NOTE
HOMING IN ON NEGLECT: A CASE STUDY OF VISUAL
SEARCH
Peter W. Halligan and John C. Marshall
(Rivennead Rehabilitation Centre, Abingdon Road, Oxford, and University Department
of Clinical Neurology, Radcliffe Infinnary, Woodstock Road, Oxford, UK)

INTRODUCTION

It is well-known that the effects of left-sided cueing in patients with left neglect are
both variable and transitory. With appropriate instructions and a sufficiently strong cue,
it is usually possible to direct the patient's "attention" to stimuli on the left. In some
cueing-paradigms one can be fairly certain that the patient has "seen" the full extent of
the stimulus-array, but even in these circumstances the patient continues to "neglect" stimuli (or parts thereof) that have apparently been "perceived" (in some sense of the tenn).
The phenomenon can be observed most clearly in studies of line bisection. In some
investigations where the patient is required to report a letter placed at the left end of a
stimulus line prior to bisection, no improvement over uncued trials is found (Heilman
and Valenstein, 1979). In other studies, such cueing reduces but does not eliminate left
neglect (Riddoch and Humphreys, 1983; Nichelli, Rinaldi and Cubelli, 1989).
A similar effect may be found on cancellation tasks, although we are not aware of
any study that has explicitly investigated the issue. Nonetheless, in the standard deployment of the Albert Test, the examiner is first required to draw over each of the 40
stimulus lines with a red pencil. The method thus calls "the subject's attention to all the
lines and also explicitly (shows) the limits of the test" (Albert, 1973). In a modification
employed by Fullerton, McSherry and Stout (1986), some of the lines are pointed out to
the patient, "including those to extreme right and extreme left." Despite these manipulations, the patient with severe left neglect wi\l still fail to cancel stimuli on the left of
the stimulus page.
We now report the first systematic exploration of how much a patient with left neglect can "know" about the overall dimensions of a cancellation test whilst "neglecting"
a substantial number of the smaller targets that he is required to cross out. The question
at issue is whether a particular fonn of "panoramic" cueing to the full extent of a stimulus display has any effect upon the direction of attention to items within that display .
CASE HISTORY

R.B., a 50 year-old, left handed man (with right handed parents), was admitted to
hospital on 16 December 1991, following a right hemisphere stroke. CT scan, on 17
December 1991, revealed a large right middle cerebral artery infarct (Figure 1). The infarct involved primarily the right parietal cortex. Admitted to Rivennead on 16 March,
1992, R.B.'s major problems included left hemiplegia, left homonymous hemianopia, left
sensory loss, and gross left vi suo-spatial neglect.
Although well oriented in time and place, R.B. had little insight into his neglect. On
the six conventional subtests of the Behavioural Inattention Test (BIT} (Halligan, Cockburn and Wilson, 1991), R.B.'s score of 25/146 indicated severe left sided neglect. On
the behavioural tests he failed to read the left side of telephone numbers and numbers
located on the left side of a clockface. He also neglected the left side of words and
Cortex, (1993) 29,167-174

168

P. W. Halligan and J.e. Marshall

Fig. 1 - CT scan on 17 December
1991, showing a large right middle
cerebral artery infarct (right is on the
left of the image).

sentences when reading or copying text. Visual acuity was 6/9 in both eyes (glasses
corrected). Saccades were slightly hypometric and there was some difficulty in maintaining steady fixation.

EXPERIMENT 1

Materials and Methods

This experiment has 4 separate steps. Each step included the requirement that the
patient cross out (cancel) all the targets of the line crossing test from the BIT (Halligan
et aI., 1991). Head and eye movements were in no way constrained, but moving the
stimulus sheet was not permitted. The targets consisted of 40 black lines of different
orientations arrayed across an A4 page (279 X 210 mm); the stimulus sheet was centred
on the sagittal midplane of the patient's head and trunk. The target lines (25 mm long
by 1 mm wide) are distributed across the page in such a fashion that they can be scored
as 7 "columns" from left to right. The central column (column 4) is composed of 4
target lines; all other columns contain 6 lines. The patient was seated at a comfortable
distance from the test sheet displayed on the desk top. No knowledge of results was
provided to the patient and no time pressure was imposed. The patient stopped when he
was satisfied with his own performance. The (unaffected) right hand was always used
and each step began with the right hand in its natural position to the right of the stimulus
page. Unlike the standardised BIT condition of this task, the patient's attention was never
drawn to the central columns. The experiment (all 4 steps) was repeated three times.
The first step (standard condition) required the patient to cross out all the target lines
he could see on the page. The second step was the same as the first except that immediately prior to the start of the task, the patient was asked to mark each comer of
the stimulus page with a cross. The third step was the same as the first except that this
time the patient was asked to indicate the comers of the stimulus page after he had
completed crossing out all the lines he could see on the page. The fourth step was identical to the first step. If the two intermediary steps had any effect, this fourth step would

Homing in on neglect

169

TABLE I

Experiment 1: Number of Lines Crossed Out: Mean and Standard Deviation
Standard
Cross out lines

Mark 4 corners/
Cross out lines

Cross out lines/
Mark 4 corners

Standard
Cross out lines

Left

Right

Left

Right

Left

Right

Left

Right

0
(0)

13.0

0
(0)

12.3
(2.5)

0
(0)

11.3

0
(0)

10.3
(1.2)

(1.7)

(1.2)

access whether or not there was a carry-over effect on performance. After this experiment
had been repeated 3 times, the patient was required to indicate the size of the stimulus
page on which the target lines had been displayed. Blank sheets of A4 (279 X 210 mm)
and AS (148 X 210 mm) were presented one above the other and the patient was asked
to indicate the sheet that was the same size as the stimulus page. The toplbottom position
of the two sheets was randomly counterbalanced throughout 6 trials. Both stimulus sheets
were centred near the patient's sagittal midplane and spatially aligned so that .the right
edges were justified.

Results
Omission scores from the 4 presentations of line crossing were calculated. In each
of the four conditions R.B. failed to cross out all the target lines located on the left side
of the stimulus page. In steps 2 and 3, however, R.B. showed no difficulty in locating
and marking the 4 comers of the stimulus page. The results (means of the three trials)
are shown in Table 1. It is clear that R.B.'s apparent awareness of the spatial extent of
the A4 stimulus page made no difference to his cancellation of the targets contained
within the area whose comers he had accurately indicated. For the task requiring judgment of the stimulus page size, R.B. always chose the correct (A4) sheet (6/6).

EXPERIMENT 2

Materials and Methods
This experiment was similar to Experiment 1. The only difference was the size of
the page on which the stimuli lines were displayed. The same spatial array of lines was
used but transposed (centrally) onto a larger A3 (422 X 298 mm) page. The first step
required the patient to cross out all the target lines as before. The second step was similar to the first except that immediately prior to the start of the task, the patient was
asked to mark the comers of the stimulus sheet. The third step was identical to the first
step. The fourth step required the patient to mark the extent of the spatial display of
target lines (rather than the comers of the stimulus page) before cancelling the lines. The
fifth step was identical to the first step. Finally, the sixth step required the patient to
judge: (1) the size of the stimulus page on which the target lines were positioned; (2)
the size of the space within which the target lines were contained. For this step, blank
A3 (422X298 mm), A4 (279X21O mm) and AS (148X21O mm) sheets were presented
one above the other and the patient was asked to indicate the stimulus sheet that corresponded to the two judgments required. All three stimulus sheets were centred near the
patient's sagittal midplane and spatially aligned so that the right edges were justified. The
vertical order of the three sheets was randomly counterbalanced over five trials for each
type of judgement.

P. W. Halligan and J.e. Marshall

170
Fig. 2 - An exampLe of
R.B. 's peiformance on step
four of Experiment 2. The
patient indicated (with numbers) the extent of the spatial array of line targets that
he couLd see on the A3 page.
The array only is shown, not
the full extent of the (A3)
page.

dYY

---r
~
\I

v

-/

"<
--;---~

./

G~

~ Q

----- -----------

--

-----

~

~

y

J- --I-- ~

\
/

~

/

~@

~

~

/

~

--

~

-----

Results

As in the first experiment, R.B. always correctly marked the comers of the stimulus
page. He accordingly seems not to show the type of directional hypokinesia (or hypometria) reported in Bottini, Sterzi and Vallar (1992). Despite this, R.B. continued to neglect some 70% of the targets on the stimulus sheet (Table II). Requested in step four to
mark the extent of the spatial array of target lines, R.B . marked the spatial array as
smaller than it was. Figure 2 shows R.B.'s performance on step four. The patient indicated (with numbers) that the extent of the spatial array of line targets constituted the
three rightmost columns of the spatial array. Even within this restricted space, however,
there are omissions on two of the three columns that lie within the extent of the spatial
array that the patient has marked. When requested (as in Experiment 1) to indicate the
size of the stimulus page on which the spatial array of target line had been presented,
R.B. consistently (SIS) chose the correct (A3) page. However, when requested to indicate
the size of the spatial array of target lines from the three page sizes (A3, A4 and AS)
R.B. chose the smaller (AS) page on 4/5 trials. On one trial, he chose the correct A4
page.
EXPERIMENT 3

Materials and Methods

A new line crossing task was devised. 42 target lines were presented on an A4 page.
The targets were pseudo-randomly distributed but in such a fashion that they could be
TABLE II

Experiment 2: Number of Lines Crossed Out
Standard
Cross out lines

Mark 4 comers/
Cross out lines

Standard
Cross out lines

Mark edges of
display/Cross out
lines

Standard
Cross out lines

Left

Right

Left

Right

Left

Right

Left

Right

Left

Right

o

9

o

12

o

12

o

15

o

13

171

Homing in on neglect

scored as seven "columns" of six target lines each. Half the target lines were red and
the other half were black. The red and black lines were pseudo-randomly placed so that
there were equal numbers on the left and on the right; and to the top and bottom of
the page. This experiment has six separate steps. The first step simply required the patient
to cross out all the target lines on the page irrespective of colour. In the second step in
order to establish that RB. could see the total number of target lines on the page he
was asked to state the colour of each of the 42 lines when the examiner systematically
pointed each one out (from the rightmost to the leftmost column). R.B. had no difficulty
with this task and was very quick to name each colour correctly. Having traversed the
total spatial extent of the target array, · the patient was requested as before to cancel all
the target lines that he could see on the page. Step three then repeats the (un cued) first
step. Step four replicated step two (naming the colours and then crossing out the targets).
Step five was identical to the first step (uncued). Finally, step six required the patient to
indicate first the size of the page on which the target lines were placed and second the
size of the spatial array of lines contained on that page. The page sizes presented were
the same as in Experiment 1.
Results

The purpose of Experiment 3 was again to ascertain whether after RB. had explicitly
covered (and made an overt response to) the full extent of the spatial array of target
lines, he would subsequently detect and cross out more of the target lines. As in the
previous experiments R.B. could clearly respond to the full extent of the stimulus page
when cued to do so. He had no difficulty in naming correctly the colour of the lines
irrespective of spatial location when they were pointed out to him. However, the cancellation results (Table III) show that, after this cueing, RB. persisted in neglecting over
50% of the targets' located on his left.
When requested (as in the previous experiments) to indicate the size of the stimulus
page on which the black and red target lines had been presented, R.B. consistently chose
the correct (A4) sheet (3/3). Requested to indicate the size of the spatial array contained
on the stimulus page, RB. consistently chose the incorrect (AS) sheet (3/3).
TABLE III

Experiment 3: Number of Lines Crossed Out

Standard
Cross out lines

Name colours/
Cross out lines

Standard
Cross out lines

Name colours/
Cross out lines

Standard
Cross out lines

Left

Right

Left

Right

Left

Right

Left

Right

Left

Right

o

II

o

22

o

17

o

15

o

II

DISCUSSION

Experiment 1 shows that RB. can correctly mark the full extent of the A4 stimulus
page both before (step 2) and after (step 3) he attempts to cancel the target lines. His
manifest ability to do so, however, has little or no effect upon his inadequate performance on the cancellation task itself. RB. also "remembers" the size of the target display
(step 5) when the display covers the full extent of the stimulus page.
Experiment 2 shows that R.B. can correctly mark the full extent of a larger A3
(422 X 298 mm) stimulus page, and confirms that so doing has little or no effect upon
cancellation performance (compare the results of steps 1, 2 and 3). But if required to
mark the comers of the target display (when that display is smaller than the stimulus
page), R.B. "neglects" a substantial part of the left extent of the target display in both

172

P. W. Halligan and J.e. Marshall

the corner marking task itself and the subsequent cancellation task. He can "remember"
the size (A3) of the stimulus page correctly (step 6), but apparently misremembers the
size of the target display (A4) as the smaller size A3. This latter result is consistent with
the size that R.B. marked as the target display when he viewed it in step 4.
Experiment 3 confirms that R.B . can perceive all the targets scattered over an A4
page: when they are individually pointed out to him he correctly names the colour of
each line. As in the previous experiments, however, he continued to "neglect" over 50%
of the target lines when required to cross them out immediately after having named their
colour. In the size judgement task (step 6), R.B . correctly "remembers" the size of the
stimulus page (A4), as he had done in Experiments I and 2, but misremembers the size
of the target display as smaller than it was (consistent with the results of Experiment 2).
The data-pattern, then, is remarkably reliable and consistent across the three experiments. The interpretation of this pattern should, we propose, proceed along the following
lines.
In one current metaphor, visual attention has been likened to a "zoom lens" (Eriksen
and St. James, 1986; Humphrey and Bruce, 1989). Such a system can be set to cover
a wide range of field with poor resolution of detail, or a constricted field with good
resolution (and all intervening ranges) . When R.B. is "looking for" the boundaries of one
(fairly) large figure (an A3 or A4 page), he has no difficulty in detecting (and making
a manual response to) the corners of that (relatively) "wide range" stimulus. There is no
obvious perceptual or "hypokinetic" deficit in left space. Neither is there any gross memory failure for the size of that space.
However, when the task demands that attention is focussed onto a smaller area (in
order to resolve and cancel the individual line targets), the "lens" has a systematic bias
to the right. This bias leads to failure to cancel left sided targets in all three experiments;
and failure to mark correctly the top and bottom left corners of the target display in
Experiment 2. Attention is initially weakly distribut\'!d over the largest (relevant) field;
when any "focussing in" occurs, the rightward bias becomes manifest, and affects both
corner marking and target cancellation. The "remembered" target field (Experiment 2) is
then the extent of the narrowed and (right-) biased field. This latter effect still obtains
even when objectively the spatial extent of the stimulus page and the target display are
(almost) identical (Experiment 3). This result thus goes beyond the well established phenomenon that neglect gets worse with increasing demands on visual selective attention
(Kaplan, Verfaellie, Meadows et aI. , 1991).
Accordingly, these effects provide a theoretical link between unilateral neglect and
simultanagnosia (Luria, 1959). As Bisiach and Rusconi (1990) and Young, Hellawell and
Welch (1992) report, patients with left neglect may correctly trace the entirety of a line
drawing of one object and thereby trace through another object situated on the left side
of the first object without consciously "noticing" the presence of the second object. As
Balint (1907) observed, a mild degree of such "simultanagnosia" is perfectly normal:
"It is a well-known phenomenon that we do not notice anything happening in our
surroundings while being absorbed in the inspection of something; focussing our attention
on a certain object may happen to such an extent that we cannot perceive other objects
placed in the peripheral parts of our visual field ... " (Translation by Husain and Stein,
1988).

Gross clinical simultanagnosia (usually seen after bilateral occipito-parietal damage)
differs from this normal constriction of attention in that one of two "objects" can be
"neglected" even when they overlap and are "seen" in foveal vision. The report by Luria
(1959) of a patient who could perceive at anyone instant only one of the two triangles
that make up a Star-of-David is a convincing demonstration of pathological simultanagnosia in the above sense of the term.
Our current conjecture is that it may be profitable to think of unilateral neglect as
a form of lateralized simultanagnosia for objects represented at a "small" spatial scale
(Farah, 1990). R.B. can see the global but not the local features of left space (Robertson
and Lamb, 1991) unless, as in Experiment 3, his attention is specifically drawn to them.
The patients reported by Bisiach and Rusconi (1990) can see the house but not the fire

Homing in on neglect

173

in left space. As Humphreys and Bruce (1989) point out, the zoom lens metaphor captures " ... the way we can choose to see either a forest or leaves at different times". It
seems to us that a zoom lens with a rightward bias captures at least one important aspect
of the behaviours that have gone under the clinical label of left vi suo-spatial neglect.
Consider, for example, neglect dyslexia (Ellis, Flude and Young, 1987). The most
frequent error-type preserves the overall length of the word (measured in letters) but letter
substitutions occur on the left (e.g. FACADE - "arcade"). Similarly, when drawing or copying a clock-face, the circle is usually complete but the numbers on the left may be
missing. Line bisections are grossly skewed to the right but the centre of a square or
circle may be correctly estimated (Halligan and Marshall, 1991). In all these examples,
there is no "neglect" when the lens is set for a wide range of field; "left neglect" occurs
when the field is narrowed in order to resolve a "small" stimulus (e.g. the end-point of
a line).
In "simultanagnosia" as traditionally understood, "higher" levels of perceptual analysis
"select" one object and block the simultaneous perception of another. This is reminiscent
of "extinction", but without the lateralized aspect thereof. In neglect, the whole can be
perceived but "higher" levels of perceptual analysis can only "select" a lateralized subpart
of the whole for processing in focal attention. Coslett and Saffran (1991) note that "the
most puzzling aspect of the performance of patients with simultanagnosia" is this:
"When confronted with a complex visual array these patients often report 'seeing'
only a single item, yet when confronted with a large depiction of a single 'object' that
is co-existensive with the multi-item array, such as a drawing of an elephant, subjects
generally report seeing the entire object rather than components of the object (e.g., an
ear, trunk, etc.) despite the fact that many of the components may be in themselves
'objects'?".
A similar (but not identical) puzzle may be one aspect of neglect. Our patient R.B.
can "see" the whole array but only a lateralized subportion of the "objects" that make
up the array.
ABSTRACT
We report three studies of visual search in a patient with left neglect after a right
parietal infarct. Manipulations whereby the patient's attention is explicitly drawn to the
spatial extent of the page on which the stimuli for cancellation are displayed have little
or no effect on the number of targets cancelled. The results are consistent with a "zoom
lens" model of attentional deployment. "Neglect" on search tasks seems to reflect a systematic bias that emerges when global attention is constricted to permit analysis of local
features.
Acknowledgements. This work was supported by the Medical Research Council and
Remedi.

REFERENCES
ALBERT, M.L. A simple test of visual neglect. Neurology, 23: 658-644, 1973.
BALINT, R. A nezes lelki benulasa optikai ataxia, a figyelem terbeti zavara. Orvosi Hetilap, 1: 209236, 1907.
BISIACH, E., and RUSCONI, M.L. Breakdown of perceptual awareness in unilateral neglect. Cortex,
26: 643-649, 1990.
BOTIINI, G., STERZI R., and VALLAR, G. Directional hypokinesia in spatial hemineglect; a case
study. Journal of Neurology, Neurosurgery and Psychiatry, 55: 562-565, 1992.
eOSLETI, H.B., and SAFFRAN E. Simultanagnosia: To see but not two see. Brain, 114: 1523-1545,
1991.

174

P. W.Halligan and J.e. Marshall

ELLIS A.W., FLUDE, B.M., and YOUNG, A.W. "Neglect Dyslexia" and the early visual processing
of letters in words and nonwords. Cognitive NeuropsychoLogy, 4: 439-464, 1987.
ERIKSEN, C.W., and ST. JAMES, J.D. Visual attention within and around the field of focal attention:
a zoom lens model. Perception and Psychophysics, 40: 225-240, 1986.
FARAH, M.1. VisuaL Agnosia. Cambridge, Mass.: MIT Press, 1990.
FULLERTON, K.1., MCSHERRY, D., and STOUT, R.W. Albert's Test: a neglected test of perceptual
neglect. The Lancet, 1: 430-432, 1986.
HALLIGAN, P.W., COCKBURN, J., and WILSON, B. The behavioural assessment of visual neglect.
NeuropsychoLogicaL Rehabilitation, 1: 5-32, 1991.
HALLIGAN, P.W., and MARSHALL, J.e. Figural modulation of visuospatial neglect: A case study.
Neuropsychologia, 29: 619-628, 1991.
HEILMAN, K.M., and V ALENSTEIN, E. Mechanisms underlying hemispatial neglect. Annals of NeuroLogy, 5: 166-170, 1979.
HUMPHREYS, G.W., and BRUCE, V. VisuaL Cognition. London: Lawrence Erlbaum, 1989.
HUSAIN, M., and STEIN, J. Rezso Balint and his most celebrated case. Archives of NeuroLogy, 45:
89-93, 1988.
KAPLAN, R.F., VERFAELLlE, M., MEADOWS, M.-E., CAPLAN, L.R., PESSIN, M.S., and DE WITT, D.
Changing attentional demands in left hemispatial neglect. Archives of NeuroLogy, 48, 1263-1266,
1991.
LURIA, A.R. Disorders of "simultaneous perception" in a case of bilateral occipito-parietal brain
injury. Brain, 82: 437-449, 1959.
NICHELLI, P., RINALDI, M., and CUBELLI, R. Selective spatial attention and length representation in
normal subjects and in patients with unilateral spatial neglect. Brain and Cognition, 9: 57-70,
1989.
RIDDOCH, M.1., and HUMPHREYS, G.W. The effects of cueing on unilateral neglect. Neuropsychologia, 21: 589-599, 1983.
ROBERTSON, L.C., and LAMB, M.R. Neuropsychological contributions to part-whole organization.
Cognitive PsychoLogy, 23: 299-332, 1991.
YOUNG, A.W., HELLAWELL, D.1., and WELCH, J. Neglect and visual recognition. Brain, 115: 5)71 , 1992.
P.W. Halligan, Rivennead Rehabilitation Centre. Abingdon Road. Oxford OXI 4XD. U.K.