c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7

ava i lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /cor tex

Research report

Leftward motion restores number space in neglect

Elena Salillasa,b,*, Alessia Granaa,c, Montserrat Juncadellad, Imma Ricod andCarlo Semenzae,f

aPsychology Department, University of Trieste, ItalybBiology Department, University of Texas at San Antonio, San Antonio, TX, USAcInstitute of Physical Medicine and Rehabilitation, Gervasutta Hospital, Udine, ItalydHospital Universitari de Bellvitge, Barcelona, SpaineDepartment of Neuroscience, University of Padova, ItalyfI.R.C.C.S. Ospedale S. Camillo, Lido di Venezia, Italy

a r t i c l e i n f o

Article history:

Received 3 December 2007

Reviewed 14 January 2008

Revised 21 April 2008

Accepted 28 September 2008

Action editor Jason Mattingley

Published online 14 November 2008

Keywords:

Attention

Hemispatial neglect

Number cognition

Number comparison

RDK

* Corresponding author. Biology DepartmentE-mail address: elena.salillas@utsa.edu (E

0010-9452/$ – see front matter ª 2008 Elsevidoi:10.1016/j.cortex.2008.09.006

a b s t r a c t

In the present study, a group of patients with left-sided neglect performed a number

comparison task that co-occurred either with coherent motion in different directions or

with random motion. Their performance was compared to that of a healthy control group

and to a group of patients with right hemisphere damage (RHD) but no signs of neglect. The

presence of leftward motion alleviated the difficulties that neglect patients typically show

for a number smaller than the reference number 5 (i.e., number 4). Moreover, the standard

distance effect was only present when the task co-occurred with leftward motion. These

effects were not present in a group of participants with RHD without neglect or in a control

group. The present data extend the effects of optokinetic stimulation (OKS) to represen-

tational neglect, suggesting that an external redirection of attention by the perception of

motion may restore the altered access to the representation of the mental number line in

neglect.

ª 2008 Elsevier Srl. All rights reserved.

1. Introduction mark only those items in the ipsilesional side. Neglect

Neglect is a failure to report, respond or orient to meaningful

stimuli presented to the side opposite a brain lesion when

this failure cannot be attributed to elemental sensory or

motor defects (Heilman, 1979). Neglect patients may fail to

draw the petals on the contra-lesional side of a presented

flower; when asked to bisect a line, they commonly make

their mark towards the ipsilesional side of the line; or, when

asked to cancel stimuli distributed across a page, they usually

, University of Texas at S. Salillas).er Srl. All rights reserved

patients may also show ipsilesional displacement of

egocentric frames of reference, such as the perceived mid-

sagittal plane. The best known aspect of the syndrome is

visual extrapersonal neglect as opposed to neglect of the body

surface (personal or bodily space). Neglect may also occur in

the internal representational space. These patients fail to

recall from memory the left-sided items (with respect to

a required viewpoint) items in well-known places, like the

Duomo Square in Milan in Bisiach and Luzzatti’s (1978)

an Antonio, One UTSA Circle, San Antonio, TX 78249, USA.

.

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7 731

seminal paper, or fail to describe from memory the contra-

lesional details of a room (Denny-Brown and Banker, 1954).

This is known as representational neglect and is the focus of

the present study.

Research on internal number representation has recently

focused on spatial neglect with the general goal of finding

evidence of the relation between numbers and space. Neglect

in internal number representation would be then a pure case of

representational neglect. Zorzi et al. (2002) presented spoken

numeric intervals (e.g., 1–5) to neglect patients and asked them

to make a judgment about the central number of the interval.

Just as these patients do with line bisection, they systemati-

cally displaced the centre of the number interval to the right

(for example, stating that five is halfway between two and six).

For Zorzi et al. (2002), this finding demonstrated the spatial

nature of the mental number line and its striking functional

isomorphism to physical lines. Other studies have shown that

neglect patients have difficulties in processing a number

‘‘located’’ on the left side of a reference number (i.e., a smaller

number) against which a number comparison has to be made

(Vuilleumier et al., 2004). When the reference number changes,

patients show the same difficulty with numbers smaller than

the new reference number. Vuilleumier et al. (2004) concluded

that different spatial representations are constructed, where

different numbers are neglected depending on the number

taken as reference in a comparison task. Representational

neglect-like symptoms in number bisection were recently

shown also in healthy participants when repetitive trans-

cranial magnetic stimulation (rTMS) was applied over the right

posterior parietal cortex (Gobel et al., 2006).

In brief, these studies imply that neglect pathology affects

numbers in a similar manner to other types of spatial infor-

mation. Thus, numbers seem to be encoded in spatial terms,

and this spatial representation or access to it is altered in

neglect. The present study is based on the prediction that

attention to contralesionally moving stimuli, using a tech-

nique called ‘‘OKS’’ (see below), can ameliorate this patho-

logical representation in neglect. (e.g., Pizzamiglio et al., 1990;

Mattingley et al., 1994; Rossetti et al., 2004).

Unilateral neglect has been experimentally remediated using

sensorial-exteroceptive (mainly visual and vestibular) and/or

proprioceptive stimulation. These methods are essentially

bottom-up methods which do not require explicit awareness of

the deficit, unlike visual scanning training (Diller and Weinberg,

1997). Of relevance for the present study is OKS, which is based

on visual displays of numerous stimuli moving coherently

towards the patient’s neglected side. This technique positively

affects performance in neglect. It has been shown that OKS

temporarily modulates focal attention and improves line bisec-

tion (Mattingley et al., 1994; Pizzamiglio et al., 1990). It alleviates

subjective visual straight-ahead deviation (Karnath, 1996),

visual size distortion and distance coding (Kerkhoff et al., 1999;

Kerkhoff, 2000) and position sense (Vallar et al., 1993, 1995).Thus

OKS produces temporary beneficial effects on sensory and

motor deficits (Vallar et al., 1997), subjective straight-ahead and

extrapersonalneglect. Onthe other hand, repetitive OKS(R-OKS)

has shown long-lasting effects up to two weeks after treatment

in cancellation tasks, line bisection, manual line bisection, size

distortion and omissions in text reading (Kerkhoff, 2006). These

restorative effects of R-OKS also generalize to other modalities

such as audition (Kerkhoff, 2003, 2006). Many of these studies

(Pizzamiglio et al., 1990; Vallar et al., 1993, 1995; Karnath, 1996)

show that OKS in the ipsilesional direction (i.e., rightwards) has

negative effects, with a decline in performance relative to no

stimulation while some studies have not shown these negative

effects (e.g., Vallar et al., 1993). Generally, these OKS effects are

interpreted as a more or less transient displacement of the

‘‘attentive focus’’ towards a specific side of the spatial field (left

vs right), depending on the direction of optokinetic motion.

To date, no study has addressed the possibility that the

perception of motion can restore the altered numerical

representation in neglect. Another well-known visuo-motor

technique, prism adaptation (PA) (Rossetti et al., 2004) can

improve the bisection of number intervals in unilateral

neglect. After exposure to wedge-prisms that shifted the

optical field 10� to the right, they showed that this visuo-motor

adaptation can in fact improve bisection of number intervals.

This finding is taken as an effect of visuo-motor adaptation on

an internal representation.

Therefore, given the evidence of disrupted numbers repre-

sentation in neglect (Zorzi et al., 2002; Vuilleumier et al., 2004;

Priftis et al., 2006, 2008), the present study investigates the

effect that covert attention to coherent motion stimuli has on

this disrupted number representation. Random dot kineto-

grams (RDKs) with coherent rightward, leftward or non-

coherent direction were presented as background, at the same

time that a number comparison to the reference number 5 was

performed. Importantly, participants had to fixate on a central

fixation point during the experiment, so that the single pattern

of a dot displacement could not be tracked; in addition,

coherent motion direction changed randomly from trial to trial.

In this way, the effect of covert attention to brief (4 sec)

coherent motion stimuli on the number comparison process

was measured. This experimental manipulation of motion and

numbers was presented to a group of neglect patients, a group

of age-matched right hemisphere damaged (RHD) patients

without neglect and an age-matched healthy control group.

The present study explores the effects of sensory visual

stimulation into representational neglect. If visual perception

of coherent motion has an impact on a neglected representa-

tion of numbers, a link between ‘‘representation based’’ and

‘‘attention based’’ accounts of neglect can be suggested. If the

mechanisms of attention that operate on external stimuli are

altered in neglect, as argued by the ‘‘attention based’’ account

(e.g., Kinsbourne, 1987, 1993; Heilman et al., 1993), can the

external compensation for this bias affect the allocation of

attention to the neglected internal representation of numbers?

Given previously reported effects of OKS on neglect

symptoms, an improvement in performance with leftward

motion should be expected, while rightward motion may lead

to deterioration. We also include a condition of random-non-

coherent motion (noise) with all dots in the background image

moving randomly. More than being a baseline, this condition

alters the horizontal opposing vectors proposed by Kins-

bourne (1987, 1993) and should lead to poor performance. This

condition has not been manipulated before in studies of OKS.

Previous studies always consisted of the entire dot array

moving in an ipsi- or contra-lesional direction. In line with

previous studies of OKS, we predict that in neglect patients,

exogenous, covert attention to coherent motion at high levels

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7732

of coherence will have directional effects in a number

comparison task, this may be shown by two indexes:

- The distance effect (Moyer and Landauer, 1967) is a classic and

robust effect that occurs when making number judgments,

in which smaller distances between numbers require more

processing time than larger distances (e.g., comparing 4

and 5 takes more time than comparing 2 and 5). This effect

is taken as evidence of the transformation of numbers into

analogue magnitudes that are subsequently compared.

Thus the normal pattern is an increase in reaction time (RT)

when the numbers to be compared are closer to one

another. If attention to the spatial representation of

numbers, in the form of a mental number line, is altered by

rightward motion as in external stimuli, the distance effect

may result altered as well by rightward motion. The default

rightward bias would be increased leading to a failure to

access and subsequently compare the number to the

reference. As a consequence, the distance effect would be

generally distorted. A disruption of the horizontal vector

system (Kinsbourne, 1987, 1993) by non-coherent motion

should lead to the worst performance in the distance effect,

since these mechanisms of attention for numbers repre-

sentation would have been altered on the top of an already

disrupted cognitive system. In absence of motion stimuli,

neglect patients show a normal distance effect (see Vuil-

leumier et al., 2004, Experiment 1): thus, any alteration of

this pattern with backward random motion could be

explained as a disruption to the horizontal vector system.

- Neglect of numbers smaller than the reference number. Accord-

ing to the results of Vuilleumier et al. (2004), increased

comparison times for numbers smaller than the reference

number should arise in the neglect group. These increased

RTs indicate that numbers on the left of the reference are

neglected, as it happens with other spatial external stimuli.

Table 1 – Neuropsychological assessment. Demographic, clinic

Patients Age(years)

Education(years)

Lesionsite

Lesionetiology

Timesincelesion

(months)

BIT(cut off 12

RHD N1 44 8 FTP IS 6 56

RHD N2 67 17 FT HS 9 65

RHD N3 60 10 FTP IS 5 22

RHD N4 50 8 TP IS 48 127

RHD N5 61 14 D IS 6 122

RHD N6 73 3 D HS 4 86

RHD N7 66 8 D HS 3 126

RHD 1 52 16 TPD Tr 50 141

RHD 2 57 8 FTP IS 4 142

RHD 3 44 14 TP IS 14 140

RHD 5 50 18 FTP HS 12 146

RHD 4 78 8 FP HS 7 140

RHD 6 48 14 TD IS 5 145

RHD 7 50 10 T IS 12 146

RHD, Right Hemisphere Damaged patient; N, with Neglect; F, frontal; T, t

Trauma: D: deep structures, subcortical.

Specifically, in a range of numbers from 1 to 9 and a given

reference number 5, higher RTs arose for the number 4 but

not for 1, 2, or 3. This may occur because smaller distances

to the reference number require more time to be compared

(distance effect). Therefore 4 should be compared to 6, its

homolog on the ‘‘right side’’ of this mental number line.

Importantly, leftward motion should show restorative

effects for the smaller number 4, compared to 6. When

pairing the comparison process with leftward motion,

differences between 4 and 6 should disappear. Instead,

rightward motion would have the worst effects, leading to

the largest difference between 4 and 6, due to the redirec-

tion of attention to the hyper-attended space.

In sum, this pattern of results would point to: (1) restorative

effects of leftward motion in representational neglect and (2)

similar attentional mechanisms operating over external and

representational-numerical stimuli (Dehaene et al., 2003).

These findings would support a link between representation

based and attention based accounts of neglect.

2. Methods

2.1. Participants

Twenty-two right-handed individuals participated in this

study: 14 with RHD (7 with left spatial neglect, mean age: 60.1

years, SD: 10.1; and 7 without neglect, mean age: 54.1, SD: 11.2)

and 8 healthy controls matched in terms of age and education

to the neglect patients (mean age: 59.3; SD: 10.3). Except for

one case of trauma, the etiology of the lesion was determined

by vascular accidents in the territory of the middle cerebral

artery (Table 1).

al and psychometric data of the right hemisphere patients.

9)The

bell test(cut off 32)

The balloon test(cut off B< 17,

Laterality< 55%)

Unilateralrepresentational

neglect

Personalneglect

TotA

TotB

LateralityB

13 9 5 0% þ �12 19 7 43% þ �7 5 1 0% þ þ

27 20 12 33.3% þ þ28 19 17 41.2% � �16 15 12 16.6% þ �28 20 12 33.3% þ þ

35 20 20 50% � �34 20 18 50% � �35 20 16 56% � �35 20 20 50% � �33 20 17 47% � �32 20 20 50% � �33 20 20 50% � �

emporal; P, parietal; IS, ischemic stroke; HS, hemorrhagic stroke; Tr,

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7 733

All patients were tested with the Behavioral Inattention

Test (BIT, Wilson et al., 1987), the Bell Test (Gauthier et al.,

1989) and the Balloon Test (Edgeworth et al., 1998). Unilateral

representational neglect was detected with a test equivalent

to Bisiach and Luzzatti’s (1978) ‘‘Duomo di Milano’’ Test.

Personal neglect was either shown as denial of ownership of

contra-lesional limbs or it was tested by asking patients to find

and reach for their contra-lesional hand (see Bisiach et al.,

1986; Berti et al., 1996). All patients were preserved in the

comprehension of Arabic numerals as shown by their above

cut-off performance in a number comparison task and in

a parity judgment task taken from the Number Processing and

Calculation Battery of Delazer et al. (2003). Demographic,

clinical and psychometric data for the brain damaged partic-

ipants are reported in Table 1. Informed consent was given by

each participant.

2.2. Stimuli

The stimulus chosen for extracting direction of motion was

a stationary moving element called RDK. An RDK consists of

a large number of moving dots randomly positioned within

a restricted area. Each dot is assigned to a particular motion

vector. With these stimuli, a variable percentage of dots can be

moved towards a single coherent direction (signal), while the

rest of the dots carry on moving in random directions (noise).

The perception of motion is based on the joint displacement of

the single elements, and the probability of a single dot being

displaced in a determined direction is called coherence of

element displacement. The dots appear and disappear in the

window at different positions and due to the large number of

dots (thousands) it is impossible for the observer to compute

the relative position of any one dot (Bosbach and Prinz, 2004;

Shadlen and Newsome, 2001). This addresses a criticism that

has been made on other static moving elements as gratings.

The RDKs were created using VisionEgg, which runs under

Python 2.3. A pool of 50 frames was generated as bitmap files.

The speed of the dots was 20�/sec. The total number of dots was

2000, and their size was 2 pixels. As onedotdisappeared from the

screen another appeared at another random location beginning

from theopposite edge.The bitmapswere then transformedinto

movies that lasted 4 sec each, using JPGVideo Version 1.05.0.0

(Independent JPEG group), and then into mpeg files.

Three different movies were created. The first two had 40%

coherence: in one the coherent motion had a horizontal right

direction, while in the other the coherent motion had hori-

zontal left direction. The background in these movies was

noisy, that is, the remaining 60% of the dots were moving

randomly in the window with an orientation of 45�. The third

RDK was one of 0% coherence: in this case, all the dots were

moving randomly as described above. The background of the

RDKs was black and the dots were white.

The digits (2, 3, 4, 6, 7 and 8), presented in yellow, had a size

of .6� and appeared while the movie was playing. A yellow

cross (also with a size of .6�) was present in the centre of the

kinematogram throughout the entire trial, except when the

number substituted the cross. The screen was set to black and

its resolution was 1024� 768 pixels.

Some aspects of the stimuli have to be considered: firstly,

a cross was presented in the centre of the screen with the

function of controlling eye movements. Secondly, in order to

make sure the participants did not fix on or attend to the onset

location of individual dots, the number of dots was high and the

overall pattern did not move. The dots moved across a limited

central window with dimensions 600� 500. Thirdly, the likeli-

hood of presence or absence of a dot at a particular location was

equal, and this information was never useful for the compar-

ison task. Overall, these factors avoided eye movements or the

tracking of the dots and consequently, an explanation of

possible effects in terms of ocular patterns associated to the

direction of motion stimuli. These stimuli characteristics entail

differences with traditional OKS in which a low number of dots

or bars move coherently and patient is allowed to track motion.

2.3. Procedure

The experiment was controlled from a portable PC-compatible

Vaio WGN- S1XP and was programmed using Presentation

software. Participants sat in front of the screen at a distance of

60 cm and were asked to perform a go-no-go task. In some

blocks they had to respond to numbers larger than 5 and in

other blocks they had to respond to numbers smaller than 5.

In go-no-go tasks, it is unlikely to obtain Simon effects

(Ansorge and Wuhr, 2004) that may occur also with directional

motion stimuli (Bosbach and Prinz, 2004). Other stimulus-

response congruency effects such as SNARC are also avoided

with a go-no-go task. This way, a possible general facilitation

of performance for rightward motion is avoided, as patients

respond with their right hand due to hemi-paresis.

The sequence of each trial was as follows: a cross appeared in

the centre of the screen and remained the only stimulus until the

irrelevant motion (RDK) began after 1000 msec. Participants were

asked to fixate on this central cross and maintain this central

fixation for the duration of the experiment. The number

appeared after 1000 msec of the beginning of the movie and

remained in the centre of the screen until a response was made,

orafter3000 msec.Sincethemovie lasted4000 msec, thenumber

coincided with motion during 3000 msec. or until response (see

Fig. 1). Participants used the right hand for responses.

The total number of trials was 432, divided into six blocks.

Response based on whether the comparison to be made was

larger or smaller than five was varied across the blocks: in

three blocks the go-response corresponded to numbers larger

than five and in the other three blocks the response corre-

sponded to numbers smaller than five. The order of the blocks

was randomized. In the end, 12 data points per condition were

obtained. In order to ensure the participants’ attention, the

experiment was divided into three sessions, with two blocks

in each session. Each block lasted a maximum of 6 min and

the participant was allowed to rest at any time during the

experiment or in the intervals between blocks.

3. Results

One neglect participant made a maximum of 12.87% of errors.

The average amount of errors in the two RHD patients was

7.75% for neglect patients and 3.21% for right hemisphere

patients without neglect. In all groups, errors were equally

distributed across participants in false positive and false

Fig. 1 – a) Experimental paradigm and b) Example of a RDK.

1 This group showed lower comparison times for rightwardmotion compared to leftward motion [F(1,7)¼ 6.04, MSE¼ 926,p¼ .04], something we have reported in another study (Salillasand Semenza, in preparation).

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7734

negatives: no systematic bias was observed. In all three

groups, the RTs for error responses were substituted with the

average RT of the condition in each subject. Points above and

below 1.5 standard deviations were replaced with the mean RT

for the condition for each subject. This method of excluding

outliers was necessary due to some sporadic failures to sustain

attention for the two groups of patients. Thus, extremely high

data points were not necessarily associated with the experi-

mental manipulation. The percentage of outliers was 6% and

was randomly distributed across conditions.

The average comparison times were then entered in a four-

way ANOVA, 3 (group: neglect, right damaged without neglect

and control, entered as a between subjects factor)� 2 (numer-

ical size: larger/smaller than 5)� 3 (distance: d3, d2, d1)� 3

(motion: to the left/to the right/random noise). The analysis

revealed a main effect of group [F(1,2)¼ 17.36, p< .001] with

controls having faster RTs, followed by the right hemisphere

damaged patients and then by the neglect group. These overall

differences in number comparison times have been previously

reported by Vuilleumier et al. (2004) using the same task.

Moreover, a group�numerical size interaction was shown in

this analysis [F(2,19)¼ 6.94, p¼ .005, h2¼ .42], with a general

tendency in neglect patients to have higher RTs for numbers

smaller than five while this was not the case in the other two

groups, also in line with the study of Vuilleumier. Finally,

a four-way group� distance�motion�numerical size inter-

action [F(8,76)¼ 2.36, p¼ .025, h2¼ .20] was shown. In order to

analyze this four-way interaction and the group�numerical

size interaction, a separate analysis was done for each group by

a 2 (number)� 3 (distance)� 3 (motion) ANOVA.

3.1. Control group

The control group showed an expected standard distance

main effect [F(2,14)¼ 21.52, h2¼ .75, p< .001] with slower

comparison times in smaller distances.1 No difference in

comparison times between numbers larger and smaller than

the reference number 5 was found and no interaction with

motion direction was found either. No other interactions or

main effects were statistically significant.

3.2. RHD without neglect group

The RHD group showed a main effect of distance [RHD:

F(2,12)¼ 6.1, p¼ .015, h2¼ .5]. As in the control group no main

effects of numerical size were found. No other interactions or

main effects were statistically significant for this group.

3.3. Neglect group

As expected from the group by numerical size interaction, and

unlike the other two groups, a main effect of numerical size

was found [F(1,6)¼ 8.46, p¼ .027, h2¼ .58] in the neglect group,

in which numbers smaller than five generated higher

comparison times than numbers larger than five.

Importantly,a three-waynumericalsize� distance�motion

interaction was found in this group [F(2,24)¼ 3.59, p¼ .02,

h2¼ .37]. This interaction initially confirmed our predictions of

neglect for numbers smaller than the reference. We directly

tested it by contrasting comparison times for number 4 (on the

left of reference number 5) with those for number 6 (on the right

of reference number 5). We focused on this pair (4 vs 6) following

the results of Vuilleumier et al. (2004), where significant differ-

ences were only found between these numbers in the group of

neglect patients.

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7 735

As previously mentioned, no main effect of numerical size

appeared in the other groups and the comparison times

between the pair 6–4 were not statistically significant in any of

the other two groups, therefore post-hoc analyses were only

performed for the neglect group. Comparison times between

the pair 6–4 in each condition of motion were contrasted

(Vuilleumier et al., 2004) (see Fig. 2). Statistically significant

differences in the rightward motion (t¼ 2.87, p¼ .028, h2¼ .58)

and noise motion (t¼ 2.95, p¼ .026, h2¼ .59) conditions were

observed, but this pair was not different when leftward

motion was presented ( p¼ .1). A finer analysis of the differ-

ence in comparison times for the pair 6–4 was performed

through the calculation of the difference between comparison

times for the number 4 minus the comparison times for the

number 6. The average difference for each condition of motion

was 212.1 msec in rightward motion, 146.18 in random motion

and 66.4 in leftward motion. Then t-tests were calculated

Fig. 2 – Mean comparison times plotted with G1 SE as

a function of the distance to the reference number 5 (d1, d2,

d3) and the condition of motion. a) Neglect group, b) RHD

control group and c) control group.

between the different conditions of motion leading to

a significant difference between random motion and leftward

motion (t¼ 2.48, p¼ .048) and a marginal difference between

rightward motion and leftward motion (t¼ 2.29, p¼ .06). The

difference between rightward and random motion was not

statistically significant ( p¼ .36). Therefore, no remarkable

negative effects appeared in the condition of rightward

motion.

In order to test the effects of directional motion in the

distance effect, a numerical size (2)� distance (3) analysis was

performed for each level of motion, which showed diverse

patterns for the distance effect in the different levels of

motion. With rightward motion, an interaction between

distance and numerical size [F(2,12)¼ 4.63, p¼ .03, h2¼ .45]

appeared due to the highest comparison times with rightward

motion in number 4 (see above) this does not mean a regular

distance effect with rightward motion as can be seen in Fig. 2.

With non-coherent motion, no effect involving distance

appeared. Conversely, when leftward motion was presented,

a regular distance effect did appear in the comparison times

[F(2,12)¼ 5.25, p¼ .014, h2¼ .47].

4. Discussion

In our study the perception of motion had an effect on the

number-space representation for neglect patients, measured

through a number comparison task. Indeed, the number

comparison task effectively reflected the spatial characteris-

tics of number representation for the neglect group in the

present study. This task is proposed to entail access to the

analogue magnitude representation (Dehaene, 1992) with

spatial characteristics. As shown by Vuilleumier et al. (2004),

the comparison times to the different numbers in the number

space are selectively increased depending on their position

with respect to the reference number. We further show that in

neglect, concurrent presentation of motion can interact with

the distance effect typically shown through this task. There-

fore, the effects of motion perception in the neglected number

space are twofold:

Firstly, the rightward attentional bias in the numbers

representation, indexed by slower comparison times for

numbers on the ‘‘left’’ side of the reference number, disap-

pears when the number is presented with leftward coherent

motion. This difficulty with number ‘‘4’’, i.e., increased

comparison times, may be another reflection of the altered

access to numbers representation. This could somehow be

compared to the patients’ problems in cancellation tasks over

the external space. Leftward motion allowed neglect patients

to compensate for the rightward bias: this is reflected by the

decrease in difficulty in processing numbers to the left of the

reference number (Vuilleumier et al., 2004). Although no

negative effects of rightward motion were found, when

rightward motion or non-coherent motion were perceived,

increased comparison times to number ‘‘4’’ were found with

respect to number ‘‘6’’.

Secondly, the distance effect, taken as an index of correct

access to the analogue quantity system, is affected by right-

ward and non-coherent motion in neglect, it is only present

when leftward motion occurs during the comparison process.

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7736

When additional attentional biases caused by the perception

of rightward or non-coherent motion are added to an already

altered attentional system, access to numbers is further dis-

rupted and thus the distance effect is not shown. On the

contrary, redirecting attention to the neglected side by left-

ward motion helped the patients to access the numerical

representation.

Direction of motion did not affect processing numbers on

left of the reference number (number ‘‘4’’), or the distance

effect for right hemisphere damaged participants or the

control group. Therefore the interaction cannot be attributed

to damage the right hemisphere per se. Thus, it is the specific

altered network behind the neglect symptoms that must be

affected and in turn compensated by the perception of

external moving stimuli.

The present data agree with previous literature on OKS

that show that leftward motion stimulation improves spatial

perception deficits in neglect (Mattingley et al., 1994; Pizza-

miglio et al., 1990; Kerkhoff, 2006). Importantly, the present

data extend our knowledge of the effects of OKS into

representational neglect. That is, the literature addressing

the effects of OKS in neglect has focused on its effects in

tasks that employ external stimuli and, although it has been

crucially shown that this sensorial stimulation can have an

impact on a cognitive deficit such as neglect, the present

study provides the first evidence of its impact on an internal

representation.

One study has shown effects of PA upon the bisection of

number intervals (Rossetti et al., 2004). Importantly, no visuo-

motor adaptation is required in the present experiment, sug-

gesting that the correcting mechanism could be perceptual

(differing from Rossetti et al., 2004). Rossetti’s study employed

PA to compensate for the right bias, which in turn was

translated into better performance on number bisection. On

the contrary, in our study the compensation exerted by the

perception of coherent motion cannot be explained by a link

based on action between number and space. As we described

before, the characteristics of our stimuli and the experimental

setup prevented eye movements to track the dots in the RDKs.

In general, the present study extends what has been sug-

gested by Rossetti and collaborators and other studies

regarding number representation in neglect (Zorzi et al., 2002;

Priftis et al., 2006; Vuilleumier et al., 2004), and signals once

more the parallels between numbers representation and

space. Fundamentally it shows that similar attentional

processes operating on external stimuli operate also on the

number representation (Dehaene et al., 2003). This conclusion

is strongly supported by our findings. Note that in some cases

electrophysiological responses have been recently found for

internal and external cueing (Salillas et al., 2008).

Finally, the present results further extend the link between

external stimulation and attention to an internal representa-

tion, agreeing with the idea of attention based and represen-

tation based accounts of neglect are complementary. Our data

show altered numerical representation in neglect patients,

signaled both by the altered distance effect and by the effect of

numerical size (Bisiach and Luzzatti, 1978). The mechanism of

recovery by external leftward motion has an attentional

source. Therefore, the present data also relate to attention

based models. According to Kinsbourne’s (1987, 1993) model,

each hemisphere shifts attention towards the contralateral

visual hemispace by inhibiting the other hemisphere. In other

words, each hemisphere generates a vector directing atten-

tion in the horizontal plane towards the opposed direction.

The vector from the left hemisphere is of higher intensity than

the vector from the right hemisphere and each vector inhibits

the other. In this framework, shifts of attention caused by

leftward motion could be operating as a compensating hori-

zontal vector to the left. Rightward motion, instead, respects

the vectors system but fails in redirecting attention. On the

contrary, it redirects attention in the direction of the strongest

overwhelming vector in neglect. Our data show comparison

times to numbers on the left of reference number 5 tended to

be the slowest in this condition. In addition, other studies of

OKS, although not all of them, have shown negative effects of

rightward motion (see these different outcomes in Pizzamiglio

et al., 1990; Vallar et al., 1993, 1995; Karnath, 1996). The

absence of coherent motion in the random motion condition

leads to the most altered distance effect, which may indicate

disruption of the vector system in attention to the numbers

representation. Other attention based models may also

explain the data. For example, the model proposed by Heil-

man et al. (1993) assumes differences between the brain

hemispheres in their involvement of attention, but, unlike in

Kinsbourne’s model, the bias is not due to a hyperactive left

hemisphere. Rather it assumes asymmetric capacities in the

exploration of space. The right hemisphere has a special role

in space-related behavior, with orientation of attention to

both hemispaces, while the left hemisphere orients unilater-

ally to the right. Thus, neglect would result from hypoarousal

of the right hemisphere. Leftward motion might reactivate the

hypoactive right hemisphere, by redirecting attention to the

left hemi-field (see Chokron et al., 2007). In this case, however,

rightward motion should lead to the worst case scenario, our

data shows a tendency in this direction where the highest

comparison times to number ‘‘4’’ were present with rightward

motion. Finally, leftward motion would entail compensation

in the failure to disengage from ipsilesional stimuli, as

proposed by Posner et al. (1984). The latter two models

however, would not explain the effects of non-coherent

motion in the distance effect in our data.

This link between attention based and representation

based accounts is consistent for the few sensory stimulation

methods that have had an impact in representational neglect

(vestibular stimulation in Geminiani and Bottini, 1992; Rode

and Perenin, 1994; transcutaneous electrical nervous stimu-

lation in Guariglia et al., 1998). In other words, our data clearly

indicate that the external redirection of attention by direc-

tional motion has an effect on an altered direction of attention

with regard to internal representations. An interesting ques-

tion would be to explore whether R-OKS could make repre-

sentational recovery longer lasting as has been shown in other

tasks.

Acknowledgements

We wish to thank Professor Paolo Di Benedetto and Doctor

Emanuele Biasutti of the Istituto di Medicina Fisica e Riabili-

tazione ‘‘Gervasutta’’, Azienda per i Servizi Sanitari, no. 4

c o r t e x 4 5 ( 2 0 0 9 ) 7 3 0 – 7 3 7 737

‘‘Medio Friuli’’, Italy for their support throughout this study.

We are also grateful to Nicole Wicha and other three anony-

mous reviewers for their valuable comments on this manu-

script. The present research has been supported by the

European Union Marie Curie Action Contract ‘‘NUMBRA’’

504927 to Carlo Semenza.

r e f e r e n c e s

Ansorge U and Wuhr PA. response-discrimination account of theSimon effect. Journal of Experimental Psychology: HumanPerception and Performance, 30: 365–377, 2004.

Berti A, Ladavas E, and Della Corte M. Anosognosia forhemiplegia, neglect dyslexia, and drawing neglect: clinicalfindings and theoretical considerations. Journal of theInternational Neuropsychological Society, 2: 246–440, 1996.

Bisiach E and Luzzatti C. Unilateral neglect of representationalspace. Cortex, 14: 129–133, 1978.

Bisiach E, Vallar G, Perani D, Papagno C, and Berti A. Unawarenessof disease following lesions of the right hemisphere:anosognosia for hemiplegia and anosognosia for hemianopia.Neuropsychologia, 24: 471–482, 1986.

Bosbach S and Prinz W. A Simon effect with stationary movingstimuli. Journal of Experimental Psychology: Human Perception andPerformance, 30: 39–55, 2004.

Chokron S, Dupierrix E, Tabert M, and Bartolomeo P.Experimental remission of unilateral spatial neglect.Neuropsychologia, 45: 3127–3148, 2007.

Dehaene S. Varieties of numerical abilities. Cognition, 44: 1–42, 1992.Dehaene S, Piazza M, Pinel P, and Cohen L. Three parietal circuits of

number processing. Cognitive Neuropsychology, 20: 487–506, 2003.Delazer M, Girelli L, Grana A, and Domahs F. Number processing

and calculation-normative data from healthy adults. ClinicalNeuropsychology, 18: 76–88, 2003.

Denny-Brown D and Banker BQ. Amophosyntesis from left parietallesions. Archives of Neurology and Psychiatry, 71: 302–313, 1954.

Diller L and Weinberg J. Hemi-inattention in rehabilitation: theevolution of a rational remediation program. Advances inNeurology, 18: 63–82, 1997.

Edgeworth J, Robertson IH, and McMillan T. Balloons Test. ThamesValley Test Company, 1998.

Gauthier L, Dehaut F, and Joanette Y. The bells test: a quantitativeand qualitative test for visual neglect. International Journal ofClinical Neuropsychology, 11: 49–54, 1989.

Geminiani G and Bottini G. Mental representation and temporaryrecovery from unilateral neglect after vestibular stimulation.JournalofNeurology,Neurosurgery,and Psychiatry, 55:332–333,1992.

Gobel S, Calabria M, Farne A, and Rossetti Y. Parietal rTMSdistorts the mental number line: stimulating ‘‘spatial’’ neglectin healthy subjects. Neuropsychologia, 44: 860–868, 2006.

Guariglia C, Lippolis G, and Pizzamiglio L. Somatosensorystimulation improves imagery disorders in neglect. Cortex, 34:233–241, 1998.

Heilman KM. Neglect and related disorders. In Heilman KM andValenstein E (Eds), Clinical Neuropsychology. New York: OxfordUniversity Press, 1979: 268–307.

Heilman KM, Watson RT, and Valenstein E. Neglect and relateddisorders. In Heilman KM and Valenstein E (Eds), ClinicalNeuropsychology. New York: Oxford University Press, 1993.

Karnath HO. Optokinetic stimulation influences the disturbedperception of body orientation in spatial neglect. Journal ofNeurology, Neurosurgery, and Psychiatry, 60: 217–220, 1996.

Kerkhoff G, Schindler I, Keller I, and Marquardt C. Visualbackground motion reduces size distortion in spatial neglect.Neuroreport, 10: 319–323, 1999.

Kerkhoff G. Multiple perceptual distortions and their modulation inleft sided visual neglect. Neuropsychologia, 38: 1073–1086, 2000.

Kerkhoff G. Modulation and rehabilitation of spatial neglect bysensory stimulation. Progress in Brain Research, 42: 257–271, 2003.

Kerkhoff G. Repetitive optokinetic stimulation induces lastingrecovery from visual neglect. Restorative Neurology andNeuroscience, 24: 357–369, 2006.

Kinsbourne M. Mechanisms of unilateral neglect. In Jeannerod(Ed), Neurophysiological and Neuropsychological Aspects of SpatialNeglect. New York: North Holland, 1987: 69–86.

Kinsbourne M. Orientational bias model of unilateral neglect:Evidence from attentional gradients within hemispace. InRobertson IH and Marshall JC (Eds), Unilateral Neglect: Clinicaland Experimental Studies. Psychology Press, 1993: 63–86.

Mattingley JB, Bradshaw JL, and Bradshaw JA. Horizontal visualmotion modulates focal attention in left unilateral spatialneglect. Journal of Neurology, Neurosurgery, and Psychiatry, 57:1228–1235, 1994.

Moyer RS and Landauer TK. The time required for judgments ofnumerical inequality. Nature, 215: 1519–1520, 1967.

Pizzamiglio L, Frasca R, Guariglia C, Inoccia C, and Antonucci G.Effect of optokinetic stimulation in patients with visualneglect. Cortex, 26: 541–554, 1990.

Priftis K, Zorzi M, Meneghello F, Marenzi R, and Umita C. Explicitvs. implicit processing of representational space in neglect:dissociations in accessing the mental number line. Journal ofCognitive Neuroscience, 18: 680–688, 2006.

Priftis K, Piccione F, Giorgi F, Meneghello F, Umilta C, and Zorzi M.Lost in number space after right brain damage: a neuralsignature of representational neglect. Cortex, 44: 449–453, 2008.

Posner MI, Walker JA, Friedrich FJ, and Rafal RD. Effects of parietalinjury on covert orienting of attention. Journal of Neuroscience,4: 1863–1974, 1984.

Rode G and Perenin MT. Temporary remission of representationalhemineglect through vestibular stimulation. Neuroreport, 5:869–872, 1994.

Rossetti Y, Jacquin-Curtois S, Rode G, Ota C, Michel C, andBoisson D. Does action make the link between number andspace representation? visuo-manual adaptation improvesnumber bisection in unilateral neglect. Psychological Science, 15:426–430, 2004.

Salillas E, EL-Yagoubi R, and Semenza C. Sensory and cognitiveprocesses of shifts of spatial attention induced by numbers: anERP study. Cortex, 44: 406–413, 2008.

Salillas E and Semenza C. Effects of attention to motion onnumber processing. Behavioral and ERP evidence. Inpreparation. Part of these data were presented at theNUMBRA/ESCOP Summer School ‘‘Neuroscience of NumberProcessing’’, Erice, Sicily, Italy, 3–10 July 2005.

Shadlen MN and Newsome WT. Neural basis of a perceptualdecision in the parietal cortex (area LIP) of the rhesus monkey.Journal of Neurophysiology, 86: 1916–1936, 2001.

Vallar G, Antonucci G, Guariglia C, and Pizzamiglio L. Deficits inposition sense, unilateral neglect and optokinetic stimulation.Neuropsychologia, 31: 1191–1200, 1993.

Vallar G, Guariglia C, Magnotti L, and Pizzamiglio L.Optokinetic stimulation affects both vertical and horizontaldeficits in position sense in unilateral neglect. Cortex, 31:669–683, 1995.

Vallar G, Guariglia C, Nico D, and Pizzamiglio L. Motor deficits andoptokinetic stimulation in patients with left hemineglect.Neurology, 49: 1364–1370, 1997.

Vuilleumier P, Ortigue S, and Brugger P. The number space andneglect. Cortex, 40: 399–410, 2004.

Wilson B, Cockbum J, and Halligan P. Development of a behavioraltest of visuospatial neglect. Neurology, 23: 658–664, 1987.

Zorzi M, Priftis K, and Umilta C. Brain damage: neglect disruptsthe mental number line. Nature, 417: 138–139, 2002.