Dissecting the component deficits of perceptual imbalance in visual neglect: Evidence from horizontal–vertical length comparisons

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Research report

Dissecting the component deficits of perceptual imbalancein visual neglect: Evidence from horizontalevertical lengthcomparisons

Pom Charras a,b,*, Juan Lupianez a, Raffaella Migliaccio b,c, Monica Toba b,Pascale Pradat-Diehl d, Christophe Duret e and Paolo Bartolomeo b,c

aUniversidad de Granada, Spainb Inserm-UPMC UMRS 975, Hopital Pitie-Salpetriere, Paris, FrancecDepartment of Psychology, Catholic University, Milan, ItalydAPHP e Hopital Pitie-Salpetriere Service de Medecine Physique et Readaptation, Paris, FranceeService de Reeducation et de Readaptation Fonctionnelle, Clinique Les Trois Soleils, Boissise le Roi, France

a r t i c l e i n f o

Article history:

Received 17 August 2010

Revised 22 September 2010

Accepted 13 January 2011

Action editor Robert McIntosh

Published online 3 February 2011

Keywords:

Neglect

Left underestimation

Right overestimation

Pre-attentive mechanisms

Symmetry

a b s t r a c t

Introduction: Signs of left unilateral neglect often occur after damage to the right hemi-

sphere and entail a lefteright imbalance in stimulus processing. Typically, neglect patients

deviate rightward when bisecting lines. An underestimation of the left portion of the line

and/or a right overestimation could explain this effect.

Methods: Here we dissected their respective contribution by asking participants to compare

a vertical segment to a horizontal segment, either on the left or on the right. We also tested

whether neglect patients exhibited the symmetry law, whereby normal participants

underestimate symmetrically bisected lines as compared to asymmetrically bisected lines.

Results: Controls and patients underestimated symmetric figures. Depending on the degree

of lefteright horizontal competition, neglect patients underestimated left-sided stimuli or

overestimated right-sided stimuli.

Conclusions: We concluded that two independent deficits contribute to neglect signs:

a deficit in attentional orienting to the left, which can be worsened by left hemianopia,

perhaps depending on impaired functioning of right-hemisphere attentional networks,

and a tendency for attention to be captured by right-sided stimuli, possibly resulting from

the activity of an isolated left hemisphere. Finally, the symmetry law was preserved in

neglect patients, and thus appears to be driven by pre-attentive mechanisms.

ª 2011 Elsevier Srl. All rights reserved.

1. Introduction

Patients with right posterior brain damage often show signs of

left visual neglect, which typically entail a lefteright imbalance

in stimulus processing (Vallar, 2001; Bartolomeo, 2007; Mort

et al., 2003). For example, when asked to bisect horizontal

lines, neglect patients shift the subjectivemidpoint towards the

right (Schenkenberg et al., 1980). This effect has often been

* Corresponding author. Departamento de Psicologıa Experimental y de Fisiologıa del Comportamiento, Campus de Cartuja, S/N. C.P.18071, Granada, Spain.

E-mail address: [email protected] (P. Charras).

Available online at www.sciencedirect.com

Journal homepage: www.elsevier.com/locate/cortex

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2

0010-9452/$ e see front matter ª 2011 Elsevier Srl. All rights reserved.doi:10.1016/j.cortex.2011.01.008

Author's personal copy

accounted for by neglect of the left-sided part of the line.

Patients seemto ignore theextreme left part of the lineand thus

underestimate its length ascompared to the right. But left-sided

impairment can also be accompanied by a bias towards right-

sided objects (D’Erme et al., 1992; Gainotti et al., 1991; Natale

et al., 2005; Bartolomeo and Chokron, 2002) that might also

contribute to the lefteright imbalance. For example, right-sided

objects exert a sort of “magnetic attraction” on patients’ gaze

and attention (Gainotti et al., 1991; see alsoDeRenzi et al., 1989),

as if they automatically captured attention (Gainotti et al., 1991;

Bartolomeo and Chokron, 1999; Marshall and Halligan, 1989). A

deterioration of the processing of left-sided stimuli by themere

presence of right-sided stimuli also occurs in visual extinction

(Baylis et al., 1993; Ward and Goodrich, 1996; Humphreys et al.,

1994). Therefore, even though the ipsilesional visual field is

sometimes considered as being “intact”, it is important to

underline that right-sided stimuli are not correctly or normally

processed, but can rather be over-enhanced or over-prioritized

(Snow and Mattingley, 2006, 2008; Kinsbourne, 1993;

Bartolomeo and Chokron, 1999; Bartolomeo et al., 1999).

Consistent with this proposal, the rightward bias in line bisec-

tion can be reduced, although not eliminated, by cuing the left

endpoint of the line (Heilman and Valenstein, 1979; Riddoch

and Humphreys, 1983; Bonato et al., 2008). The fact that cuing

the left endpointof the linedoesnoteliminatebutonlyweakens

the rightward bias, suggests that the lefteright imbalance

cannot rely exclusively on neglect of the left part of the line.

Consequently, the role of the right part in the global lefteright

imbalance has to be reconsidered (Ishiai et al., 2000; see also

McIntosh et al., 2005). Recently, Urbanski and Bartolomeo (2008)

focused on the role of the right portion in the lefteright imbal-

ance by using an imaginal endpoint task in which patients had

to place the left and right endpoints of an imaginary line for

which the centre was provided. To test the importance of the

rightpart, theorder ofmarking (left-to-rightor right-to-left)was

manipulated: participants had to place first the left endpoint

and then the right, or vice-versa. Their results clearly showed

that although the typical rightwards error persisted in the right-

to-left condition, it was eliminated or even reversed in the left-

to-right condition (i.e., when patient first drew the left endpoint

and then the right one). This finding suggests that right over-

processing in the framework of a competition between right-

and left-sided objects might be an important determinant of

lefteright imbalance in neglect.

More recently, Charras et al. (2010) proposed a novel

methodological approach to separately assess the relative

roles of left underestimation and right overestimation in the

final lefteright imbalance, characteristic of neglect. They used

a line extension task, in which a vertical line was centrally

displayed on a sheet of paper. Participants were asked to trace

a horizontal segment, perpendicular to the vertical (L shape)

either on the left or right side. Thus, it was possible to disso-

ciate the relative roles of left underestimation and right

overestimation in the lefteright imbalance. The results

showed that performance of neglect patients was biased on

the left side, consistent with a left underestimation, but was

identical to controls on the right side.

In the present study, our first objective was to replicate this

result using a line length comparison task in which partici-

pants had to judgewhich one of the two linese vertical versus

horizontal e was the longest. Vertical lines were always cen-

trally presented, while horizontal lines were displayed either

to the left or to the right of the vertical line. Therefore,

a vertical line was either compared to a left-sided horizontal

line to test the left underestimation (see Fig. 1, left L) or to

a right-sided horizontal line to test the right overestimation

(see Fig. 1, right L). On the basis of the results of Charras et al.

(2010), we expected to observe left underestimation rather

than right overestimation. Note that in these two L configu-

rations, there is no lefteright competition, as the horizontal

line is either displayed towards the left or the right of the

vertical line. A further aim was to induce a lefteright

competition in the configurations. To do so, we added

configurations in which the vertical line remained centrally

presented but horizontal lines were bisected, such that one

part falls in the left side of space and the other in the right side

(see left and right Ts and Xs1 in Fig. 1). Previous studies with

normal participants revealed that vertical lines are slightly

overestimated in the left and right Ts as compared to hori-

zontal lines, but horizontal lines are overestimated in the left

and right Xs as compared to vertical lines (Charras and

Lupianez, 2009, 2010). Therefore, we were able to test the

role of lefteright competition in the case of vertical emphasis

(in the T configurations), or of horizontal emphasis (in the X

configurations). Note that the vertical or horizontal emphasis

was assumed to directly result from the vertical

Fig. 1 e It represents the eight types of figure used in the

experiment. The L configurations enabled to test the left

and right biases separately. In the T and X configurations,

we introduce a lefteright competition by bisecting the

horizontal line (either symmetrically or asymmetrically).

Importantly, in the T configuration, the vertical line was

overestimated; while the X configuration, the horizontal

line was overestimated.

1 Note that the terminology “X configuration” refers to the crossshapes as depicted in Fig. 1.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2 541

Author's personal copy

overestimation in the T configurations or from the horizontal

overestimation in the X configurations.2

In two further configurations, the centre T and X, the hori-

zontal line was symmetrically bisected. In the visual percep-

tion literature, it is well known that symmetry drastically

influences line length estimation. The horizontal vertical illu-

sionwith so-called “centreT” stimuli (see Fick, 1851; Kunnapas,

1955, 1957; Finger and Spelt, 1947; Avery and Day, 1969) best

demonstrates this kind of perceptual distortion (Fig. 1).

In this configuration the horizontal line is perceived as

being shorter than the vertical line, even though they are of

the same physical length. This horizontal underestimation is

mainly caused by the symmetric bisection of the horizontal

line. Indeed, symmetric bisections are underestimated

(Kunnapas, 1955; Wolfe et al., 2005; Charras and Lupianez,

2009). But importantly, as soon as the two lines are symmet-

rically bisected (see centre X in Fig. 1), the visual distortion is

abolished, and none of the lines is overestimated by normal

participants (Charras and Lupianez, 2010).

Therefore, the present study also aimed at testing whether

neglect patients, like controls, would underestimate the hori-

zontal line in the centre T configuration, but would show no

bias in the centre X, thus suggesting that the symmetry law

might rely on different mechanisms, perhaps at a lower pro-

cessing level than those implicated inpatients’ attentional bias.

Further issues of interest include the anatomical bases of

biased length estimation. The lesional bases of biased line

bisection have often been referred to posterior brain lesions

(Binder et al., 1992; Rorden et al., 2006). Thiebaut de Schotten

et al. (2005) were recently able to directly probe the anatom-

ical bases of rightward deviation on line bisection by using

electrical cortical and subcortical stimulation in human

neurosurgical patients. They found small but significant

deviations with temporary inactivation of the supramarginal

gyrus of the parietal lobe and of the temporaleparietal junc-

tion; however, the most dramatic deviations occurred in one

patient upon temporary disconnection of a branch of the

superior longitudinal fasciculus (SLF), consistent with

previous evidence from stroke patients (Doricchi and

Tomaiuolo, 2003). In further work using diffusion tensor

imaging (DTI) tractography in stroke patients, damage to the

inferior fronto-occipital fasciculus (IFOF) was also found to be

present in neglect patients, but absent in patients with similar

cortical lesions in the right hemisphere and no signs of neglect

(Urbanski et al., 2008). In the present study, we obtained high-

definition magnetic resonance imagings (MRIs) for all the

patients. Lesions were mapped and localized in relation both

to cortical and subcortical grey matter structures (Rorden

et al., 2007) and to the course of major white-matter fasciculi

(Catani and Thiebaut de Schotten, 2008).

The present study had two principal aims: (1) to disen-

tangle the relative contribution of left underestimation and

right overestimation to line bisection errors in visual neglect;

(2) to explore the integrity of early perceptual processes in

neglect, such as those implicated in the symmetry law. We

used a new methodology, which allowed us to separately test

and quantify left and right biases by comparing vertical lines

to either left or right horizontal lines (see Charras et al., 2010).

2. Methods

2.1. Participants

Eight right-handed patients with unilateral lesions in the right

hemisphere [seven men; mean age, 57 years; standard devi-

ation (SD), 14.5] participated in this study. All of them showed

signs of moderate to severe left unilateral spatial neglect on

a battery of paper-and-pencil tests including the following

standard clinical tests: line cancellation (Albert, 1973), bell

cancellation (Gauthier et al., 1989), letter cancellation and line

bisection (Azouvi et al., 2002). Four patients had intact visual

fields on confrontation; the remaining four had left homony-

mous hemianopia (HLH), confirmed by Goldmann perimetry

in two cases. Clinical and anatomical data for each patient are

represented in Table 1, with the detailed description of both

grey and white matter lesion involvement. Eight right-handed

healthy participants without neurological impairment (seven

men; mean age, 54 years; SD, 12) also consented to participate

in the experiment.

2.2. Brain imaging and lesion analysis

Brain MRI scans included T1, T2, fluid attenuated inversion

recovery, and diffusion images obtained with standard

parameters on a 3T General Electric scanner. Lesion extent

was determined for each patient by manually drawing the

lesion borders images directly onto the original 3D T1 MRI, by

using a 325� 203 mm pen tablet (Wacom Intous4 large, http://

www.wacom.com). To reduce lesion-induced registration

errors, spatial normalization was performed using a mask

that excluded the damaged areas of the brains, thereby pre-

venting these areas from biasing the transformation (Brett

et al., 2001). The 3D brain scan and lesion volume were then

normalized to the standard Montreal Neurological Institute

(MNI) template in Statistical Parametric Mapping-5 (http://

www.fil.ion.ucl.ac.uk/spm) running under Matlab 7.5 (http://

www.mathworks.com). After normalization, the brain lesion

was segmented and its borders were redefined in the

normalized brain. MRIcron software (www.mricro.com/

mricron; Rorden and Brett, 2000) was used to estimate the

lesion and to label the lesion sites based on the Automated

Anatomical Labelling map volume (Tzourio-Mazoyer et al.,

2002, see their Table 1). Furthermore the MRIcron software

was applied to generate group-specific lesion overlay plots for

all patients (Fig. 4A) and subtraction plots displaying voxels

selectively involved in patients with hemianopia (all neglect

patients minus patients without hemianopia, see Fig. 4B).

Finally, in order to identify the involvement of specific white

matter fiber tracts, recent work by Catani and Thiebaut de

Schotten (2008) based on DTI-tractography on normal partic-

ipants, was used as a reference. This study is in the same

space as the MNI reference brain. The overlapping between

the lesion map and specific white-matter tracts in the right

hemisphere was visually assessed. The tracts we considered

2 In the present manuscript, the under or overestimations ofthe vertical or horizontal lines are not absolute, but ratherrelative and highly dependent of the verticalehorizontalcomparison process at hand in the task.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2542

Author's personal copy

Table

1e

Itprese

nts

thedem

ogra

phical,clin

icalandanatom

icaldata

ofpatients

withth

eir

perform

ance

onvisuosp

atialtests.

PatientGender/

age,y

Onse

tof

Illnes

s(m

onth

s)

Visual

field

defect

Greymatterlesionsite

White

matter

lesion

site

Lesion

volume

(cm

3)

Etiology

Linebisec

tion

5cm

(mm

of

rightw

ard

deviation)

Linebisection

20cm

(mm

of

rightw

ard

deviation)

Line

cance

llation

(max30/30)

Bell

cance

llation

(max15/15)

Letter

cance

llation

(max30/30)

HUH

M/45

12

LHH

Inferiorandmiddle

temporalgyri,su

perior

temporalpole,hippoca

mpusand

parahippoca

mpalarea,insu

la,prece

ntral/

postce

ntralco

rtex,ca

udate

nucleus,

putamen,th

alamus,

calcarineco

rtex

(BA

18e19)

IFOF,

ILF&

OR

90.02

H�3

19

30/29

8/15

15/27

HAA

M/75

5LHH

Superiorandmiddle

temporalgyri,insu

la,

rolandic

opercolum,putamen,ca

udate

nucleus,

middle

occ

ipitalgyru

s(BA

19)

SLF,IFOF,

ILF&

OR

23.45

I.5

17.5

30/29

9/14

4/23

ZOJ

F/76

4LHH

Fusiform

andlingualgyri,ca

lcarine

cortex(BA

17e18)

IFOF

14.5

I3

46

29/30

13/15

28/25

FOE

M/47

1LHH

Superiorandmiddle

temporalpole,

superiorandmiddle

temporalgyri,

insu

la,putamen,th

alamus,

hippoca

mpus

andparahippoca

mpalarea,inferior

parietalco

rtex,angularand

supramarginalgyri,middle

occ

ipital

gyru

s(BA

18)

SLF,IFOF,

ILF&

OR

162.71

H�7

26

29/29

5/14

18/27

LOA

M/52

4None

Rolandic

operculum,middle

temporal

pole,inferior,

middle

andsu

ptemporal

gyri,parahippoca

mpalarea,insu

la,

putamen,su

pramarginalandangulargyri

SLF&

IFOF

95.92

I1.5

3.5

30/30

12/15

28/30

SSA

M/43

4None

Putamen

IFOF&

OR

17.3

H4.5

3.5

27/30

0/11

0/14

MAF

M/44

7None

Inferior,

middle

andsu

periorfrontalgyri,

orb

itofrontalco

rtex,prece

ntral/postce

ntral

cortex,motorandsu

pplementary

motor

areas,

parace

ntrallobule,insu

la,amygdala,

superior,

middle

andinferiortemporal

gyri,insu

la,su

periorandinferiorparietal

areas,

anteriorandmiddle

cingulate,

putamen,ca

udate

nucleus

SLF,IFOF,

ILF&

OR

522.82

I.2

1.9

20/27

12/11

12/30

COM

M/70

3None

Putamen,ca

udate

nucleus

IFOF

6.38

I2

11

//15/15

29/30

Abbreviations:

H¼hemorrhagic,I¼

isch

aemic.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2 543

Author's personal copy

were the optic radiations (OR), whose disconnection typically

results in contralateral hemianopia, and tracts whose

disconnection has been implicated in the occurrence of

neglect signs, such as the SLF (see Doricchi and Tomaiuolo,

2003; Bartolomeo et al., 2007), the IFOF (Urbanski et al., 2008)

and the inferior longitudinal fasciculus (ILF; Bird et al., 2006).

2.3. Stimuli and procedure

Participants sat in front of a laptop screen (12.1 inches) at an

approximate distance of 50 cm.

E-Prime software (Schneider et al., 2002) controlled the

presentation of the stimuli, timing operations, and data

collection. At the beginning of each trial, the word “Pret?”

(“Ready?” in French) appeared at the centre of the screen and

the trial began after pressing the space bar. A fixation point (a

white asterisk) was displayed for 1000 msec on the black

background and was followed by the target stimulus, a white

configuration composed of a horizontal line and a vertical line.

Target presentation time decreased from block to block,

ranging from 3000 to 150 msec (3000 msec, 1000 msec,

500 msec 150 msec and 150 msec), depending on performance

at judging the centre X. If accuracy for the centre X configu-

ration was superior or equal to 80%, then presentation time

decreased (note that all participants’ performance for the

centre Xwas good enough to reach the 5th and last block). The

task was to judge which one of the two perpendicular lines

was longer. The participants vocally reported the answer

ehorizontal or verticale and the “experimenter” pressed the

corresponding button on a computer mouse. Participants

were instructed to respond as accurately as possible.

The horizontal line was always of the same standard

length (41 mm or 4.7�) whereas the vertical line could vary

from 22.5 mm (2.6�) to 59.5 mm (6.8�). From a standard vertical

length line (41 mm), several lengths were calculated, ranging

from 55% to 145% of the standard line. In the 1st and 2nd

blocks of trials, horizontal length varied from 55% to 145% in

six steps (55, 75, 95, 105, 125 and 145%). Depending on accu-

racy, task difficulty increased so that horizontal line in the 3rd

block ranged from 75% to 125% (75, 85, 95, 105, 115 and 125%).

As previously mentioned, all participants appropriately per-

formed the task so that task difficulty increased. All the lines

subtended .8� in width. The vertical line was always centrally

presented. Although the horizontal line was always presented

perpendicularly to the vertical one, its position along the

vertical line varied and provided three types of configuration.

The L configurations: in these two configurations, the two

lines form an L shape, the horizontal line being displayed

either to the left or the right of the vertical line (left L or right L

respectively, see Fig. 1).

The T configurations: the horizontal line was displayed at

the bottom of the vertical line. In one configuration, the

horizontal line was bisected in the middle by the vertical line

(centre T). And in the two remaining configurations, the

horizontal line was asymmetrically bisected. The largest part,

which represented 3/4 of the total length, could be either on

the left (left T) or on the right (right T, see Fig. 1).

The X configurations: in the configuration called “centre X”,

the two lines crossed in their middle, thus forming a cross

shape. When the horizontal line was displaced towards the

left, the configuration was the “left X”, while a right

displacement led to the “right X” (see Fig. 1).

2.4. Design

The present experiment had two within participant factors:

Configuration and Vertical line length. Vertical line length

could take six different values, three inferior and three supe-

rior to the standard horizontal line (see section Stimuli for

further details). We used eight configurations: two L configu-

rations (left L and right L); three T configurations (left T, centre

T, right T) and three X configurations (left X, centre X, right X).

The current experiment was divided in four sessions of 120

trials (total of 480 trials). Participants performed eight practice

trials before proceeding with every experimental session. The

eight configurations were judged 60 times each. Every exper-

imental session had five blocks, time presentation (3000, 1000,

500, 150 and 150 msec) and length range decreasing with

accuracy across blocks (between 55% and 145% in blocks 1 and

2; between 75% and 125% in blocks 3, 4 and 5).

3. Data analysis

The method of the constant stimuli was used. One stimulus

considered as standard remained identical in length in all

trials (here the horizontal line), while the comparative stim-

ulus varied (here the vertical line). From the six possible

lengths of the vertical line, we computed the point of subjec-

tive equality (PSE). This measure represents the estimated

length of the vertical line in order for it to be perceived as

identical to the standard horizontal line. For the sake of

simplicity, the PSE was transformed into a percentage of

illusory extent. So, we obtained a percentage of under or

overestimation for the eight configurations. Note that the

closer to zero the value is, the less biased length perception is.

Negative values refer to horizontal underestimation whereas

positive ones refer to horizontal overestimation.

4. Results

Data were analyzed separately as a function of Side (centre,

left, right) and submitted to Group (Controls vs Neglect) by

Configuration (L, T, X) mixed analysis of variances (ANOVAs).

We also conducted planned single sample t-tests to test the

reliability of the under and overestimations.

4.1. Line length perception

4.1.1. Centred horizontal line (centre T and X)We performed the first analysis for the two configurations in

which the horizontal line was displayed in the middle (centre

T and centre X) to test whether as already observed with

normal participants, the horizontal line was underestimated

in the centre T but well estimated in the centre X (Charras and

Lupianez, 2009, 2010). The main effect of Group was far from

significance, F(1,14)< 1, suggesting that controls and patients

performed in a similar way. There was a main effect of

Configuration (T vs X), F(1,14)¼ 156.7, p< .001. Participants

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2544

Author's personal copy

underestimated the horizontal line in the centre T but not in

the centre X. Group and Configuration interacted, F(1,14)¼7.42, p¼ .02, because neglect patients showed a tendency to

underestimate more the horizontal line in the centre T than

controls, F(1,14)¼ 3.39, p¼ .09. Planned comparisons

confirmed that both patients and control participants under-

estimated the horizontal line in the centre T but not in the

centre X [t(7)¼ 10.48, p< .001 and t(7)¼�9.31, p< .001 for the

centre T; t(7)¼ .58, p¼ .57, ns and, t(7)¼�1.05, p¼ .32, ns for

the centre X respectively for control participants and neglect

patients].

4.1.2. Left-sided horizontal lineWe conducted an analysis with the items with leftward-

oriented horizontal lines in the three configurations (L, T, X).

There were principal effects of Group, F(1,14)¼ 9.02, p¼ .009,

as well as of Configuration, F(2,28)¼ 21.73, p< .001. The two

factors did not interact, F(2,28)¼ 1.48, p¼ .24, suggesting that

neglect patients, as compared to controls, underestimated the

length of the left-sided horizontal line, independently of the

configuration. This result was highly expected on the basis of

neglect patients’ tendency to ignore left-sided items. The

effect of Configuration resulted from the horizontal line being

estimated as longer in the left X, than in the left T, F(1,14)¼47.69, p< .001, consistentwith the symmetry law, according to

which an asymmetrically bisected line (here the horizontal

line) is overestimated as compared to a symmetrically bisec-

ted line (the vertical line). Importantly, the symmetry law

influenced in a similar way the performance of both patients

and control participants (see for example in the centre and

right panels of Fig. 2, the difference in length perception

between the T and X configurations).

4.1.3. Right-sided horizontal lineFor the three configurations (L, T, X) inwhich thehorizontal line

was right-sided, there were a main effect of Group [F(1,14)¼4.62, p¼ .05], a main effect of Configuration [F(2,28)¼ 46.83,

p< .001] and a highly significant interaction [F(2,28)¼ 5.96,

p¼ .007]. This suggests that even though both groups were

sensitive to line symmetry, neglect patients had anexaggerated

horizontal overestimation in the right X. Planned comparisons

confirmed that patients and control participants had similar

performance for the right L andrightT [F(1,14)¼ 1.43,p¼ .25 and

F(1,14)< 1, p¼ .82], but not for the right X, where neglect

patients showed an enhanced horizontal overestimation

[F(1,14)¼ 10.92, p¼ .005].

4.2. Effect of left HLH on line length perception

Half of the neglect patients in our study presented visual field

deficits. We performed an additional statistical analysis in

which we investigated the effect of visual field defects on line

estimation. It has long been known that left HLH dramatically

increases patients’ rightwards error in line bisection

(Bartolomeo, 1987; Doricchi and Angelelli, 1999), presumably

because at each fixation it nullifies the perceptual salience of

the left part of the line, thus further increasing the biased

competition in favour of the right part (Bartolomeo et al., 2003).

We compared the performance of patients with and without

hemianopia to test the competition hypothesis. First, we con-

ducted a repeated-measures ANOVA with the factors Hemi-

anopia [neglect patients with hemianopia (NHþ) vs neglect

patients without hemianopia (NH�) and Configuration for

centred horizontal lines (centre T and centre X) to test whether

NH� and NHþ patients performed differently when the hori-

zontal line was symmetrically bisected. Hemianopia did not

influence performance [F(1,6)< 1, p¼ .73, ns], but the configu-

ration did [F(1,6)¼ 78.01, p< .001]. Importantly, the interaction

Hemianopia�Configurationwas far fromsignificance [F(1,6)<

1, p¼ .93], suggesting that both groups of patients under-

estimated the horizontal line in the centre T while showing

relativelyunbiasedperception for the centreX [t(3)¼ .46,p¼ .67

and t(3)¼ .27, p¼ .80 respectively, see left panel of Fig. 3]. Thus,

both NH� and NHþ patients correctly perceived the left-sided

50% of each horizontal line (if not, they would have under-

estimated the horizontal lines length in the centre X).

We then conducted two repeated-measures ANOVAs with

the factors Configuration (L, T or X) and Hemianopia (NHþ vs

NH�) as a function of the side of the horizontal line. For left-

sided lines, there were main effects of Hemianopia [F(1,6)¼17.05, p¼ .006] and of Configuration [F(1,6)¼ 11.82, p¼ .001],

but no interaction [F(1,6)¼ 2.39, p¼ .13]. As expected, NHþpatients showed a stronger bias to underestimate left-sided

lines than NH� patients, and this occurred for the three

configurations. The effect of Configuration showed that the

left underestimation decreased for the left X configuration as

Fig. 2 e It shows the mean percentages of horizontal underestimation (negative values) and/or horizontal overestimation

(positive values) and SDs as a function of Side and Configuration for control participants and neglect patients.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2 545

Author's personal copy

compared to the left T and L. However, for NH� patients, the

horizontal line was eventually overestimated in the left X

[t(3)¼ 4.64, p¼ .02] whilst NHþ patients persisted in

underestimating the horizontal line in this configuration

[t(3)¼�3.90, p¼ .03]. Although this distortion almost reached

40%, it was not clear whether it was due to sensory or

attentional deficits. Indeed, the lack of vertical under or over-

estimation in the centre X indicated that NHþ patients were

able to correctly process the 50%of thehorizontal line thatwas

displayed on the left. Taking into account that in the left X, 75%

of the horizontal line was left-sided (instead of 50% in the

centre X), sensory deficits could only explain 25% of the

Table 2 e It presents the percentages of horizontal under or overestimation as a function of Configuration for the controlgroup and for individual patients without visual defect (NHL) and with HLH (NHD).

lo

rt

no

C

HN

- +H

N

Left L Right L Patient Left L Right L Patient Left L Right L

2 -1

SSA -15 -1 HAA -47 20LOA -5 16 HUH -47 -2MAF -58 -12 FOE -69 23COM 7 15 ZOJ -33 -9

Left T Centre T Right T Left T Centre T Right T Left T Centre T Right T

-8 -27 -8

SSA -22 -50 -12 HAA -45 -35 -2LOA -10 -23 -7 HUH -51 -30 -22MAF -41 -38 -28 FOE -77 -46 2COM -4 -25 7 ZOJ -61 -37 -16

Left X Centre X Right X Left X Centre X Right X Left X Centre X Right X

13 -3 14

SSA 10 -17 4 HAA -18 8 32LOA 30 18 38 HUH -49 -15 29MAF 22 -4 43 FOE -62 2 42COM 16 20 61 ZOJ -27 11 43

Fig. 3 e It shows the percentages of horizontal underestimation (negative values) and/or horizontal overestimation (positive

values) and SDs as a function of Side and Configuration for control participants, neglect patients without hemianopia (NHL)

and neglect patients with hemianopia (NHD).

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2546

Author's personal copy

underestimation out of the 40%. Therefore, visual field deficits

leave15%oftheeffectunexplained.However,singlesample t-test

showed that the underestimation in the left X was not statisti-

cally different from an underestimation of 25% [t(3)¼�1.40,

p¼ .25]. Inotherwords, thisunderestimationwasnotstatistically

different fromwhat would cause the sensory deficits.

For right-sided lines, there was a main effect of Configu-

ration, F(1,6)¼ 33.85, p< .001, but no effect of Hemianopia,

F(1,6)< 1, p¼ .88, and no interaction, F(1,6)< 1, p¼ .95.

Thus, NHþ patients showed a more severe deficit on the

left side than NH� patients did, but they had similar perfor-

mance when the horizontal line was centred or right-sided

(Table 2).

4.3. Precision of line length perception

As the method of the constant stimuli was used, we could

calculate, additionally to the PSE, an indicator of response

precision called the just noticeable difference (JND). The JND

corresponds to the minimum value necessary to discrimi-

nating two stimuli. Consequently, the higher the JND is, the

lower the precision of line length perception is. We conducted

a repeated-measures ANOVA with the factors Group (control,

NH� and NHþ) and Configuration for centred horizontal lines

(centre T and centre X) to test whether response precisionwas

modulated by these factors. As one could expect, we observed

a main effect of Group on the JND: response precision was

better for controls than for NH� or NHþ patients [F(2,13)¼7.65, p¼ .006]. However, no significant difference was

observed between NH� and NHþ patients (F< 1). A main

effect of Configuration was also observed [F(1,13)¼ 17.26,

p¼ .001], indicating that JNDwas smaller for the centre X than

for the centre T. This finding consistently reveals that

unbiased perception (centre X) is more precise than biased

perception (centre T). The interaction between Group and

Configuration was far from significance (F< 1).

Then we conducted another repeated-measures ANOVA

with the factors Group (control, NH� and NHþ) and Configu-

rationbut for left-sidedhorizontal lines (left L, left T and left X).

We observed a main effect of Group [F(2,13)¼ 27.63, p< .0001]

revealing that controls showed a higher precision than NH�patients [even though this effectwas tendencial; F(1,13)¼ 4.07,

p¼ .06], and that NH� showed a higher precision than NHþpatients [F(1,13)¼ 26.87, p¼ .0004]. The effect of Configuration

was not significant [F(2,26)¼ 1.13, p¼ .39] in opposition to the

interaction between these two factors [F(4,26)¼ 2.67, p¼ .054].

The interaction suggests that response precision for NHþpatients drastically decreased when there was a vertical

emphasis [left T as compared the left L or X; F(1,13)¼ 10.12,

p¼ .007]. Finally, we conducted a repeated-measures ANOVA

with the factors Group (control, NH� and NHþ) and Configu-

ration but for right-sided horizontal lines (right L, right T and

right X). We found a main of Group [F(2,13)¼ 12.65, p¼ .0008],

showing that response precision of controls was much higher

than that of NH� and NHþ patients. No difference was

observed between NH� and NHþ patients (F< 1). The effect of

Configuration and the interaction between Group and Config-

urationwere significant [F(2,26)¼ 19.58, p< .0001 and F(4,26)¼5.49, p¼ .002 respectively]. The interaction revealed that for

both NH� and NHþ patients, response precision decreased

when there was a horizontal emphasis, that is when the

horizontal linewas greatly overestimated (rightX as compared

to right T and right L; for NH�, F(1,13)¼ 19.80, p¼ .0006; for

NHþ, F(1,13)¼ 7.11, p¼ .02; for controls, F< 1).

The precision analysis is very consistent with the line

length perception analysis (PSE) in the sense that unbiased

Fig. 4 e Analysis of brain lesions in the two patient groups. (A) Group-specific overlay plots (all patients). (B) Subtraction

analysis displaying voxels selectively involved in patients with hemianopia (all neglect patients e neglect patients without

hemianopia). Results are superimposed on slices of the MNI standard brain. The number of overlapping lesions is indicated

by colour-coding increasing frequencies from light blue to red.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2 547

Author's personal copy

perception is characterized by a higher precision as compared

to biased perception. NH� and NHþ patients generally

showed lower precision than controls, and performance of

these two groups differed when horizontal lines were left-

sided, but not when they were right-sided. In conclusion, the

precision analysis mainly supports the idea that biased

perception is characterized by a decrease in response preci-

sion (that is an increased JND).

4.4. Lesion site

Superimposed brain lesions of all patients are shown in

Fig. 4A. The number of overlapping lesions is colour coded,

with increasing frequencies ranging from light blue (n¼ 1) to

red (n¼ 8). There was no single greymatter region damaged in

all patients. Maximum overlapping lesion area involved

partially the putamen (seven patients out of eight) and the

caudate nucleus (four patients). Within the cortex, damage

partially overlapped on the insula (five patients) and the

temporal cortex (middle temporal gyrus in five patients,

superior and inferior temporal gyri in three patients). In

contrast to these partial grey matter overlaps, all the patients

showed an area of maximum overlap involving the white-

matter bundle of the external capsule, passingmedially to the

insula, and consistent with the location of the anterior portion

of the IFOF (Catani and Thiebaut de Schotten, 2008). An

involvement of the IFOF was also found in all patients when

separately assessed, along with the SLF (four out of eight

patients) and the ILF (four patients) (Table 1).

Fig. 4B shows the results of the subtraction analysis in

order to display the additional damaged areas in neglect

patients showing left hemianopia. As expected, lesions in

these patients overlapped along the course of the OR.

5. Discussion

5.1. Summary of results

The present results provide new evidence concerning the

lefteright processing imbalance in neglect patients. Across all

three configurations (L, T and X), left-sided lines were under-

estimated as compared to right-sided lines. A novel aspect of

our approach allowed us for the first time to analyze and

quantify the relative contribution of left underestimation and

right overestimation to theperceptual lefteright imbalance. For

the L configurations, the lefteright imbalance clearly resulted

from left underestimation, andnot froma right overestimation,

thus replicating the results of Charras et al. (2010) in a line

extension task. Importantly, for these L configurations, there

was no lefteright competition. For the T configurations, in

which the horizontal line was bisected but the vertical

emphasis (cf. vertical overestimation) prevented the lefteright

horizontal integration, we observed a similar pattern of results.

The lefteright imbalancewas causedby a left underestimation,

but not by a right overestimation. Indeed, in these configura-

tions, the horizontal line was displayed in such a way that the

left and right sides did not compete to be integrated. None-

theless, when the configurations involved a lefteright hori-

zontal competition and integration (X configurations), the

patternofdatadifferedandsuggested that right overestimation

didcontribute topatients’ biasedperformance. Indeed, it seems

that for bothNHþ andNH� patients, the lefteright competition

weakened the left bias (the horizontal underestimation in the

left X configuration is smaller than in the left T or L configura-

tions), but produced a right over-emphasis (the horizontal line

is largely overestimated in the right X).

5.2. Length estimation and the orienting of spatialattention

Taken together, the present results indicate that biased

performance of neglect patients on line bisection can result

from the combined influence of left underestimation and right

overestimation. However, the occurrence and magnitude of

these two biases were influenced by lefteright competition on

the horizontal axis: whereas left underestimation occurred in

the absence of any lefteright horizontal competition, right

overestimation only arose in the context of such a competition

(i.e., right X). Lefteright horizontal competition, perhaps

favouring lefteright integration, decreased left impairment

(i.e., leftX) but also inducedneglect patients tooverestimate the

right part (i.e., right X). These findings are consistent with the

description of peculiar deficits in orienting of spatial attention

occurring in left visual neglect. For example, when performing

Posner-like cuing paradigms (Posner, 1980; Posner and Cohen,

1984), neglect patients typically present reduced detection

facilitation on the left side and no inhibition of return on the

right side (Bartolomeo et al., 2001; Sieroff et al., 2007), together

with an increased disengagement time for left targets preceded

by right-sided cues (Posner et al., 1984; Losier and Klein, 2001;

D’Erme et al., 1992; Bartolomeo et al., 2001). These results can

be interpreted in terms of a deficit of exogenous orienting to

left-sided objects, and a disengagement deficit of exogenous

attention from right-sided objects (Bartolomeo and Chokron,

2002; Rastelli et al., 2008; see also Urbanski and Bartolomeo,

2008).

Impaired exogenous orienting towards the left, perhaps

resulting from dysfunction of fronto-parietal attentional

networks in the right hemisphere (Thiebaut de Schotten et al.,

2005; Corbetta et al., 2005), might be responsible for the left

underestimation. On the other hand, the right disengagement

deficit, perhaps depending on the activity of an isolated left

hemisphere (Bartolomeo et al., 2007), might produce the right

overestimation. Consistent with this hypothesis, left HLH

caused by damage to the OR in the right hemisphere, resulting

in disconnection of primary visual cortex, modulated the size

of the left underestimation (in consistence with previous

findings, Bartolomeo, 1987; D’Erme et al., 1987; Doricchi and

Angelelli, 1999; Doricchi, 2002), but not that of the right over-

estimation. Perhaps hemianopic patients showed a more

pronounced deficit on the left side because the impairment of

the right hemisphere responsible for the left underestimation

was enhanced by the deficit in primary visual areas, which

totally deprived left-sided stimuli even of their residual

attention-catching power (see Bartolomeo et al., 2003; Ferber

and Karnath, 2001). Note, however, that the presence of

hemianopia did not affect the magnitude of right over-

estimation, consistent with the hypothesis that this deficit

may result from the activity of an isolated left hemisphere. It

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2548

Author's personal copy

thus suggests that these two patterns of deficit might be

independent of one another and leads to the conclusion that

the most distinctive sign of left unilateral neglect, uncon-

taminated by hemianopia, might be right attraction, and not

left neglect, which may reflect in some cases the combined

influence of left neglect and hemianopia. Indeed the single

most common sign of neglect in a large patient population

(n¼ 206) tested with an extensive paper-and-pencil battery

was to start on the right side a cancellation task (Azouvi et al.,

2002).

5.3. Length estimation and the metrics of spatialrepresentations

Theories based on the hypothesis of a distorted representa-

tion of space can provide different accounts of rightward

deviations in line bisection. For example, it has been sug-

gested that in neglect, spatial coordinates lose their normal

isometric character and progressively expand from the right

to the left (Bisiach et al., 1996, 1998, 1999). This putative

horizontal distortion is not confined to the left hemispace, but

affects both hemispaces. As a consequence, the anisometry

hypothesis can account for both left underestimation and

right overestimation in neglect. Thus, left underestimation

and right overestimation would result from a single deficit of

space representation, affecting both hemi-fields in a graded

fashion. Consequently, this account predicts that left-sided

segments are underestimated independently of the presence

of right-sided objects, and vice-versa for right over-

estimations. Contrary to this reasoning, our results provide

evidence that right overestimations are highly dependent on

competition between horizontal segments on the two sides of

space (see also Urbanski and Bartolomeo, 2008). Thus, unless

it is assumed that the distortion of space representation is

modulated by the deployment of attention, the anisometry

hypothesis cannot explain why in the present data right

overestimations only occurredwhen the left and right parts of

the line had to be integrated into a whole.

5.4. Neglect and the symmetry law

Thepresent results extend to a larger set of recent findings about

symmetry and line length perception in the normal population,

according to which symmetric bisections produce an underes-

timationas compared toasymmetricbisections (seeCharrasand

Lupianez, 2009, 2010; Wolfe et al., 2005). In the present study,

control participants highlyunderestimated thehorizontal line in

the centre T figure, consistent with the horizontalevertical illu-

sion (Fick, 1851; Kunnapas, 1955, 1957; Finger and Spelt, 1947;

Avery and Day, 1969); however, underestimation decreased

when the horizontal line was asymmetrically bisected (left and

right Ts; see alsoWolfe et al., 2005; Charras and Lupianez, 2010).

Concerning the X configurations, lines were appropriately

perceived in the centre X (i.e., no biaswas observed) but, as soon

as the horizontal line became asymmetrically bisected (left and

rightXs), itwasoverestimated. Importantly, neglect patientshad

similar performance to healthy controls on these stimulus

configurations, which clearly indicates that the symmetry law

was preserved in the present patients.

In agreement with our findings, de Montalembert and

Mamassian (2010) recently claimed that neglect patients were

influenced by the horizontalevertical illusion, in the sense that

vertical lines were overestimated as compared to horizontals

when displayed in an inverted T configuration (called here

centre T). In their study, they manipulated various two-line

configurations, with horizontal lines being displayed either on

the left or right of thevertical line.However, their results arenot

completely consistentwith ours, because deMontalembert and

Mamassian (2010) did not find a clear lefteright imbalance

across configurations. When presented with L configurations,

either left- or right-sided, their neglect patients did show

a vertical overestimation, consistent with the vertical bias

(referred to as the anisotropy component in their study);

however, there was no lefteright imbalance. This result is

somewhat surprising, given that we used similar experimental

paradigms and found a clear difference between left- and right-

oriented configurations, with left underestimation. It is note-

worthy that we replicate here a pattern of results previously

obtained with a partially different group of patients, who were

asked to complete L configurations in a line extension task

(Charras et al., 2010). The discrepancy between our results and

those of de Montalembert and Mamassian might depend on

methodological differences. In de Montalembert and Mamas-

sian study, stimuli remained on the screen for 1000 msec on

each trial,whereas inour study stimulusdurationprogressively

decreased from 1000msec on the first trial to 150 msec on the

last trial. The prolonged exposure time in the deMontalembert

and Mamassian study may have decreased the sensitivity of

their task. On the other hand, they did observe a lefteright

imbalance when using rotated T configurations in which the

horizontal section could either be left- or right-sided (rotated

90�, right horizontal T; rotated 270�, left horizontal T). The

authors suggested in the discussion that the imbalance

between right horizontal T and left horizontal T configurations

was caused by right overestimation, but this claim does not

seem to be supported by experimental evidence. Both a left

underestimation and a right overestimation could explain this

lefteright imbalance. It is important to note that the left and

right horizontal T configurations used in the de Montalembert

andMamassian study are physically very similar to the left and

right X configurations used here. Using our methodology, we

were able to show that the lefteright imbalance in the left and

right X configurations was due to both a left underestimation

and a right overestimation. Therefore, more experimental

evidence is needed to determine what caused the lefteright

imbalance in the horizontal T configurations in the de Mon-

talembert and Mamassian study.

In the following paragraph, we discuss the finding that the

symmetry law is preserved in neglect patients. Indeed, this

preservation raises the issue of whether the symmetry law is

driven by perceptual mechanisms at a different level of pro-

cessing with respect to mechanisms of attentional bias in

neglect. It is, however, important to note that the lefteright

imbalance observed in neglect patients’ performance and the

symmetry law had additive effects. Indeed, despite the

preserved symmetry law (for T and X configurations), neglect

patients showed a clear lefteright imbalance in line length

perception, revealing that left-sided lines were underestimated

as compared to right-sided lines (for T configurations), or that

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2 549

Author's personal copy

right-sided lines were overestimated as compared to left-sided

lines (for X configurations inNH� patients), or, finally, that both

phenomena occurred (for X configurations in NHþ patients).

The finding that a/symmetry influenced both control and

neglect participants indicates that the detection of a/

symmetry might be driven by low perceptual level, perhaps

pre-attentive, mechanisms. It has often been found that

behaviours resulting from pre-attentive mechanisms are

intact after right parietal damage and neglect or extinction

(Mattingley et al., 1997; Halligan and Marshall, 1995;

Shomstein et al., 2010). Indeed, neglect patients’ perfor-

mance is influenced by visual illusions such as the Brentano

version of the MullereLyer illusion (Mattingley et al., 1995;

Vallar et al., 2000; Daini et al., 2002). Gestalt laws such as

grouping or connectedness are known to remain intact after

right parietal damage (Mattingley et al., 1997; Robertson et al.,

2003; Esterman et al., 2000). For instance, perceptual grouping

helps neglect patients to detect left-sided targets (Robertson

et al., 2003), perhaps because left and right are integrated

into a “perceptual whole”. The authors concluded that pre-

attentive mechanisms responsible for grouping enable the

conceptualization of left and right as a unique object, thus

preventing or at least decreasing left neglect. The symmetry

law explored here might rely on similar mechanisms. The

detection of a/symmetry may shape the attentional distribu-

tion all around the locus of bisection, in order to map the

whole configuration. Consequently, the distribution of atten-

tionmight encourage the orienting of attention to the left, and

thus reduce the left underestimation. Biased perception of

a horizontal metal rod can be modulated in a similar way by

the intention to act: neglect-based deviation to the right of

centre was decreased when patients reached towards the rod

as if to pick it up in the centre, compared towhen they pointed

to its centre (Robertson et al., 1995).

Note, however, that even though the attentional distribu-

tion improved left processing, it could not wholly restore left

orienting in the case of neglect plus hemianopia.3

We thus claim here that line a/symmetry can improve line

processing by reducing the lefteright imbalance. However,

this idea seems to be contradictory to what is typically

observed in the landmark task (Milner, et al., 1993; Harvey

et al., 1995). In the landmark task, observers are presented

with a line divided by a small vertical marker and have to

judge which one of the two parts is longer. Usually, neglect

patients still show a lefteright imbalance in this task, i.e., they

tend to judge the right-sided portion as being longer. The

apparent discrepancies between the landmark task and our

results could be accounted for by the fact that the dividing

segment is relatively short in the landmark task, whereas in

the present X configurations it was grossly comparable to the

length of the horizontal segment, with vertical lines being

approximately twice shorter to twice longer than horizontal

lines. Relatively long vertical lines might have helped patients

to further spread their attentional focus (see Halligan and

Marshall, 1994), to the extent of including a relatively long

part of the left segment and thus decreasing the perceptual

imbalance.

5.5. Anatomical considerations

Although the present patient sample was relatively small, the

analysisof theanatomyof lesions thanks tohigh-resolutionMRI

did produce reliable results. No single grey matter area was

invariably damaged in the present sample of neglect patients.

This finding is consistent with the view that the lesion over-

lapping method may not always provide precise data on ana-

tomo-clinical relationships (Bartolomeo, 2006; Godefroy et al.,

1998; Catani and Mesulam, 2008). On the other hand, all

patients presented damage in a white-matter location consis-

tent with the course of the IFOF. Accumulating evidence corre-

lates neglect signswith dysfunction of fronto-parietal networks

in the right hemisphere, linkedbybranches of theSLF (Thiebaut

de Schotten et al., 2005; Doricchi and Tomaiuolo, 2003; Corbetta

et al., 2005; Bartolomeo et al., 2007; Doricchi et al., 2008).

However, IFOF disconnection in the right hemisphere has also

been described in a group of neglect patients with relatively

ventral lesions, which only marginally impinged upon the

parietal lobe and the SLF (Urbanski et al., 2008). IFOF discon-

nection might provoke signs of left neglect by impairing the

topedown influences of frontal regions to visual cortex, or by

depriving frontal areas of visual input from the left visual field.

Thus, the present case could be a typical example of failure to

observe consistent grey matter correlates of a neurological

deficit, because the deficit actually stems from damage to

different parts of a long-range white-matter fasciculus (Catani

and Mesulam, 2008). Finally, as expected, damage to the OR

was only observed in hemianopic patients, thus validating our

anatomo-clinical approach.

5.6. Conclusions

We described a new approach for investigating the global

lefteright imbalance, characteristic of left unilateral neglect, by

dissociating the relative contribution of the left and right biases

(see also Charras et al., 2010). The results indicate that left

impairment occurs independently of lefteright competition, in

opposition to the right bias. Indeed, right over-enhancement

appeared to be consecutive to perceptual lefteright integration;

when no lefteright competition was present, neglect patients

performed in a similar way as controls on the right side. In

addition, left HLHmodulated the left but not the right bias, thus

making the case for the independenceof these twobiases. Based

on the present results, we propose distinct neural bases for right

overestimation, resulting from the activity of an isolated left

hemisphere, and left underestimation, dependent on impaired

functioning of right-hemisphere attentional networks.

Acknowledgements

This research was supported by the Spanish “Ministerio de

Educacion y Ciencia” (by a predoctoral grant -AP-2006-3911- to

thefirst author, and a research grant -PSI2008-03595PSIC- to the

second author). Please direct correspondence concerning this

3 As previously detailed, the left underestimation did notsimply result from sensory deficit like hemianopia, which couldaccount for 25% of underestimation, whilst the effect reachedabout 40%.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2550

Author's personal copy

article to: Pom Charras, Departamento de Psicologıa Experi-

mental y Fisiologıa del Comportamiento, Facultad de Psicolo-

gıa, Universidad de Granada, Campus Universitario de Cartuja

S/N, 18071, Granada, Spain. E-mail: [email protected].

r e f e r e n c e s

Albert M. A simple test of visual neglect. Neurology, 23:658e664, 1973.

Avery G and Day R. Basis of the horizontal vertical illusion. Journalof Experimental Psychology, 81(2): 376e380, 1969.

Azouvi P, Samuel C, Louis-Dreyfus A, Bernati T, Bartolomeo P,Beis J-M, et al. Sensitivity of clinical and behavioural tests ofspatial neglect after right hemisphere stroke. Journal ofNeurology, Neurosurgery and Psychiatry, 73(2): 160e166, 2002.

Bartolomeo P. Aspetti dell’emi-inattenzione spaziale nelle lesioniemisferiche: Fattori che influenzano la bisezione di linee [aspects ofspatial hemi-inattention: Factors influencing line bisectionperformance]. Secondary Titl. Roma: Universita Cattolica, 1987.

Bartolomeo P. A parieto-frontal network for spatial awareness inthe right hemisphere of the human brain. Archives ofNeurology, 63(9): 1238e1241, 2006.

Bartolomeo P. Visual neglect. Current Opinion in Neurology, 20(4):381e386, 2007.

Bartolomeo P and Chokron S. Left unilateral neglect or righthyperattention. Neurology, 53(9): 2023e2027, 1999.

Bartolomeo P and Chokron S. Orienting of attention in leftunilateral neglect. Neuroscience and Biobehavioral Reviews,26(2): 217e234, 2002.

Bartolomeo P, Chokron S, and Sieroff E. Facilitation instead ofinhibition for repeated right-sided events in left neglect.NeuroReport, 10(16): 3353e3357, 1999.

Bartolomeo P, Sieroff E, Decaix C, and Chokron S. Modulating theattentional bias in unilateral neglect: The effects of the strategicset. Experimental Brain Research, 137(3/4): 424e431, 2001.

Bartolomeo P, Thiebaut de Schotten M, and Doricchi F. Leftunilateral neglect as a disconnection syndrome. CerebralCortex, 17(11): 2479e2490, 2007.

Bartolomeo P, Urbanski M, Chokron S, Chainay H, Moroni C,Sieroff E, et al. Neglected attention in apparent spatialcompression. Neuropsychologia, 42(1): 49e61, 2003.

Baylis GC, Driver J, and Rafal RD. Visual extinction and stimulusrepetition. Journal of Cognitive Neuroscience, 5(4): 453e466, 1993.

Binder J, Marshall R, Lazar R, Benjamin J, and Mohr JP. Distinctsyndromes of hemineglect. Archives of Neurology, 49(11):1187e1194, 1992.

Bird CM, Malhotra P, Parton A, Coulthard E, Rushworth MF, andHusain M. Visual neglect following right posterior cerebralartery infarction. Journal of Neurology, Neurosurgery andPsychiatry, 77(9): 1008e1012, 2006.

Bisiach E, Neppi-Modona M, Genero R, and Pepi R. Anisometry ofspace representation in unilateral neglect: Empirical test ofa former hypothesis. Consciousness and Cognition, 8(4):577e584, 1999.

Bisiach E, Pizzamiglio L, Nico D, and Antonucci G. Beyondunilateral neglect. Brain, 119(3): 851e857, 1996.

Bisiach E, Ricci R, and Modona MN. Visual awareness andanisometry of space representation in unilateral neglect: Apanoramic investigation by means of a line extension task.Consciousness and Cognition, 7(3): 327e355, 1998.

Bonato M, Priftis K, Marenzi R, and Zorzi M. Modulation ofhemispatial neglect by directional and numerical cues in theline bisection task. Neuropsychologia, 46(2): 426e433, 2008.

Brett M, Leff AP, Rorden C, and Ashburner J. Spatial normalizationof brain images with focal lesions using cost functionmasking. NeuroImage, 14(2): 486e500, 2001.

Catani M and Mesulam M. The arcuate fasciculus and thedisconnection theme in language and aphasia: History andcurrent state. Cortex, 44(8): 953e961, 2008.

Catani M and Thiebaut de Schotten M. A diffusion tensor imagingtractography atlas for virtual in vivo dissections. Cortex,44(8): 1105e1132, 2008.

Charras P and Lupianez J. The relevance of symmetry in linelength perception. Perception, 38(10): 1428e1438, 2009.

Charras P and Lupianez J. Length perception of horizontal andvertical bisected lines. Psychological Research, 74(2): 196e206, 2010.

Charras P, Lupianez J, and Bartolomeo P. Assessing the weights ofvisual neglect: A new approach to dissociate defectivesymptoms from productive phenomena in length estimation.Neuropsychologia, 48(11): 3371e3375, 2010.

Corbetta M, Kincade MJ, Lewis C, Snyder AZ, and Sapir A. Neuralbasis and recovery of spatial attention deficits in spatialneglect. Nature Neuroscience, 8(11): 1603e1610, 2005.

D’Erme P, DeBonis C, and Gainotti G. Influenza dell’emi-inattenzione e dell’emianopsia sui compiti di bisezione dilinee nei pazienti cerebrolesi. Archivio di Psicologia, Neurologiae Psichiatria, 48: 165e189, 1987.

D’Erme P, Robertson IH, Bartolomeo P, Daniele A, and Gainotti G.Early rightwards orienting of attention on simple reactiontime performance in patients with left-sided neglect.Neuropsychologia, 30(11): 989e1000, 1992.

Daini R, Angelelli P, Antonucci G, Cappa SF, and Vallar G.Exploring the syndrome of spatial unilateral neglect throughan illusion of length. Experimental Brain Research, 144(2):224e237, 2002.

de Montalembert M and Mamassian P. The verticalehorizontalillusion in hemi-spatial neglect. Neuropsychologia, 48(11):3245e3251, 2010.

De Renzi E, Gentilini M, Faglioni P, and Barbieri C. Attentionalshift towards the rightmost stimuli in patients with left visualneglect. Cortex, 25(2): 231e237, 1989.

Doricchi F. The contribution of retinotopic and multimodalcoding of space to horizontal space misrepresentation inneglect and hemianopia. In Karnath HO, Milner AD, andVallar G (Eds), The Cognitive and Neural Bases of Spatial Neglect.New York: Oxford University Press, 2002.

Doricchi F and Angelelli P. Misrepresentation of horizontal spacein left unilateral neglect: Role of hemianopia. Neurology,52(9): 1845e1852, 1999.

Doricchi F, Thiebaut de Schotten M, Tomaiuolo F, andBartolomeo P. White matter (dis)connections and gray matter(dys)functions in visual neglect: Gaining insights into thebrain networks of spatial awareness. Cortex, 44(8):983e995, 2008.

Doricchi F and Tomaiuolo F. The anatomy of neglect withouthemianopia: A key role for parietalefrontal disconnection?NeuroReport, 14(17): 2239e2243, 2003.

Esterman M, McGlinchey-Berroth R, and Milberg WP. Parallel andserial search in hemispatial neglect: Evidence for preservedpreattentive but impaired attentive processing.Neuropsychology, 14(4): 599e611, 2000.

Ferber S and Karnath HO. Size perception in hemianopia andneglect. Brain, 124(3): 527e536, 2001.

Fick A. De errore quodam optico asymetrica bulbi effecto. Marburg:J. A. Kochin.

Finger FW and Spelt DK. The illustration of the horizontal verticalillusion. Journal of Experimental Psychology, 37(3): 243e250, 1947.

Gainotti G, D’Erme P, and Bartolomeo P. Early orientation ofattention toward the half space ipsilateral to the lesion inpatients with unilateral brain damage. Journal of Neurology,Neurosurgery and Psychiatry, 54(12): 1082e1089, 1991.

Gauthier L, Dehaut F, and Joanette Y. The bells test: A quantitativeand qualitative test for visual neglect. International Journal ofClinical Neuropsychology, 11(2): 46e54, 1989.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2 551

Author's personal copy

Godefroy O, Duhamel A, Leclerc X, Saint Michel T, Henon H, andLeys D. Brain-behaviour relationships. Some models andrelated statistical procedures for the study of brain-damagedpatients. Brain, 121(8): 1545e1556, 1998.

Halligan PW and Marshall JC. Right-sided cueing can ameliorateleft neglect. Neuropsychological Rehabilitation, 4(1): 63e73, 1994.

Halligan PW and Marshall JC. Grounding figural attention in leftneglect. Neurocase, 1: 79e82, 1995.

Harvey M, Milner AD, and Roberts RC. An investigation ofhemispatial neglect using the landmark task. Brain & Cognition,27(1): 59e78, 1995.

Heilman KM and Valenstein E. Mechanisms underlyinghemispatial neglect. Annals of Neurology, 5(2): 166e170, 1979.

Humphreys GW, Romani C, Olson A, Riddoch MJ, and Duncan J.Non-spatial extinction following lesions of the parietal lobe inhumans. Nature, 372: 357e359, 1994.

Ishiai S, Koyama Y, Seki K, and Izawa M. Line versusrepresentational bisections in unilateral spatial neglect. Journalof Neurology, Neurosurgery and Psychiatry, 69(6): 745e750, 2000.

Kinsbourne M. Orientational bias model of unilateral neglect:Evidence from attentional gradients within hemispace. InRoberston IH and Marshall JC (Eds), Unilateral Neglect: Clinicaland Experimental Studies. Hillsdale, NJ: Lawrence ErlbaumAssociates, Inc, 1993: 63e86.

Kunnapas TM. An analysis of the “vertical horizontal illusion”.Journal of Experimental Psychology, 49(2): 134e140, 1955.

Kunnapas TM. Vertical horizontal illusion and surrounding field.Acta Psychologica, 13: 35e42, 1957.

Losier BJ and Klein RM. A review of the evidence for a disengagedeficit following parietal lobe damage. Neuroscience andBiobehavioral Reviews, 25(1): 1e13, 2001.

Marshall JC and Halligan PW. When right goes left: Aninvestigation of line bisection in a case of visual neglect.Cortex, 25(3): 503e515, 1989.

Mattingley JB, Bradshaw JL, and Bradshaw JA. The effects ofunilateral visuospatial neglect on perception of muller-lyerillusory figures. Perception, 24(4): 415e433, 1995.

Mattingley JD, David G, and Driver J. Preattentive filling-in ofvisual surfaces in parietal extinction. Science, 275(5300):671e674, 1997.

McIntosh RD, Schindler I, Birchall D, and Milner AD. Weights andmeasures: A new look at bisection behaviour in neglect.Cognitive Brain Research, 25(3): 833e850, 2005.

Milner AD, Harvey M, Roberts RC, and Forster SV. Line bisectionerrors in visual neglect: Misguided action or size distortion?Neuropsychologia, 31(1): 39e49, 1993.

Mort DJ, Malhotra P, Mannan SK, Rorden C, Pambakian A,Kennard C, et al. The anatomy of visual neglect. Brain, 126(9):1986e1997, 2003.

Natale E, Posteraro L, Prior M, and Marzi CA. What kind of spatialattention is impaired in neglect? Neuropsychologia, 43(7):1072e1085, 2005.

Posner M, Walker J, Friedrich F, and Rafal R. Effects of parietalinjury on covert orienting of attention. Journal of Neuroscience,4(7): 1863e1874, 1984.

Posner MI. Orienting of attention. Quarterly Journal ExperimentalPsychology, 32(1): 3e25, 1980.

Posner MI and Cohen Y. Components of visual orienting. InBouma H and Bouwhuis D (Eds)Attention and performance.London: Lawrence Erlbaum, 1984: 531e556.

Rastelli F, Funes MJ, Lupianez J, Duret C, and Bartolomeo P. Leftneglect: Is the disengage deficit space- or object-based?Experimental Brain Research, 187(3): 439e446, 2008.

Riddoch MJ and Humphreys GW. The effect of cueing onunilateral neglect. Neuropsychologia, 21(6): 589e599, 1983.

Robertson LC, Mirjam E, and Knight R. Grouping influences inunilateral visual neglect. Journal of Clinical and ExperimentalNeuropsychology, 25(3): 297e307, 2003.

Robertson IH, Nico D, and Hood BM. The intention to act improvesunilateral left neglect: Two demonstrations. NeuroReport, 7(1):246e248, 1995.

Rorden C, Berger MF, and Karnath HO. Disturbed line bisection isassociated with posterior brain lesions. Brain Research, 1080(1):17e25, 2006.

Rorden C and Brett M. Stereotaxic display of brain lesions.Behavioural Neurology, 12(4): 191e200, 2000.

Rorden C, Karnath HO, and Bonilha L. Improving lesion-symptommapping. Journal of Cognitive Neuroscience, 19(7): 1081e1088,2007.

Schenkenberg T, Bradford DC, and Ajax ET. Line bisection andunilateral visual neglect in patients with neurologicimpairment. Neurology, 30(5): 509e517, 1980.

Schneider W, Eschman A, and Zuccolotto A. E-prime User’s Guide.Pittsburg: Psychology Software Tools Inc., 2002.

Shomstein S, Kimchi R, Hammer M, and Behrmann M. Perceptualgrouping operates independently of attentional selection:Evidence from hemispatial neglect. Attention, Perception &Psychophysics, 72(3): 607e618, 2010.

Sieroff E, Decaix C, Chokron S, and Bartolomeo P. Impairedorienting of attention in left unilateral neglect: Acomponential analysis. Neuropsychology, 21(1): 94e113,2007.

Snow JC and Mattingley JB. Goal-driven selective attention inpatients with right hemisphere lesions: How intact is theipsilesional field? Brain, 129(1): 168e181, 2006.

Snow JC and Mattingley JB. Central perceptual load does notreduce ipsilesional flanker interference in parietal extinction.Neuropsychology, 22(3): 371e382, 2008.

Thiebaut de Schotten M, Urbanski M, Duffau H, Volle E, Levy R,Dubois B, et al. Direct evidence for a parietalefrontal pathwaysubserving spatial awareness in humans. Science, 309(5744):2226e2228, 2005.

Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F,Etard O, Delcroix N, et al. Automated anatomical labeling ofactivations in SPM using a macroscopic anatomicalparcellation of the MNI MRI single-subject brain. NeuroImage,15(1): 273e289, 2002.

Urbanski M and Bartolomeo P. Line bisection in left neglect:The importance of starting right. Cortex, 44(7):782e793, 2008.

Urbanski M, Thiebaut de Schotten M, Rodrigo S, Catani M,Oppenheim C, Touze E, et al. Brain networks of spatialawareness: Evidence from diffusion tensor imagingtractography. Journal of Neurology, Neurosurgery and Psychiatry,79(5): 598e601, 2008.

Vallar G. Extrapersonal visual unilateral spatial neglect and itsneuroanatomy. NeuroImage, 14(1): 52e58, 2001.

Vallar G, Daini R, and Antonucci G. Processing of illusion of lengthin spatial hemineglect: A study of line bisection.Neuropsychologia, 38(7): 1087e1097, 2000.

Ward R and Goodrich S. Differences between objects andnonobjects in visual extinction: A competition for attention.Psychological Science, 7(3): 177e180, 1996.

Wolfe U, Maloney LT, and Tam M. Distortions of perceived lengthin the frontoparallel plane: Tests of perspective theories.Perception and Psychophysics, 67(6): 967e979, 2005.

c o r t e x 4 8 ( 2 0 1 2 ) 5 4 0e5 5 2552