Brain (1990) 113, 191-205

PERIPERSONAL AND VERTICAL NEGLECT
by PAUL A. SHELTON, DAWN BOWERS
and KENNETH M. HEILMAN
{From the Department of Neurology, Veterans Administration Medical Center and University of
Florida College of Medicine, Gainesville, Florida, USA)

SUMMARY

INTRODUCTION

An individual with neglect fails to respond to novel or meaningful stimuli despite the
preservation of elementary sensory and motor functon (Heilman et al., 1985ft). Neglect
has been attributed to deficits both in attentional and intentional (motor preparatory)
processes. The behavioural manifestations of neglect are usually most profound in
extrapersonal and body space contralateral to hemisphere lesions, most often left
hemispace as defined by the midsagittal plane. Head and body midline are important
determinants of the neglected spatial field, and the visual half-field has relatively less
influence (Bowers etal., 1981; Heilman et al., 1987).
Several observations, however, suggest that neglect may occur in other fields of
extrapersonal space and therefore that attentional-intentional processes may be organized
in a more complex manner. First, many subjects with left hemispatial neglect fail to
cancel most targets in the left lower quadrant (Morris et al., 1985). Secondly, extinction and neglect of stimuli in the lower visual fields have recently been reported in man
(Rapcsak et al., 1988). Finally, neglect of pericutaneous stimuli has been reported in
monkeys (Rizzolatti et al., 1983). We recently evaluated a man for multimodal
neglect of upper space and distant extrapersonal space and compared him with 12 control
subjects.
Correspondence to: Dr Paul A. Shelton, Section of Neurology. University of Manitoba, Clinical Neurophysiology
Laboratory, St Boniface Hospital, 409 Tachd Avenue, Winnipeg, Manitoba, Canada R2H 2A6 (Present address).
© Oxford University Press 1990

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

When bisecting radial lines visually, normal subjects err towards distant peripersonal space, and when
bisecting vertical lines visually, they err towards upper vertical space. In contrast, when bisecting lines
under tactile-proprioceptive guidance, subjects err towards near peripersonal space, suggesting that normally
attention is preferentially distributed away from the body during visual exploration but distributed towards
the body during tactile exploration. A patient with ischaemic lesions, however, involving both inferior
temporal lobes neglected far peripersonal and upper vertical space. He also demonstrated a motor bias
away from the neglected space. These findings suggest that in man attention is spatially directed in three
orthogonal axes (horizontal, vertical and radial), that attention may normally be unequally distributed in
each of these axes, and that neglect may occur in not only the horizontal axis but also in the radial and
vertical axes.

P. A. SHELTON AND OTHERS

192

METHODS
Subject

B
30-

30-

0-

0-30

0

X
• •
•

0

30

Fio. 1. Visual mislocalization by reaching for a point of light (A) viewed in left upper visual quadrant, 30 cm above
eye level; B, viewed in right upper visual quadrant; and c, fixated 30 cm above eye level, x = target; dots = response;
zero vertical displacement is eye level, zero horizontal displacement is body midline. The intersection of coordinate
lines indicates the point of visual fixation in each case.

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

A 66-yr-old right-handed white man with an eighth grade education who worked as a house painter was
transferred to the Veterans Administration Medical Center in Gainesville, Florida, from an outside hospital
following a stroke. He experienced the sudden onset of headache, visual loss, and right-sided weakness
in the evening, and the following morning he was found to be blind with mild left-sided paresis and decreased
arousal. His deficits improved over several days, and he was left with inconsistent visual field loss and
a severe anterograde amnesia. The aetiology of his stroke was presumed to be embolic from a cardiac
source; mitral annulus calcification but no thrombus was demonstrated on echocardiography. He had a
pacemaker inserted for atrial fibrillation and sinus pauses of 2 - 3 s and he was maintained on digoxin,
quinidine, verapamil and warfarin.
Six weeks after his stroke he was alert and cooperative but distractable, with motor impersistance for
eye closure and arm extension. He was generally hypokinetic, and there was a paucity of spontaneous
speech. Language was fluent with good repetition and naming. Comprehension of conversational speech
was normal but he had some difficulty with complex sequential commands. He was disorientated for time
and place, had a severe anterograde amnesia for verbal and figural material, and a relatively extensive
retrograde memory deficit extending 20 yrs into the past. Verbal fluency was 1 item in 60 s; praxis, simple
calculations, right/left orientation, and finger recognition were normal. Insight, general knowledge,
vocabulary, abstract reasoning and problem solving were compatible with his educational and occupational
level. He had a stable bilateral upper hemivisual field loss, which was complete except for the ability to
detect and localize light. Pursuit eye movements in all directions were broken into small saccades. Visual
function improved to the point where he was able to discriminate high contrast gratings with a period of
less than 30 min arc, and he localized target dots of less than 12 min arc diameter. Optokinetic nystagmus
was elicited with 25 min arc gratings to right, left and upward stimuli, resulting in downward saccades.
He traced drawings of standard line widths with his right index finger but was unable to identify figures
on the usual charts for estimating visual acuity. He accurately localized all light stimuli under visual guidance
for targets in the lower visual fields by reaching out with his right hand, but made a systematic error towards
the lower right when reaching for targets in the upper visual fields. His error was 3—4 cm for targets
in the right upper quadrant and 30—40 cm for targets in the left upper quadrant. He was aware of his
errors and perplexed by them. To isolate the effect of spatial field on reaching behaviour, target location
in the visual field was controlled by having the patient fixate the target. When fixating targets placed in
various spatial fields, he accurately localized them in lower hemispace (below eye level) but made systematic
errors in reaching for targets in upper hemispace (fig. 1). He was unable to differentiate relative line

PER1PERSONAL AND VERTICAL NEGLECT

\

193

\ ; w \ \

\

x

orientations of less than 90° but could differentiate circle size differences of 2 mm. A complete description
of his apperceptive visual agnosia, colour agnosia, and prosopagnosia will be reported elsewhere.
This man demonstrated left hemispatial neglect on line bisection and cancellation tasks (fig. 2). In addition,
he failed to cancel targets in the upper quadrants and misbisected vertical lines below midpoint. He neglected
upper and left portions of drawings, in copying figures and in tracing, but completed them when the stimulus
figures were inverted. Numbers clustered in the right lower quadrant of his drawing of a clock (fig. 3).
He had mild bilateral paratonic rigidity ('Gegenhalten'), worse on the right, and mild weakness of the
right deltoid and triceps and left upper extremity generally. The tendon reflexes were brisk in the left
upper and lower extremities and the plantar responses were extensor bilaterally. There was mild dysmetria
and intention tremor in the left arm, and he had mildly decreased proprioception in the fingers of the right
hand. There was, however, no dysmetria in the right arm as evidenced by normal finger-nose testing with
eyes closed. Crude sensation, tactile object and number recognition, and tactile and proprioceptive point
localization on the body and in space were normal as he was able to locate the opposite hand in space

FIG. 3. Drawing of clock.

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

FIG. 2. Line cancellation task presented free-field on desk top (transverse plane), 30 cm from body surface.

194

P. A. SHELTON AND OTHERS

with eyes closed. CT scans documented the evolution of bilateral inferior temporal lobe and deep occipital
lobe ischaemic infarction, worse on the left, but no evidence of parietal lobe pathology (fig. 4).
He was seen in follow-up 6 months after his first evaluation. At this time his visual acuity was testable
with a Snellen eye chart because of improvement in number recognition, and his acuity had improved
to 20/100 in the right eye and 20/200 in the left. His visual field had enlarged to approximately 60° above
the horizontal meridian at their maximum using moving or stationary solid targets, but varied in size from
day to day. On some days he had only 20° of vision in the superior visual quadrants, and he demonstrated
significant fatigue over the course of each testing session such that a region of visual inattention for stationary
stimuli gradually replaced the visual extinction in upper visual fields, and finally no visual function could
be documented above the horizontal meridian at the end of the test session on some days. Upper visual
quadrant extinction was clearly demonstrated for stimuli (90° movement of the examiner's index finger)
presented simultaneously in upper and lower visual quadrants in each of the right and left hemifields. Rightsided tactile extinction was present, and right visual extinction was demonstrable only for stimuli presented

simultaneously in both superior quadrants. There was no extinction of peripheral targets when central and
peripheral stimuli were presented simultaneously in the lower visual fields. Testing for extinction of distal
versus proximal stimuli in peripersonal space was more difficult and limited by the degree of preserved
superior quadrant vision. Nonetheless, on some occasions an upper visual field stimulus was no longer
extinguished when presented 30 cm closer to the subject's face than a simultaneous stimulus in the lower
visual quadrant. Visual mislocalization toward the lower right was still present but of lesser degree.
Controls
There were 12 control subjects, 8 male and 4 female. Eight were selected from the orthopaedic and
internal medicine inpatient service (controls 1 —8) after excluding active central nervous system disease,
a past history of stroke, seizures, dementia, malignancy, major psychiatric disorder, or evidence of decreased
arousal or neglect on screening neurological examination. The mean age of the hospital controls was 57 yrs
and their mean educational level was 10 yrs. Four healthy volunteers also participated (controls 9—12)
and had a mean age of 35 yrs.
To evaluate the nature of this patient's neglect more fully, his performance was tested using a line bisection
task under a variety of spatial conditions, and we compared his performance with that of control subjects.

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

FIG. 4. CT scan at the level of A, inferior temporal lobes, and B, calcarine cortex 6 weeks poststroke, revealing hypodense
lesions in posterior medial temporal lobes (fusiform and lingular gyri), left greater than right, and dilated occipital horn
of the left lateral ventricle.

PERIPERSONAL AND VERTICAL NEGLECT

195

Apparatus and procedures
Subjects were asked to bisect single black lines of 2.5 mm width and of six different lengths (20, 22,
24, 26, 28, 30 cm), and centred on a 21.6x35.5 cm page (visuomotor condition). Stimuli were presented
in blocks of 12 trials, 2 of each length, in a pseudorandomized fashion. Stimuli were located in 12 different
areas of egocentric extrapersonal space for the patient, along 3 orthogonal axes (fig. 5). Horizontal lines
were presented in the usual fashion with pages in the transverse plane flat on a tabletop and with lines
orientated along the intersecting frontal (coronal) plane 30 cm in front of the subjects (horizonal condition).

Vertical lines were presented on pages 30 cm in front of the subjects at the intersection of the frontal and
midsaggital plane (vertical condition). Radial lines were presented on pages in the transverse plane at table
level with lines orientated along the intersection of the midsagittal plane (radial condition). Stimuli were
placed in 3 locations along the horizontal axis: with midpoint approximately 30 cm right, 30 cm left, or
at the midline. Stimuli were also placed in 3 locations along the radial axis: with pages adjacent to the
body surface (near), with the line midpoint approximately 30 cm (middle), or 60 cm (far condition) from
the subject; and in 6 locations along the vertical axis, from 15 cm above eye level to 60 cm below eye
level. Three blocks of 12 trials were presented in each radial location, and one block of 12 trials were
presented in the horizontal and vertical locations. The order of presentation of blocks of trials was
counterbalanced to control for the effects of practice and fatigue, and test sessions were limited to 1 h.
It was required that the subject explore the full extent of the test line before attempting to bisect it, there
was no time limit for completion of the task, and he/she was reminded of the instructions when necessary.
A tactuomotor line bisection task was given to the subjects by using wooden rods of 2.5 mm width and
again of 6 lengths (20, 22, 24, 26, 28, 30 cm). One block of 12 trials was presented in each of the following
locations: far-radial, eye level vertical, and right and left-horizontal. The subject's index finger was placed
on the end of the rod, and he was instructed to explore the rod and find the midpoint while blindfolded.
Again, the subject was required to explore the full extent of the stimulus and to identify both ends of the
rod before attempting to bisect it. They were started at one end of the rod in half of the trials and at the
opposite end in the other half, the order of the starting point being varied pseudorandomly.
The test procedures for the controls were the same as for the patient except that only one block (12
trials) was presented in each of 6 locations for the visuomotor task: near and far space along the radial

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

FIG. 5. The region of egocentric extrapersonal space
studied; stimuli aligned along three orthogonal axes.

196

P. A. SHELTON AND OTHERS

dimension, right and left hemispace, and at eye level and 30 cm below eye level in the vertical dimension.
Rods were presented in 4 locations for the tactuomotor task: near, far, eye level, and 30 cm below eye
level. In addition to the noncued line bisections in the visuomotor condition, both the patient and the control
subjects were presented with an equal number of cued line bisections where subjects were required to
name letters at the line ends before performing the bisection to document that the subject had attended
to the entire line.
Analysis

RESULTS
As shown in Table 1, for the 3 visual line bisection tasks, results of the ANOVAs
indicated that controls significantly misbisected radial [ F ( l , l 1) = 19.63, P = 0.001]
and vertically placed lines [ F ( 1 , U ) = 36.38, P = 0.0001], whereas horizontal lines
were not significantly misbisected [ F ( l , l l ) = 2.96, P = 0.11]. All remaining effects
and interactions were nonsignificant, that is, cueing did not affect bisection performance
across any of the spatial conditions [radial: F ( l , l l ) = 2 . 1 1 , P = 0.17; vertical:
F ( l , l l ) = 1.26, P = 0.28; horizontal: F ( l , 11) = 0.02, P = 0.90]. The data from the
two conditions were pooled for each location in subsequent analyses. There was a trend
for vertically presented lines to be more accurately bisected when positioned below eye

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

Visual line bisection error was measured to an accuracy of ± 1 mm from the true midpoint, and the
mean error and SD in cm were calculated for each condition. Tactile bisections were measured to an accuracy
of 0.25 cm. The direction of the error along each dimension was arbitrarily assigned a (+) or ( - ) value.
Errors to the right of midpoint of horizontal lines, above the midpoint of vertical lines, and beyond the
midpoint of radial lines, were assigned (+) values whereas errors to the left, below, or near the midpoint
for radial lines were given ( - ) values. To facilitate interpretation of the results, the definition of egocentric
extracorporeal space was elaborated to specify regions of space with particular relevance in terms of
behaviour. Extracorporeal space was divided into peripersonal space, within arm's reach, and extrapersonal
space, beyond arm's reach. Peripersonal space was further divided into near space, immediately adjacent
to the body surface, middle and far peripersonal space. Location may be specified in a polar coordinate
system, along a radius extending from body centre and orientated at a given angle from the anterior midsagittal
line in a plane transverse to the long axis of the body. Such a system does not diminish the importance
of bilateral symmetry in the concept of extracorporeal space but radial distance from the individual may
influence performance independently within a given hemispace.
The mean line bisection error (LBE) in cm was calculated for each block of trials for both patient and
controls across the 5 bisection tasks (visual-vertical, visual-radial, visual-horizontal, tactile-vertical, tactileradial). The control data from these 5 tasks were analysed using repeated measures analysis of variance
(ANOVA). For all analyses, the mean LBE was compared with zero (i.e., the actual midpoint) by constructing
a null set of data.
Separate ANOVAs were performed for the 3 visual bisection tasks, using Position (eye level vs below
eye level for the vertical task; near vs far for the radial task; left vs right for the horizontal task); and
Cue (cued, uncued), and Data (actual, null) as within-subject factors. Separate ANOVAs were also performed
for the two tactile tasks using Position (eye level vs below eye level for the vertical task; near vs far for
the radial task) and Data (actual, null) as the within-subjects factors.
The mean age of the control group was 1 SD less than the age of the patient (66 yrs); the control group
was therefore divided into subjects older (n = 5) and younger (n = 7) than 55 yrs. The mean age of the
older group was 65.4 yrs (SD = 8.73), and the two groups were compared by ANOVA. As there was
no significant difference in LBE between the young and older control subjects on any of the 5 tasks just
described, the data for the whole control group were pooled for subsequent comparisons with the patient.

P E R I P E R S O N A L AND V E R T I C A L N E G L E C T

197

TABLE 1. LINE BISECTION ERROR: MEAN OF 12 CONTROL
SUBJECTS

Visual
Near
Far
Eye level
Below
Tactile
Near
Far
Eye level
Below

Mean
(cm)

SD

I*

P value

0.65
0.90
0.75
0.54

0.66
0.66
0.42
0.40

3.26
4.54
5.95
4.52

<0.01
<0.001
<0.0001
<0.001

-0.09
-0.92
-0.34
-0.38

0.68
0.99
1.42
1.11

-0.40
-2.78
-0.72
-1.03

n.s.
<0.05
n.s.
n.s.

level (X = 0.54 cm, SD = 0.40) than when positioned at eye level (X = 0.75,
SD = 0.42), indicating a gradient of response similar to that seen in hemispatial neglect
[F(l,ll) = 4.60, P = 0.055].
The control group's means for the 2 tactile bisection tasks (vertical, radial) are presented
in Table 1. Results of the ANOVAs indicated that control subjects significantly
misbisected radially positioned lines [F(l,9) = 11.12, P = 0.0087]. In addition, there
was a trend for radial lines in near peripersonal space to be more accurately bisected
than lines in far peripersonal space [F(l,9) = 3.47, P = 0.095], indicating another
gradient of response like that observed in hemispatial neglect. No significant main effects
or interactions were observed in the vertical tactual line bisection task [F(l ,9) = 1.02,
P = 0.034].
Finally, the visual and tactile bisection tasks were directly compared for each spatial
condition (radial, vertical) in separate ANOVAs. In both conditions, Modality (visual,
tactile) and Position (near vs far radial, eye level vs below eye level vertical) were the
within-subject factors. In these analyses, only the uncued visual data (as there was no
cued condition in the tactile tasks) and data from the 10 subjects completing both visual
and tactile tasks were used. The visual and tactile LBE for all subjects are presented
in figs 6 (radial) and 7 (vertical). Results of these analyses indicated a significant
difference in bisection performance between the two modalities in both the radial
[F(l,9) = 21.81, P = 0.0012] and vertical conditions [F(l,9) = 8.39, P = 0.018].
The LBE was of equal magnitude in each modality in far peripersonal space but in opposite
directions. There was a significant ModalityxPosition interaction [F(l,9) = 5.15,
P = 0.049], and post hoc comparisons using Tukey HSD revealed that visual far radial
differed significantly from the tactile far radial LBE (P < 0.05). Together, these results
indicate that normal subjects err toward the body on tactuomotor but away from the
body on visuomotor radial line bisections and that the error is of greatest magnitude
in far peripersonal space. Even though no significant vertical tactuomotor LBE was
demonstrated, normal subjects also erred in a consistent fashion upwards or away from
the body on the vertical visuomotor task.

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

* One sample, two-tailed t test, df = 11. ( + ) values for LBE beyond midpoint (radial)
or above midpoint (vertical); (—) values for LBE near midpoint or below midpoint.

198

P. A. SHELTON AND OTHERS

-3

6 7 8
Subjects

9

10 11 12

FIG. 6. Mean line bisection error of 12 control subjects for radial stimuli in visuomotor and tactuomotor conditions.
( + ) error beyond, ( —) error near midpoint. Closed bars = visual; open bars = tactile.

FIG. 7. Mean line bisection error of 12 control subjects for vertical stimuli in visuomotor and tactuomotor conditions.
( + ) error above, ( - ) error below midpoint. Closed bars = visual; hatched bars = tactile.

As shown in Table 2, the patient's mean LBE was compared with zero (no LBE)
using a one sample, two-tailed t test. To control for the effect of multiple t tests, the
patient data were adjusted with Bonferroni's inequality. The patient repeated the bisection
procedure in 12 conditions, and the P values reported in the table have been adjusted
to P x 12. The patient's mean error in each condition was compared with the controls
by calculating the number of SDs from the sample mean. The patient misbisected
horizontal lines to the right of midline in the visual condition (midline location:

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

5

199

PERIPERSONAL AND VERTICAL NEGLECT
TABLE 2. LINE BISECTION ERROR: PATIENT DATA
Visual
Condition
Horizontal

Radial

Vertical

Tactile

Mean
1.18
1.88
-5.56

SD
1.59
1.00
1.88

t*

n.s.
< 0.001
<0.0001

Mean
4.17
-1.54

SD

2.58
6.52
-10.26

1.93

7.46

Pc
< 0.0002

2.07

-2.58

n.s.

Near
Middle
Far

-2.34
-5.33
-7.10

1.49
2.35
1.95

-0.46**
-13.59**
21.85**

<0.000001
<0.000001
<0.000001

_
-2.74

2.13

-4.46

<0.02

+ 15t

-5.78
-5.66
-1.04
-2.98
-2.59
-1.18

2.08
2.20
1.96
2.23
2.03
2.13

-9.62
-8.89
-1.84
-4.63
-4.41
-1.91

<0.0001
<0.0001
n.s.
<0.01
<0.02
n.s.

_
-2.73
_
_

3.16

-2.99

n.s.

EL
-15
-30
-45
-60

t*

Pc

* One sample, two-tailed t test, df = 11; ** df = 35. t cm above ( + ) or below ( - ) eye level. ( + ) values for LBE
beyond midpoint (radial), above midpoint (vertical) or the right (horizontal); ( - ) values for LBE near of midpoint or
below midpoint; EL = eye level; P c = Bonferroni adjustment P x 12.

X = 1.88 cm, SD = 1.00; left hemispace: X = 1.18 cm, SD = 1.59) but erred to
the left for stimuli located in right hemispace (X = 5.56, SD = 1.88). The LBE was
significant at midline (t = 6.52, P < 0.001) and in right hemispace (t = 10.26,
P < 0.001). His error was greater than 3 SD from the control sample mean for stimuli
in left hemispace and 15 SD in right hemispace, indicating a significant degree of left
hemispatial neglect and, at least for stimuli more peripheral in the right hemispatial
field, a degree of right hemispatial neglect. He manifested only left hemispatial neglect
in the tactuomotor condition (X = 4.17, SD = 1.93, P < 0.001) and only with stimuli
in left hemispace.
There is no horizontal plane for separating upper from lower hemispace that is
comparable with the midsagittal plane that separates right from left hemispace. When
the eyes are in neutral position, the midsagittal plane also separates the right visual field
from the left. Therefore we chose eye level as the plane to separate upper from lower
hemispace because when the eyes are in neutral position this plane separates upper from
lower visual fields. In the visual modality the patient erred consistently below the midpoint
of vertical lines (eye level: X = 5.66 cm, SD = 2.20 cm, P < 0.0001; 30 cm below
eye level: X = 2.98 cm, SD = 2.23, P < 0.01) and was 15 SD from the control mean
at eye level. He demonstrated a gradient of response such that LBE was of greatest
magnitude for stimuli in upper vertical space as defined by eye level. In the tactuomotor
condition, he erred below midpoint (X = —2.73, SD = 3.16) but with the conservative
estimate of significance used (the Bonferonni adjustment), it failed to reach statistical
significance (t = - 2 . 9 9 , P < 0.12).
Finally, the patient consistently misbisected radial lines towards his body at all distances
in both visual and tactile modalities. In far peripersonal space in the visual modality
he erred by a mean 7.10 cm (SD 1.95 cm, P < 0.000001) and in the tactile modality
by 2.74 cm (SD 2.13 cm, P < 0.012). In near space in the visual modality he misbisected

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

Location
Left
Midline
Right

200

P. A. SHELTON AND OTHERS

radial lines by a mean 2.34 cm (SD 1.49 cm, P < 0.000001). He again demonstrated
a gradient of response with greater directional errors for more distant stimuli. His LBE
was greater than 4 SD from the mean of the controls in both near and far visual conditions
and 1.8 SD from the controls in the tactuomotor condition.
DISCUSSION

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

The patient described here demonstrated left as well as a degree of right hemispatial
neglect on a cancellation task and on line bisection. In addition, he clearly demonstrated
extinction of visual stimuli (finger flexion) presented to either superior visual quadrant
with simultaneous lower quadrant stimulation. The errors in line cancellation over the
top half of the page and in copying figures are analogous to the exploratory and
constructional deficits seen in hemispatial neglect and suggest neglect of upper vertical
space. When stimuli viewed in the superior visual field were brought near the patient,
they were no longer extinguished when presented simultaneously with distant lower
visual field stimuli, suggesting neglect of more distant stimuli in the superior visual
fields. Retinal image size, however, is a major determinant of depth perception, and
this study did not allow us to determine which factor is the most relevant. Near stimuli
may also be qualitatively different than far stimuli, and these differences may have
accounted for our results. He misbisected vertical lines below and radial lines near
midpoint under both visual and tactile control. His errors were consistent and differed
significantly from those of the control subjects, suggesting neglect of upper vertical
and far peripersonal space.
Because hemianopic patients do not necessarily have neglect and many patients with
neglect are not hemianopic, loss of elementary visual function cannot itself account for
these deficits in spatial behaviour (Rosenberger, 1974; Ogden, 1985). The patient traced
relatively complex figures and had no difficulty in tracing single lines whose extent
was limited to the right inferior visual quadrant. His errors and his failure to explore
the material were consistently related to the spatial and visual field in which the stimulus
was located. Therefore, whereas neglect may have influenced his visual recognition
and contributed to his agnosic errors, his visual and perceptual deficits cannot totally
account for the line bisection error in upper space. Furthermore, there is no reason
to suspect that upper visual field loss would affect line bisection performance more than
hemianopia does.
Elementary motor deficits cannot explain the patient's response bias as all testing was
done with the right hand. There was only minimal weakness in the right arm in a pattern
that cannot be ascribed to his stroke, and limb movement throughout the field where
stimuli were located was unrestricted. As he was required to explore the full extent
of test lines before tactile bisection, clearly he was able to reach midpoint in all cases.
For far radial stimuli the response was always further from the body than the true midpoint
of near or midradial lines, indicating that his responses fall short of midpoint for a reason
other than restriction of basic motor function. Similarly, in the cued visual bisection
task, the identification of figures at the line ends proved that the full extent of the line
had been perceived before bisection.

PERIPERSONAL AND VERTICAL NEGLECT

201

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

A pathophysiological substrate for vertical neglect has not been entirely elucidated.
The patient with lower altitudinal neglect previously reported (Rapcsak et al., 1988)
had bilateral parieto-occipital lesions, whereas this patient had bilateral inferior temporooccipital lesions. As indicated by Rapcsak et al. (1988), the upper hemiretina has
physiological superiority, and lower visual field stimuli may be more successful in
competing for limited attentional capacity. Subjects with a global decrease in visual
spatial attention capacity may therefore manifest upper visual inattention because of
the comparatively greater saliency of lower visual field stimuli. This alone, however,
cannot account for the multimodal nature of vertical neglect in these 2 patients. Electrophysiological studies in monkeys have demonstrated the existence of populations of lightsensitive inferior parietal (IPL) neurons with visual receptive fields confined to upper
or lower visual quadrants (Robinson et al., 1978; Motter and Mountcastle, 1981). Some
cells with quadrantic receptive fields include the fovea, and this class generally discharged
to a light stimulus in that quadrant or their activity was further enhanced by saccades
towards the light stimulus. Recording from IPL, Robinson et al. (1978) demonstrated
a preponderance of receptive fields in the lower contralateral visual field, and Motter
and Mountcastle (1981) found that cells in IPL, active during hand manipulation, had
visual receptive fields confined to the lower quadrants. These cells appear similar in
other respects to IPL neurons which are proposed to mediate the direction of attention
(Mountcastle, 1976). The paucity of cells with upper quadrant visual receptive fields
identified in area 7a suggests that neurons with these 'attentional' properties may be
located elsewhere. If so, the inferior temporal visual association cortex might contain
such cells.
In addition to his vertical neglect in cancellation, line bisection, and drawing tasks,
our patient demonstrated misreaching and consistently erred to the lower right. He not
only misreached for targets in upper visual half-fields but also in upper hemispace,
irrespective of the retinotopic location of the stimulus. Although he had a mild
proprioceptive deficit in the digits of the right hand, there was no dysmetria and he
was able to localize his opposite limb accurately in space when blindfolded. A proprioceptive deficit alone should not cause a consistent directional error. Furthermore, this
task requires an open-loop movement on the first trial in any location and requires
proprioceptive information from both the passively positioned as well as the actively
exploring limb (DeJong, 1988). Point localization through the use of proprioceptive
information was preserved, therefore misreaching was due to a defect in visually-guided
motor behaviour. Holmes (1918, 1919) studied visual localization in patients with bilateral
brain injury and found that localization was defective throughout the visual fields.
Subsequently, however, cases with defective visual localization restricted to the left
or right homonymous visual half-fields were described (Riddoch, 1935; Brain, 1941;
Bender and Teuber, 1947). Some patients were consistent in the direction of error
(Poppelreuter, 1917) but others were not (Cole et al., 1962; Ratcliff and Davies-Jones,
1972). Subjects with acute lesions characteristically misjudged direction (localization
in a frontal plane) and distance (localization in radial direction) but patients with chronic
deficits reported by Ratcliff and Davies-Jones (1972) erred only in direction. When

202

P. A. SHELTON AND OTHERS

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

subjects who misreach directed their gaze peripherally so as to fixate peripheral targets,
Cole et al. (1962) found that the misreaching disappeared. On this basis they concluded
that visual disorientation was due to a visual perceptual defect and not to motor guidance
per se (reaching). However, Levine et al. (1978) reported a case with inaccurate reaching
due to a tumour of the superior parietal lobule of the right hemisphere. When the target
was fixated but the left hand could not be seen, misreaching persisted. If the subject,
however, was allowed to view the left hand as it approached its goal, misreaching
disappeared altogether. It appears that the deficit was in the cross-modal integration
of proprioceptive and visual information about spatial location.
Unlike these previous cases, our patient mislocalized targets when he was allowed
to view his hand both in the upper hemifield and in upper hemispace when the targets
above eye level were fixated. Accurate reaching for fixated targets in lower hemispace
indicates that the spatial field of the target had an independent influence on visuallyguided reaching behaviour. Visual mislocalization was systematically towards the lower
right, away from the upper and left spatial fields where he manifested spatial neglect.
A similar but horizontal motor bias, termed directional hypokinesia, has been
demonstrated in patients with neglect from isolated parietal lobe lesions (Valenstein et al.,
1982; Heilman et al., 1985a). Activation of movements towards the neglected field
is decreased and manifests as delayed initiation of movement towards neglected space
as well as a tendency to fall short of the target (Meador et al., 1988). The pattern of
mislocalization demonstrated by this man may be due to directional hypokinesia and
a function of the attentional-intentional mechanisms rather than a visual perceptual
disturbance per se.
Unlike the patients of Cole et al. (1962) and Levine et al. (1978) whose mislocalization
disappeared for peripheral targets visualized in the central visual field, mislocalization
persisted in the central visual field in our case but was less severe. This amelioration
of misreaching in central vision could represent the same kind of improvement
demonstrated by the previously reported patients and which was believed to support
the supposition that visual disorientation is due to a defect in visual perception. The
dependence on spatial field of stimulus demonstrated here, however, is characteristic
of attentional processes. A hemispace-visual field interaction has been demonstrated
for visual extinction whereby stimuli located more peripherally in the visual field are
neglected to a greater degree (Rapcsak et al., 1987). In addition, neglect is most profound
when visual half-field and head and body hemispace coincide (Bowers et al., 1981).
Directional hypokinesia therefore appears to be the most plausible explanation for
misreaching in this man. It remains unclear to what extent intentional or visual spatial
neglect may account for misreaching and visual disorientation in general, as previous
studies have failed to control for hemispace. Directional hypokinesia may actually be
a frequent cause of misreaching but would appear to be restricted to cases with systematic
directional errors.
Ventre et al. (1984) proposed that visual mislocalization and motor aspects of
hemispatial neglect are related phenomena, 'the two syndromes overlap and are part
of the same functional, if not anatomical, entity'. Further, they proposed that both

PER1PERS0NAL AND VERTICAL NEGLECT

203

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

syndromes may result from a displacement of an egocentric coordinate system relative
to allocentric coordinates derived through sensory input and following a parietal lesion.
The results of the current study are consistent with this hypothesis in that visually-guided
localization of an external point showed a systematic directional error, whereas no such
misreaching was demonstrated towards parts of the patient's own body, where location
may be interpreted in terms of egocentric coordinates alone. In the tactile line bisection
task, the systematic directional error reappeared, and the results can be interpreted as
a displacement of the egocentric coordinate systems guiding the movement of the
exploring finger relative to the accurately perceived allocentric coordinates of the midpoint
of the rod. The original hypothesis of Ventre et al. (1984) may be explained on the
basis of a failure of cross-modal integration between retinotopic location and proprioceptive information required for accurate reaching. If the response bias is due to
mismatching of egocentric and allocentric coordinates then the same error should exist
irrespective of sensory modality. The results of the line bisection test are consistent
with the hypothesis but it is not possible to dissociate the effect of the divided attention
aspect of the task.
The study of spatial neglect has emphasized the importance of contralateral hemispace
in the control of behaviour. Brain (1941), however, suggested that the region of
extrapersonal space within grasping distance was also of special significance. More recent
studies support this contention and suggest that attentional processes mediate this bias.
Leinonen et al. (1979) and Leinonen and Nyman (1979) showed that a population of
neurons in area 7b of the parietal cortex of the monkey are multimodal and that their
activity is enhanced only by visual targets approaching their cutaneous receptive field
or by stationary stimuli within 5-10 cm of it. Most of these light-sensitive cells responded
only to movement in a specific direction and did not respond at all if the target was
greater than 1 m from the monkey. Area 7b projects in a unidirectional fashion to the
posterior bank of the arcuate sulcus (Chavis and Pandya, 1976) and Rizzolatti et al.
(1981a, b) showed that cells in this region have properties analogous to those in area
7b: 54% of the neurons responded consistently to visual stimuli only if less than 10 cm
away, and 46% responded to stimuli within arm's reach. They introduced the term
peripersonal space to denote this region of extrapersonal space and showed subsequently
that postarcuate ablations induce visual and tactile neglect of contralateral pericutaneous
stimuli and increased latency in reaching for targets in peripersonal space. Lesions in
area 7b likewise result in attentional deficits limited predominantly to contralateral
peripersonal space (Rizzolatti et al., 1985).
In the present study we demonstrated that multimodal neglect of events in far
peripersonal space occurs in man. The patient misbisected radial lines near midpoint,
under either visual or kinaesthetic guidance, and there was a gradient of response whereby
his error diminished as stimuli were placed nearer the body. The LBE indicates a deficit
in selective attention to far space and a motor bias toward the body, that is, away from
the neglected field. The attentional mechanisms in man appear to be organized in terms
of proximity to the body in addition to the horizontal and vertical dimension, and focal
hemisphere lesions can affect attention-intention to far space preferentially. The

204

P. A. SHELTON AND OTHERS

ACKNOWLEDGEMENTS
Supported in part by the Medical Research Service of the Veterans Administration and the State of Florida
Department of Health and Rehabilitative Services, Division of Aging and Adult Services.
REFERENCES
BENDER MB, TEUBER HL (1947) Spatial organization of visual perception following injury to the brain.
Archives of Neurology and Psychiatry, Chicago, 58, 721—739.
BOWERS D, HEILMAN KM (1980) Pseudoneglect: effects of hemispace on a tactile line bisection task.

Neuropsychologia, 18,491-498.
BOWERS D, HEILMAN KM, VAN DEN ABELL T (1981) Hemispace-VHF compatibility. Neuropsychologia,

19, 757-765.
BRADSHAW JL, NETTLETON NC, NATHAN G, WILSON L (1983) Head and body space to left and right,
front and rear. II. Visuotactual and kinesthetic studies and left-side underestimation. Neuropsychologia,
21, 475-486.
BRAIN WR (1941) Visual disorientation with special reference to lesions of the right cerebral hemisphere.
Brain, 64, 244-272.
CHAVIS DA, PANDYA DN (1976) Further observations on corticofrontal connections in the rhesus monkey.
Brain Research, Amsterdam, 117, 369—386.
COLE M, SCHUTTA HS, WARRINGTON EK (1962) Visual disorientation in homonymous half-fields.
Neurology, Minneapolis, 12, 257—263.
DEJONG RN (1988) Case taking and the neurologic examination. In: Clinical Neurology, Volume 1. Edited
by R. J. Joynt. Philadelphia: J. B. Lippincott, Chapter 1, p. 12.
HEILMAN KM, BOWERS D, COSLETT HB, WHELAN H, WATSON RT (1985a) Directional hypokinesia:

prolonged reaction times for leftward movements in patients with right hemisphere lesions and neglect.

Neurology, Cleveland, 35, 855-859.
HEILMAN KM, WATSON RT, VALENSTEIN E (1985i>) Neglect and related disorders. In: Clinical
Neuropsychology. Second edition. Edited by K. M. Heilman and E. Valenstein. New York: Oxford

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

localization of these processes in man is unclear. Area 7 of the monkey cortex is analogous
to the inferior parietal lobule of man, but lesions of this region in man have not been
shown to be associated with neglect of distal stimuli. It is possible that inferior temporal
structures are important in the mediation of attention to far peripersonal and upper vertical
space, especially in view of the paucity of light-sensitive inferior parietal neurons with
superior quadrantic receptive fields.
Lastly, attentional or intentional biases have been reported previously in normal
subjects. For example, normal subjects bisect horizontal lines towards the left (Bowers
and Heilman, 1980; Bradshaw et al., 1983). Our control subjects also demonstrated
a bias on radial and vertical line bisection. However, unlike the bias reported for horizontal
lines, the radial bias was modality-specific such that with the visual task the subjects
erred away from their body and with the somaesthetic task they erred towards their
body. The visual system is, by virtue of the ability to sense light, specialized for detecting
distant stimuli. In contrast, the somaesthetic modality is restricted to the detection of
stimuli on or near the body. Perhaps with this specialization there is a modality-specific
attentional bias such that attention is preferentially distributed away from the body during
visual exploration but is biased towards the body during tactile exploration.

PER1PERSONAL AND VERTICAL NEGLECT

205

University Press, pp. 243-293.
HEILMAN KM, BOWERS D, VALENSTEIN E, WATSON RT (1987) Hemispace and hemispatial neglect. In:

Neurophysiological and Neuropsychological Aspects of Spatial Neglect. Edited by M. Jeannerod.
Amsterdam: North-Holland/Elsevier, pp. 115 — 150.
HOLMES G (1918) Disturbances of visual orientation. British Journal of Ophthalmology, 2, 449-468.
HOLMES G (1919) Disturbances of visual space perception. British Medical Journal, 2, 230—233.
LEVINE DN, KAUFMAN KJ, MOHR JP (1978) Inaccurate reaching associated with a superior parietal lobe
tumor. Neurology, New York, 28, 556—561.
LEINONEN L, HYVARINEN J, NYMAN G, LINNANKOSKI I (1979) Functional properties of neurons in lateral

part of associative area 7 in awake monkeys. Part 1. Experimental Brain Research, 34, 299—320.
LEINONEN L, NYMAN G (1979) Functional properties of cells in anterolateral part of area 7 associative
face area of awake monkeys. Part 2. Experimental Brain Research, 34, 321—333.
MEADOR KJ, LORING DW, BARON MS, ROGERS OL, KIMPEL TG (1988) Hemispatial/limb hypometria.

Journal of Clinical and Experimental Neuropsychology, 10, 24.
Clinical and Experimental Neuropsychology, 7, 609.
MOTTER BC, MOUNTCASTLE VB (1981) The functional properties of the light-sensitive neurons of the
posterior parietal cortex studied in waking monkeys: foveal sparing and opponent vector organization.
Journal of Neuroscience, 1, 3—26.
MOUNTCASTLE VB (1976) The world around us: neural command function for selective attention.
Neurosciences Research Program Bulletin, 14, Supplement, 1 —47.
OGDEN JA (1985) Anterior-posterior interhemispheric differences in the loci of lesions producing visual
hemineglect. Brain and Cognition, 4, 59 — 75.
POPPELREUTER WL (1917) Die Psychischen Schadigungen durch Kopfschuss im Kriege 1914-1916: die
Storungen der Niederen und Hoheren Leistungen durch Verletzungen des Oksipitalhirns, Volume 1.
Leipzig: Leopold Voss.
RAPCSAK SZ, WATSON RT, HEILMAN KM (1987) Hemispace-visual field interactions in visual extinction.
Journal of Neurology,

Neurosurgery

and Psychiatry,

SO, 1117—1124.

RAPCSAK SZ, CIMINO CR, HEILMAN KM (1988) Altitudinal neglect. Neurology, Cleveland, 38, 277-281.
RATCLIFF G, DAVIES-JONES GAB (1972) Defective visual localization in focal brain wounds. Brain, 95,
49-60.
RIDDOCH G (1935) Visual disorientation in homonymous half-fields. Brain, 58, 376-382.
RIZZOLATTI G, SCANDOLARA C, MATELLI M, GENTILUCCI M (1981a) Afferent properties of periarcuate

neurons in macaque monkeys. I. Somatosensory responses. Behavioural Brain Research, 2, 125-146.
RIZZOLATTI G, SCANDOLARA C, MATELLI M, GENTILUCCI M (19816) Afferent properties of periarcuate

neurons in macaque monkeys. II. Visual responses. Behavioural Brain Research, 2, 147—163.
RIZZOLATTI G, MATELLI M, PAVESI G (1983) Deficits in attention and movement following the removal
of postarcuate (area 6) and prearcuate (area 8) cortex in macaque monkeys. Brain, 106, 655-673.
RIZZOLATTI G, GENTILUCCI M, MATELLI M (1985) Selective spatial attention: one center, one circuit,
or many circuits? In: Attention and Performance XI. Edited by M. 1. Posner and O. S. M. Marin.
Hillsdale, NJ: Lawrence Erlbaum, pp. 251—265.
ROBINSON DL, GOLDBERG ME, STANTON GB (1978) Parietal association cortex in the primate: sensory
mechanisms and behavioral modulations. Journal of Neurophysiology, 41, 910—932.
ROSENBERGER PB (1974) Discriminative aspects of visual hemi-inattention. Neurology, Minneapolis, 24,
17-23.
VALENSTEIN E, HEILMAN KM, WATSON RT, VAN DEN ABELL T (1982) Nonsensory neglect from

parietotemporal lesions in monkeys. Neurology, New York, 32, 1198—1201.
VENTRE J, FLANDRIN JM, JEANNEROD M (1984) In search for the egocentric reference. A neurophysiological
hypothesis. Neuropsychologia, 22, 797 — 806.
(Received October 5, 1988. Revised April 18, 1989. Accepted April 26, 1989)

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016

MORRIS R, MICKEL S, BROOKS M, SWAVELY S, HEILMAN K (1985) Recovery from neglect. Journal of

Downloaded from http://brain.oxfordjournals.org/ by guest on April 8, 2016