent, 1988), and, more generally, for image generation by left hemisphere mechanisms (Farah, 1988). Whether the concept of a fixed boundary should be interpreted in strictly 'physical' terms, i.e. direction of gaze (Ishiai, Furukawa and Tsukagoshi, 1987), is debatable in the light of demonstrations of 'imaginal neglect' (Bisiach and Luzzatti, 1978). It is certainly possible that some more central representation determines the extent that can be 'seen' by the left hemisphere. Be that as it may, our model proposes that the bisection task in left neglect is undertaken by intact regions of the left hemisphere. The relevant mechanisms perform the task correctly, given the representations of the stimulus that are available to them. When these mechanisms bisect the represented line accurately, the result will be a quartering of the actual (or 'completed') line minus half the distance from true centre to the attentional boundary. Within this model, differences between patients must be accounted for by variability in the placement of the attentional boundary (that is, the empirical value that the constant in D =
~-
C takes in each patient). If, within an individual patient,
the requisite value of C differs between the different line lengths, our conjecture is falsified; the account would be trivialized by multiple post hoc adjustments for each length. For the model to hold (in the simple form given here), the equation for the regression of displacement on line length must include a multiplier that is (not significantly different from) + 0.25. The best-fitting regression equation for P.B.'s performance (Halligan and Marshall, 1988) is - 14.52 + (0.276 X line length). All measurements are in millimetres. Clearly, this calculation is consistent with the 'quartering' and 'attentional boundary' model (and accounts for 98% of the observed variance). Accordingly, we now inquire as to whether the model is indeed fully general with respect to the variability in extent of neglect from patient to patient.
MATERIALS AND METHODS
Subjects
Twenty control subjects, ten male and ten female, without known neurological disease were recruited. Their mean age was 43.9 (range 25-74), with a standard deviation of 13.2. All were right-handed. J.H. is a fifty-six year-old, right-handed man who sustained a right hemisphere stroke. CT scan showed an area of low density with some slight mass effect affecting the right caudate nucleus and putamen, including the anterior limb of the internal capsule. These changes are consistent with a small infarct. There was no hemiparesis or hemianopia. J.H. showed mild left neglect. On letter-cancellation from the Behavioural Inattention Test (Wilson, Cockburn and Halligan, 1987), he obtained a score of 29140; cut-off for normality is 32140. On star-cancellation, from the same test battery, a score of 49154 was gained; normal cut-off is 51 I 54. P.S. is a forty-nine year-old, right-handed woman who sustained a subarachnoid haemorrhage. CT scan showed subarachnoid blood with mild hydrocephalus. Angiograpy showed that she had an aneurysm at the bifurcation of the basilar artery. Fronto-temporal craniotomy and clipping of the basilar aneurysm was undertaken, but she afterwards developed left hemiplegia with dense left homonymous hemianopia. The only neuropsychological feature of note was florid left neglect. Assessed on the Behavioural Inattention Test (Wilson, Cockburn and Halligan, 1987), P.S. obtained a score of 37 I 146; the cut-off for normality is 1301146. This aggregate is derived from performance on: Line cancellation; letter-and star-cancellation; figure copying; line bisection; and representational drawing.
Line bisection in visuo-spatial neglect
519
Stimuli and Procedure Eleven horizontal black lines were individually drawn on sheets of white paper (280 x 298 mm). Each line was approximately 1mm wide. The lines varied in length from 1" (25 mm) to 11" (279 mm) in steps of 1". Each line was presented on a separate sheet in pseudo-random order, and placed on the desk so that the objective midpoint was positioned in the sagittal midplane of the subject's trunk. The lines themselves were likewise centred on the page, both horizontally and vertically. The subjects were instructed to mark the midpoint of each line with a fine pen or pencil, using the preferred (right) hand. Viewing distance was approximately 450 mm. Control subjects performed the task (11 lines) once each; the two patients performed the task ten times, each producing 110 transections, ten at each line length. Stimulus lengths were pseudo-randomized across all 110 presentations.
RESULTS
The data (mean displacements, and standard deviations, at all line lengths) are shown in Table I. It can be seen that mean performance of the normal controls is extremely accurate at all line lengths, although the previously reported tendency towards leftward displacement is confirmed. The best-fitting regression equation is + 0.06577 - (0.00434 X line length), and accounts for 43% of the variance. Variability increases as a function of length; there is a linear correlation of r = + 0.88 (p<0.001) between standard deviation and length of the line presented. By contrast, J.H. shows a progressively increasing rightwards displacement of his bisections from a (leftward) value of - 0.2 mm at length 25 mm to a (rightward) value of + 13.9 mm at 279 mm. The linear correlation between (objective line length and signed displacement is r = + 0.96 (p<0.001). The correlation between line length and standard deviation is r = + 0.84 (p
TABLE I
Magnitudes of Signed Displacements in mm ( + Standard Deviations) Line length 279 mrn/11" 254 mm/10" 228 mrn/9" 203 mm/8" 178 mrn/7" 152 mm/6" 127 mrn/5" 102 mrn/4" 77 mm/3" 51 mm/2" 25 mrn/1"
Control Subjects (N - 1.35 (4.36) - 1.30 (3.66) - 1.10 (2. 73) -0.40(3.10) -0.15 (2.85) -0.10 (3.85) - 1.35 (2.11) -0.05 (2.70) -0.35 (1.76) -0.40 (1.14) 0.00 (0.56)
= 20)
J.H.
P.S.
13.9 (5.8) 15.7 (3.3) 12.1 (3.2) 13.4 (5.9) 11.9 (3.5) 6.7 (3.4) + 5.8 (2.3) + 2.2 (1.7) + 2.6 (1.8) + 1.0 (1.1) 0.2(1.1)
+ 94.9 (11.8) +93.3 (12.1) +80.2 ( 6.1) + 70.7 ( 7.7) +58.5 ( 2.8) +48.8 ( 4.7) +40.3 ( 4.1) +26.4( 5.1) + 12.6 ( 2.9) + 0.7 ( 4.2) - 1.9 ( 4.5)
+ + + + + +
Peter W Halligan and John C. Marshall
520
11"(279mm)
-----------+--,----,-------,---
10" (254mm)
I
9" (228mm)
8" (203mm)
/
I II I
6" (152mm)
4" (102mm)
\
/
7" (178mm)
5" (127mm)
/
I
til I /
/
I
/
I
I
I
I
I
iL,· / /
!
3" {77mm)
'I
2" (51mm)
1" (25mm)
/
_j_
----PB -------- J1
------ PS
Fig. 1 -Mean displacements obtained from three neurological patients (P B, J H, and PS) requested to bisect horizontal lines presented in free vision.
from patient P.B. (Halligan and Marshall, 1988) included for ease of reference and comparison. DISCUSSION
With respect to our original conjecture, the datum at stake is the slope of the regression line in patients J.H. and P.S. The best-fitting regression equation for I .H.'s performance is - 2.45 + (0.067 X line length), a calculation that accounts for 93% of the observed variance. For P.S., the equation is - 15.92 + (0.42 per line length), accounting for 99% of the variance. In both cases, the slope of the regression is dramatically different from that predicted by the model. The multiplier is, in one case, far too large ( + 0.42) and, in the other, far too small ( + 0.067). We thus confirm the observation of Bisiach, Bulgarelli, Sterzi and Vallar (1983) that there are differences between neglect patients in the proportional rate of bisection displacements as a function of line length; and we have now shown that these differences are too great to be accounted for by a simple model in which a line's extension to the left of an attentional boundary is 'neglected' whilst the left hemisphere (accurately) bisects the rightward extent of the line. Are there, then (at least) two types of neglect, one manifested by P.B. (whose performance does give a close fit to the model) and another shown by J.H. and P.S. for which the underlying disorder is quite distinct? Alternatively, is there a model (as yet unknown) within which variation of a single parameter could suffice to interpret the full range of empirical variation reported here? ABSTRACT
We describe two patients with left neglect who show a linear relationship between the magnitude of their line bisection displacements and the veridical length of the line
Line bisection in visuo-spatial neglect
521
presented. For lines smaller than those typically employed in previous investigations, the patients' mean performance is fairly accurate; with even smaller lines, the direction of displacement crosses over true centre to give a leftward displacement ('right neglect'). In both cases, the slope of the regression line falsifies the conjecture that (intact) left hemisphere mechanisms (correctly) bisect the (subtotal or supertotal) line length that is available in right hemispace. Acknowledgements. This research was supported by the Chest, Heart, and Stroke Association (P.W.H.) and the Medical Research Council (J.C.M.). We are also grateful to Dr. Elaine Funnell for a number of helpful discussions. REFERENCES
AxENFELD, Th. Hemianopische Gesichtsfeldstorungen nach Schlidelschiissen. Klinische Monatsblatter fur Augenheilkunde, 55: 126-143, 1915. BISIACH, E., and LUZZATTI, C. Unilateral neglect of representational space. Cortex, 14: 129-133, 1978. BISIACH, E., BULGARELLI, C., STERZI, R., and VALLAR, G. Line bisection and cognitive plasticity of unilateral neglect of space. Brain and Cognition, 2: 32-38, 1983. DE RENZI, E. Disorders of Space Exploration and Cognition. New York: Wiley, 1982. FARAH, M.J. Is visual imagery really visual?. Overlooked evidence from neuropsychology. Psychological Review, 95: 307-317, 1988. HALLIGAN, P.W., and MARSHALL, J.C. How long is a piece of string? A study of line bisection in a case of visual neglect. Cortex, 24: 321-328, 1988. HEILMAN, K.M., WATSON, R.T., and VALENSTEIN, E. Neglect and related disorders. In K.M. Heilman and E. Valenstein (Eds.), Clinical Neuropsychology (2nd ed.). New York: Oxford University Press, 1985. ISHIAI, S., FuRUKAWA, T., and TsuKAGOSHI, H. Eye fixation patterns in homonymous hemianopia and unilateral visual neglect. Neuropsychologia, 25: 675-679, 1987. RronocH, M.J., and HUMPHREYS. G.W. The effect of cueing on unilateral neglect. Neuropsychologia, 21: 589-599, 1983. SERGENT, J. An investigation into perceptual completion in blind areas of the visual field. Brain, Ill: 347-373, 1988. WARRINGTON, E.K. The completion of visual forms across hemianopic field defects. Journal of Neurology, Neurosurgery and Psychiatry, 25: 208-218, 1962. WILSON, B., CoCKBURN, J., and HALLIGAN, P.W. Behavioural Inattention Test. Titchfield, Hants.: Thames Valley Test Company, 1987. Peter W. Halligan, Rivermead Rehabilitation Centre, Abingdon Road, Oxford OX! 4XD, U.K. John C. Marshall, Neuropsychology Unit, The Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, U.K.