BRAINA JOURNAL OF NEUROLOGY

Effect of prism adaptation on left dichotic listeningdeficit in neglect patients: glasses to hear better?S. Jacquin-Courtois,1,2,3,4 G. Rode,1,2,3,4 F. Pavani,5 J. O’Shea,6 M. H. Giard,7 D. Boisson1,2,3,4 andY. Rossetti1,2,3,4

1 Inserm UMR-S 864, Espace et Action, 16 avenue Lepine, 69676 Bron, France

2 Universite de Lyon, Universite Lyon 1, France

3 Hospices Civils de Lyon, Service de Medecine Physique et Readaptation, Hopital Henry Gabrielle, route de Vourles, Saint Genis Laval, France

4 Hospices Civils de Lyon & Institut Federatif des Neurosciences de Lyon, Mouvement et Handicap, St Genis Laval et Lyon, France

5 Dipartemento di Scienze della Cognizione e della Formazione, Universita degli Studi di Trento, Rovereto, Italy

6 Oxford Centre for Functional MRI of the Brain (FMRIB), John Radcliffe Hospital, University of Oxford, Oxford OX3 9DU, UK

7 Inserm UMR-S 280, Mental Processes and Brain Activation, Centre Hospitalier le Vinatier, Batiment 452, 95 Boulevard Pinel, F-69500 Bron, France

Correspondence to: Sophie Jacquin-Courtois,

Hopital Henry Gabrielle,

20 Route de Vourles,

Saint Genis Laval 69230,

France

E-mail: sophie.courtois@chu-lyon.fr

Unilateral neglect is a disabling syndrome frequently observed following right hemisphere brain damage. Symptoms range from

visuo-motor impairments through to deficient visuo-spatial imagery, but impairment can also affect the auditory modality. A short

period of adaptation to a rightward prismatic shift of the visual field is known to improve a wide range of hemispatial neglect

symptoms, including visuo-manual tasks, mental imagery, postural imbalance, visuo-verbal measures and number bisection. The

aim of the present study was to assess whether the beneficial effects of prism adaptation may generalize to auditory manifesta-

tions of neglect. Auditory extinction, whose clinical manifestations are independent of the sensory modalities engaged in

visuo-manual adaptation, was examined in neglect patients before and after prism adaptation. Two separate groups of neglect

patients (all of whom exhibited left auditory extinction) underwent prism adaptation: one group (n = 6) received a classical prism

treatment (‘Prism’ group), the other group (n = 6) was submitted to the same procedure, but wore neutral glasses creating no

optical shift (placebo ‘Control’ group). Auditory extinction was assessed by means of a dichotic listening task performed three

times: prior to prism exposure (pre-test), upon prism removal (0 h post-test) and 2 h later (2 h post-test). The total number of

correct responses, the lateralization index (detection asymmetry between the two ears) and the number of left-right fusion errors

were analysed. Our results demonstrate that prism adaptation can improve left auditory extinction, thus revealing transfer of

benefit to a sensory modality that is orthogonal to the visual, proprioceptive and motor modalities directly implicated in the

visuo-motor adaptive process. The observed benefit was specific to the detection asymmetry between the two ears and did not

affect the total number of responses. This indicates a specific effect of prism adaptation on lateralized processes rather than on

general arousal. Our results suggest that the effects of prism adaptation can extend to unexposed sensory systems. The bottom-up

approach of visuo-motor adaptation appears to interact with higher order brain functions related to multisensory integration and

can have beneficial effects on sensory processing in different modalities. These findings should stimulate the development of

therapeutic approaches aimed at bypassing the affected sensory processing modality by adapting other sensory modalities.

Keywords: neglect; auditory; prism; adaptation; crossmodal

doi:10.1093/brain/awp327 Brain 2010: 133; 895–908 | 895

Received August 12, 2009. Revised October 23, 2009. Accepted November 16, 2009. Advance Access publication January 27, 2010

� The Author (2010). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.

For Permissions, please email: journals.permissions@oxfordjournals.org

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

IntroductionUnilateral neglect is frequently observed in right-handed patients

following right hemisphere brain damage. It is a debilitating

neurological disorder characterized by complex deficits in attention

and spatial processing. The syndrome has been defined as a loss of

awareness for stimuli, contra-lateral to the lesion and despite

residual implicit processing (Driver and Mattingley, 1998;

Marshall and Halligan, 1988). Typically, the patient fails to

report, respond to or orient towards left-sided stimuli (Heilman

et al., 1985; Cappa et al., 1987), or tends to underestimate

them on various perceptual dimensions (Halligan and Marshall,

1991; Milner et al., 1993). Neglect symptoms range from deficits

in sensori-motor processing (Heilman et al., 1985; Mattingley

et al., 1992, 1998) through to impairments in the mental repre-

sentation of space (Bisiach and Luzzatti, 1978; Rode et al., 1995;

Guariglia et al., 2005). In everyday life, neglect patients may omit

to read the left part of a journal or a book, omit to eat food from

the left half of a plate, forget to shave the left half of the face or

hit obstacles on the left when moving in the environment.

In addition to these heterogeneous symptoms and reflecting the

complexity of this multifaceted syndrome, it is now increasingly

acknowledged that unilateral neglect is not just restricted to the

visual modality, but can often affect different sensory modalities

concurrently (for review see Brozzoli et al., 2006). In particular,

numerous studies of patients with visuo-spatial neglect found con-

comitant deficits in auditory attention and spatial processing

(Bisiach et al., 1984; Soroker et al., 1997; Pavani et al., 2004).

Bisiach et al. (1984) first described impairments in auditory spatial

processing in patients with spatial neglect, i.e. auditory localiza-

tions errors, particularly when pointing to contra-lesional sounds.

They emphasized the importance of investigations in other sensory

modalities in neglect and not just vision. In more recent years,

auditory manifestations of unilateral neglect have been more fre-

quently studied (Pavani et al., 2003, 2004). There is now some

evidence that visual neglect patients can manifest contra-lesional

deficits in a variety of auditory tasks (detection, identification and

localization).

Impaired detection of contra-lesional sounds in visual neglect

patients has been observed when targets were presented together

with other competing sounds (De Renzi et al., 1989; Beaton and

MacCarthy, 1993, 1995). In such a situation of auditory bilateral

presentation, an auditory deficit has been reported for free-fields

sounds (Bender and Diamond, 1965; Heilman and Valenstein,

1972) as well as for sounds presented dichotically over head-

phones (De Renzi et al., 1984, 1989).

Just as for simple auditory detection tasks, deficits in auditory

identification tasks in neglect have more often been observed

under bilateral simultaneous stimulation, with competition thus

arising between different concurrent auditory targets. Deficits for

contra-lesional sounds have been demonstrated for both free-field

stimulation (Calamaro et al., 1995; Soroker et al., 1995, 1997;

Deouell and Soroker, 2000) and dichotic presentation over head-

phones (Hugdahl et al., 1991; Bellmann et al., 2001; Pavani et al.,

2005). These auditory deficits were typically more pronounced in

visual neglect patients than in healthy controls and, more

importantly, were more severe than in right-brain damaged

patients without neglect (Calamaro et al., 1995; Pavani et al.,

2005). The interpretation of this phenomenon was initially equiv-

ocal about whether deficits in neglect patients reflected suppres-

sion of auditory input from the contra-lesional ear or higher order

deficits in spatial processing (Beaton and MacCarthy, 1993, 1995;

Bellmann et al., 2001). More recent contributions (Pavani et al.,

2005; Spierer et al., 2007) strongly suggest a role for higher

level spatial factors in some of these auditory deficits. In combi-

nation, the available evidence converges to support the interpre-

tation of neglect as a disturbance of multisensory spatial

processing.

Sound localization tasks have also revealed predominantly

contra-lesional deficits, even for single unilateral stimuli.

Different sound-localization paradigms have been used, such as

pointing to sounds (Ruff et al., 1981; Bisiach et al., 1984) and

auditory midline perception (Vallar et al., 1995; Kerkhoff et al.,

1999; Tanaka et al., 1999). Pavani et al. (2001) demonstrated a

deficit in auditory localization for contra-lesional space processing

in visual neglect patients. In combination, these results support the

existence of some impairment in detecting changes in sound posi-

tion in visual neglect patients.

In addition, evidence suggests that non-spatial auditory deficits

may also be observed. Using an auditory test of sustained atten-

tion, Robertson et al. (1997) reported a non-spatial auditory def-

icit in right brain-damaged patients with neglect compared to

non-neglect patients, with a correlation between impairment in

this task and severity of visual neglect. Attentional limits have

been evoked to explain a deficit arising during a task of compar-

isons between brief successive sounds presented centrally (Cusack

et al., 2000). Pavani et al. (2005) reported difficulties in a task in

which words were presented bilaterally and diotically, this deficit

affecting ‘left’ and ‘right’ sounds equally on double stimulation,

suggesting a general capacity limitation in neglect patients in addi-

tion to their lateralized spatial biases.

Finally, recent studies concerning multimodal representation

have provided empirical evidence indicating that parietal cortex

and other brain structures commonly associated with neglect

(superior temporal lobe: Karnath et al., 2001; premotor and fron-

tal corticies: Husain and Kennard, 1996) may play a fundamental

role in multisensory processing of space.

Hence, auditory deficits occur frequently in right brain-damaged

patients with neglect. Although it is well established that the pres-

ence of post-stroke neglect is a poor prognostic factor for recovery

(e.g. Denes et al., 1982; Jehkonen et al., 2000), the specific

consequences for recovery and rehabilitation of the multisensory

aspects of neglect have yet to be explored. For instance, it can be

speculated that the debilitating outcome of the neglect syndrome

may be compounded when neglect-related disorders affect differ-

ent sensory modalities simultaneously. A rehabilitation protocol

that proves to be simultaneously effective on the different clinical

manifestations of the neglect syndrome would thus have a greater

impact on clinical outcome. In this respect, prism adaptation, a

rehabilitative procedure introduced 11 years ago (Rossetti et al.,

1998) appears particularly promising (Milner and MacIntosh,

2005; Luaute et al., 2006a), since it has been shown to produce

improvement in a wide range of neglect symptoms in the visual

896 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

domain (for review see Rossetti and Rode, 2002; Rode et al.,

2003, 2006a, 2007; Pisella et al., 2006; Revol et al., 2007), and

has recently shown rehabilitative efficacy for neglect-related

disorders in the tactile domain (Maravita et al., 2003; Dijkerman

et al., 2004).

Previous studies have demonstrated that a short period of

visuo-motor adaptation to a rightward prismatic shift of the

visual field can improve symptoms of hemispatial neglect in a

wide range of visuo-motor tasks such as line-bisection,

line-cancellation or copy drawing (Rossetti et al., 1998), but also

in mental imagery (Rode et al., 1998), postural imbalance (Tilikete

et al., 2001), wheel-chair driving (Jacquin-Courtois et al., 2008),

visuo-verbal measures (Farne et al., 2002), neglect dysgraphia

(Rode et al., 2006b) and number bisection (Rossetti et al.,

2004). This suggests that visuo-manual adaptation can modify

spatial representations at a higher level than at a simply sensory

level, suggesting a possible link between sensori-motor plasticity

and space representation. Prominently, while in the first study

(Rossetti et al., 1998) patients’ improvement was fully maintained

2 h later, in further case studies the effects of a single session of

prism adaptation have been reported to last up to 4 days (e.g.

MacIntosh et al., 2002; Pisella et al., 2002). Repeated sessions of

prism adaptation over several weeks have been shown to induce a

long-lasting improvement of neglect patients’ performance, not

only in standard clinical tests of neglect but also in more ecological

assessment (Frassinetti et al., 2002; Humphreys et al., 2006;

Shiraishi et al., 2008; Serino et al., 2009). The improvement

found after prism adaptation in a wide variety of visuo-spatial

tasks indicates that prism adaptation may affect the organization

of higher levels of spatial representation, such as those impaired in

neglect patients.

Although tactile effects of prism adaptation have been described

(Maravita et al., 2003; Dijkerman et al., 2004), it is possible that

such effects can be ascribed to cross-talk within the

somato-sensory system between proprioception (which is directly

modified by adaptation) and the tactile modality. However,

if visuo-manual prism adaptation improves auditory manifestations

of neglect, this would imply that adaptation alters higher level

representations of space in such a way that it can modify the

processing of sensory input that has not been directly altered by

prism exposure.

The aim of the present study was to test whether prism adap-

tation effects generalize to neglect symptoms that are indepen-

dent of any of the components directly linked to visuo-manual

adaptation. To this end, a common auditory manifestation of

neglect, auditory extinction, was quantified by a dichotic listening

task. Amelioration of auditory extinction by prism adaptation

would be particularly remarkable, because this neglect-related def-

icit emerges from a different sensory modality (audition) than

those involved in the adaptation procedure (i.e. vision and pro-

prioception), and it is assessed by a non-manual response (i.e.

verbal report). Hence, auditory extinction is tested by a task that

is fully independent from the sensory and motor components

intrinsic to the prism adaptation procedure.

In our study, auditory extinction was examined in neglect

patients before and after prism adaptation. Patients were ran-

domly assigned to two groups to assess the effect of prism

adaptation: one group received a classical prism treatment and

the other group underwent the same procedure but wore neutral

glasses that do not create any optical shift.

Methods

SubjectsTwelve right brain-damaged patients were included in this study,

according to the following criteria: (i) unilateral right hemisphere

damage with no history of previous neurological illness; (ii) left unilat-

eral visual neglect on admission to the rehabilitation unit, as assessed

by classical neuropsychological tests [line cancellation test (Albert,

1973), line bisection task (Schenkenberg et al., 1980) and a

copy-drawing task (Gainotti et al., 1972)]; (iii) left ear extinction on

the dichotic listening task (see ‘Methods’ section) (preliminary clinical

evaluation of auditory extinction was performed with the sound of

snapping fingers being delivered to both ears; left auditory extinction

represented repeated neglect of applied sound stimulus to left ear);

(iv) adequate hearing threshold as measured by pure tone audiometry

(determined over a range of frequencies between 256 and 8192 Hz)

and no asymmetric loss (510 dB difference in mean threshold between

left and right ear); and (v) right-handedness. In addition, 10 healthy

control subjects (4 male, 6 female) aged between 23 and 58 years

(mean = 37; SD = 12.60) were recruited to obtain normative data for

the dichotic listening task.

At the time of clinical examination, there was neither confusion nor

temporal or spatial disorientation. Informed consent was obtained

from all patients prior to testing; the procedure, approved by French

law on patients’ rights (4 March 2002), was in accordance with the

Declaration of Helsinki. Patients were randomly assigned to the two

groups. A finding commonly observed and reported in the literature is

that neglect patients do not overtly (neither spontaneously nor follow-

ing direct questioning) detect the visual perturbation caused by prisms,

which normally induce a spontaneous surprise reaction in healthy sub-

jects (Michel et al., 2007). Further, neglect patients do not show any

changes in skin conductance when prisms are unexpectedly introduced

in the course of a pointing task (Calabria et al., 2004). This evidence

suggests that neglect patients are not aware of the visual perturbation,

but nevertheless they adapt to it. Accordingly, our patients did not

report that the goggles produced any visual shift, even when specif-

ically questioned.

Clinical and paraclinical data, lesion topography and co-morbid dis-

orders are summarized for each patient in Table 1. All patients had a

CT or MRI scan to define the lesion topography (Fig. 1). As shown in

Table 1, the implication of the global fronto-parieto-temporal network

was more frequent in the ‘Control’ group. However, Fig. 1 shows that

lesion size tended to be larger in the ‘Prism’ group. Patients’ lesions

depicted in Fig. 1 have been quantified by a computerized imaging

system (Leica imaging system and Quantimet 500 software). This

system quantified each area marked in black on Fig. 1 and provided

a total area value for each patient in arbitrary pixel unit. A t-test

performed on the total lesion surface showed that the ‘Prism’ group

(mean = 13 933, SEM = 6278) does not differ significantly [t(10) = 1.44;

P40.17] from the ‘Control’ group (mean = 9333, SEM = 4640)

[t(10) = 1.44; P40.17]. In addition, both groups are comparable in

terms of paraventricular fibre lesion corresponding to the auditory

inter-hemispheric pathway (two patients in the ‘Control’ group,

three patients in the ‘Prism’ group) and no patient exhibited callosal

disconnection at the mid-sagittal plane. Visual and tactile extinction

Auditory extinction: glasses to hear better? Brain 2010: 133; 895–908 | 897

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

were assessed by confrontation testing. At the time of testing, 7/12

patients still showed left unilateral neglect. ‘Prism’ group (n = 6) refers

to patients who underwent prism adaptation, ‘Control’ group (n = 6)

refers to patients who wore neutral glasses. There was no difference

between the groups in scores on the visual neglect tests performed

prior to the experimental intervention [Albert test: t(10) =�1.43,

P40.18; Schenkenberg test: t(10) = 0.33, P40.74]. The two groups

also did not differ significantly in age (‘Prism’ group: mean age = 59.7

years, SEM = 6.5 and ‘Control’ group: mean age = 56.7 years,

SEM = 4.7) or in their post-stroke onset delay prior to experimental

Figure 1 Brain lesions in patients. The figure shows the location of brain damage in each patient, for both groups (‘Prism’ and ‘Control’),

according to the standard template provided by Damasio and Damasio (1989). The lesion areas depicted in black were subjected to

quantitative analysis by a computerized imaging system.

Table 1 Clinical and paraclinical data for all patients (’Prism’ and ’Control’ groups)

Group Patient Sex Age(years)

Lesion(righthemisphere)

Delay(days)

L visualneglect

Left visualextinction

Left somestheticextinction

Left auditoryextinction

At time oftesting

At time oftesting

Prism D.M. M 72 Parietal 73 No Not determined No Yes

Prism B.E. F 42 Temporo-parietal 566 No Left hemianopia Yes Yes

Prism A.E. F 75 Parieto-occipital 80 Yes Left hemianopia Yes Yes

Prism G.C. F 38 Fronto-parietal 124 Yes Left hemianopia Yes Yes

Prism L.A. M 62 Temporo-occipital 515 Yes Left hemianopia Not determined Yes

Prism M.J. M 69 Thalamic 60 Yes Yes Yes Yes

Control B.B. F 50 Parietal 30 No No Yes Yes

Control N.G. M 70 Parieto-occipital 35 No No Yes Yes

Control B.C. F 71 Parietal 109 No Not determined Yes Yes

Control D.J. M 45 Capsulolenticular 41 Yes Left hemianopia Not determined Yes

Control G.M. F 47 Fronto-temporo-parietal 53 Yes Yes No Yes

Control F.J.Y. M 57 Fronto-temporo-parietal 146 Yes Left liemiauopia Not determined Yes

898 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

testing (236 days, SEM = 96.8 for the ‘Prism’ group and 69 days,

SEM = 19.4 for the ‘Control’ group). A t-test with assumption of

unequal variances was performed on age and delay [age:

t(9.1) = 0.36, P40.5; delay: t(5.4) = 1.69, P40.1]. Note that the

numerically shorter post-stroke onset delay of the ‘Control’ group

(compared to the ‘Prism’ group) would predict a more favourable

spontaneous recovery rate. Alternatively, it may be argued that

chronic patients may already have a less severe form of neglect and

hence show the effect of an intervention more easily. However,

neglect assessment showed that this was not the case (no significant

difference observed between the two groups regarding severity of

visual neglect).

Experimental procedures

Evaluation of left visuospatial neglect at time of testing

Neglect was assessed by three classical neuropsychological tests: line

cancellation (Albert, 1973), line bisection (Schenkenberg et al., 1980)

and a copy-drawing task (Gainotti et al., 1972). The presence of

neglect was defined by a deficit in at least two of these tests.

Evaluation of left auditory extinction

The patients sat in a quiet room in front of the examiner. Stimuli were

played through earphones, at the volume judged most comfortable by

each patient, in order to study extinction in its everyday clinically rel-

evant form.

The dichotic listening task was a verbal dichotic listening test, includ-

ing 60 pairs of stimuli (total across all categories). The two stimuli of

each word pair were simultaneously presented, one to the left and the

other to the right ear, through the earphones. There were four cate-

gories of stimuli: pairs of phonologically similar disyllabic words [e.g.

blague/bla:g (joke); claque/klak (slap)], pairs of single words [e.g.

couleur/ku.lœ� (colour); lundi/l ~œ.di (monday)], pairs of numbers

(e.g. deux/dø (two); cinq/Ssk (five)] and meaningful short groups of

words [e.g. un bon repas/ ~œ/bO//�@.pa (a good meal); un gros

gateau/ ~œ// _g�o// _ga.to (a big cake)]. Patients were instructed to con-

centrate equally on both stimuli and to repeat them. To familiarize

patients with the procedure, this evaluation was preceded by a brief

period of practice with right and left presentations of eight items (four

single words to the right and four single words to the left).

Individual responses were first scored as the percentage of correctly

repeated words presented to the right and left ear (Table 2).

Performance was then assessed using two main parameters. First, a

quantitative measure was used to assess the patient’s global perfor-

mance. This global quantitative measure was simply the total number

of auditory stimuli correctly reported by the patients, pooled over the

four stimulus categories. Second, we computed a lateralization index

for each patient, a sensitive means of detecting and quantifying asym-

metry in the detection of stimuli presented to the two ears (Bellmann

et al., 2001). The lateralization index = 100� (R� L)/(R + L), where

R and L are the total number of correctly reported stimuli from the

right ear and left ear, respectively. The lateralization index metric

represents the proportionally greater number of right than left ear

stimuli detected, with the difference expressed as a percentage of

the total number of stimuli detected. In addition, incorrect responses

were examined for fusion errors. Fusion errors are verbal responses

where one phoneme from words presented on the left is combined

with words presented on the right to produce a novel combination

that was not actually presented to the patient, though each constitu-

ent part was. Fusion errors imply some degree of processing of the left Tab

le2

Audit

ory

exti

nct

ion

resu

lts

inth

etw

ogro

ups

(‘Pri

sm’

and

‘Contr

ol’

)w

ith

per

centa

ge

of

corr

ectl

yre

pea

ted

stim

uli

asa

funct

ion

of

stim

ula

tion

side

(lef

tor

right)

,ev

aluat

ion

sess

ions

(pre

-tes

t,0

hpost

-tes

tor

2h

post

-tes

t)an

dSt

imulu

sC

ateg

ory

(sin

gle

word

s,gro

ups

of

word

s,phonolo

gic

ally

sim

ilar

word

s,num

ber

s)

Gro

up

Pat

ient

Word

sG

roup

of

word

sPhonolo

gic

ally

sim

ilar

word

sN

um

ber

s

Pre

-tes

t0

hpost

-tes

t2

hpost

-tes

tPre

-tes

t0

hpost

-tes

t2

hpost

-tes

tPre

-tes

t0

hpost

-tes

t2

hpost

-tes

tPre

-tes

t0

hpost

-tes

t2

hpost

-tes

t

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Left

Rag

ht

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Left

Rig

ht

Prism

D.M

.13

97

30

90

10

97

3100

13

100

0100

10

80

30

90

20

80

40

100

50

100

30

100

Prism

B.E

.3

87

13

87

387

10

83

30

97

30

100

10

70

30

80

10

70

10

90

80

100

70

100

Prism

A.E

.7

30

733

13

27

060

073

083

030

10

60

10

50

070

060

10

80

Prism

G.C

.20

97

50

93

60

97

53

100

87

100

77

97

30

70

50

70

40

90

50

90

80

90

90

100

Prism

L.A

.47

70

50

87

57

87

33

97

37

97

33

97

60

50

40

80

40

80

60

100

80

100

80

100

Prism

M.J

.7

87

13

83

17

93

097

10

97

13

100

10

40

10

50

070

10

100

20

100

30

100

Contr

ol

B.B

.10

63

20

67

33

63

20

80

27

87

40

87

40

70

30

70

40

70

70

100

90

100

90

100

Contr

ol

N.G

.7

23

10

30

727

723

10

43

13

37

10

20

10

50

10

60

30

80

20

80

30

100

Contr

ol

B.C

.23

37

17

67

37

60

093

097

393

50

50

20

50

050

10

90

60

90

90

100

Contr

ol

D.J

.0

93

093

093

0100

097

0100

070

080

070

0100

0100

0100

Contr

ol

G.M

.23

87

30

90

20

90

47

70

30

100

13

97

30

70

20

70

20

70

50

100

30

100

20

100

Contr

ol

F.J.

Y.

070

080

783

0100

0100

10

97

080

070

080

0100

0100

0100

Auditory extinction: glasses to hear better? Brain 2010: 133; 895–908 | 899

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

stimulus rather than total extinction. The number of fusions was also

analysed as an additional variable.

Prism adaptation

Subjects in the ‘Prism’ group were exposed to a rightward optical shift

of the visual field produced by the prismatic lenses. Glacier goggles

(Julbo�) were fitted with wide-field, point-to-point wedge lenses

creating an optical shift of 10� (OptiquePeter.com, Lyon) affording

wide binocular vision. The total visual field was 110� and the

one-eye visual field was 80� (including a 50� binocular field). With

these goggles on, the visual field was uniformly displaced to the

right with minimal visual distortion. The exposure period consisted of

50 pointing responses to visual targets presented 10� to the right or to

the left of the objective body midline. During the prism exposure, each

patient was asked to point at a fast but comfortable speed; he/she

could see the target, the second half of the pointing trajectory and the

terminal error. His/her head was kept aligned with the body’s sagittal

axis by a chin-rest and controlled by an investigator throughout the

adaptation procedure. The total duration of this exposure was �8 min.

The dichotic listening task was presented three times: once in the

pre-test (prior to prism exposure) and twice as post-tests (after prism

adaptation): upon prism removal (0 h post-test) and 2 h later (2 h

post-test).

Subjects from the ‘Control’ group underwent the same procedure

except that the goggles were fitted with neutral sham glasses (the

weight of these sham goggles was matched to the real prisms by

having two pairs of 5� prisms for each eye; the bases of these two

prisms were oriented left and right in front of each eye, such that the

two deviations cancelled out, yielding no net optical shift).

Statistical analysesSimple comparisons (e.g. age, delay post-stroke) were performed with

t-tests. All other multiple comparisons of means (e.g. Group, Session

and/or Category) were performed with ANOVAs. Further planned

comparisons of the means were used when appropriate (least-square

method). All analyses were realized in the Statistica version 8.0.

(StatSoft France, 2008). Threshold for statistical significance was set

at 0.05.

ResultsNone of the patients complained about discomfort or the auditory

task’s difficulty and all testing was completed without interruption.

All patients heard and reported 100% of stimuli during the pre-

liminary monaural presentation (used as a familiarization practice)

for both right and left monaural stimuli. As validated by prelimi-

nary testing, the 5 s interval between auditory stimuli allowed

patients to perform the test at a comfortable pace, such that no

data were lost due to time constraints. To validate the occurrence

of substantial prism adaptation in the ‘Prism’ group, a simple clin-

ical procedure was used. Two open-loop pointing trials (i.e. with-

out visual feedback) were performed by each patient at the end of

prism exposure (goggles were removed). We checked that perfor-

mance on these pointing trials was shifted by at least 5� with

respect to the visual target. Using this procedure, we confirmed

that all patients in the ‘Prism’ group showed a sufficient amount

of prism adaptation.

Baseline pre-test evaluation of dichoticlistening performanceA range of measures were used to assess performance in dichotic

listening: average total number of correct responses; lateralization

index for overall task performance and for each Stimulus Category,

lateralization index = 100� (R� L)/(R + L), where R and L are the

total number of correctly reported stimuli from the right ear and left

ear, respectively; and fusion errors, verbal responses where one

phoneme from words presented on the left is used to produce a

new word that was not actually presented to the patient.

In healthy controls (n = 10), the average lateralization index was

1.02 (range from �5.26 to +4.89; SE = 0.99), coherent with the

small right ear advantage that has been described in right-handed

subjects (Michel et al., 1986).

Next, we tested for baseline differences in task performance

between the two patient groups, the ‘Prism’ group and the

‘Control’ group. Global task performance was assessed using a

one-way ANOVA. The average total number of correct responses

obtained in the ‘Prism’ group (mean = 79.7; SE = 9.8) did not differ

significantly [F(1, 10) = 0.66; P40.43] from that obtained in the

‘Control’ group (mean = 69.2; SE = 8.4). A significant left–right

effect, however, was present [F(1, 10) = 73,23, P50.001], coher-

ent with the left extinction observed in all patients (i.e. patients

omit more left than right stimuli). Figure 2 shows the left and right

scores (correct responses) obtained in the two groups, confirming

that the global performance observed before prism adaptation was

identical in the two groups.

The asymmetry of detection performance was quantified using

the lateralization index. All patients presented major left auditory

extinction with an average global lateralization index of +68.5

(SE = 10.7) in the ‘Prism’ group and of +66.4 (SE = 11) in the

‘Control’ group. Figure 3 presents lateralization index values for

Figure 2 A comparison of the two groups’ global performance

as a function of stimulation side [right (R) or left (L)] was

performed in the pre-test session with a one-way ANOVA.

No group effect was obtained but a significant left–right effect

was observed, coherent with the left extinction observed in all

patients. These results confirm that global performance in the

pre-test session did not differ between the two groups.

900 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

each Stimulus Category. Lateralization index values were numeri-

cally smaller for phonologically similar words than the other stim-

ulus categories (in particular as compared to short groups of

words). We tested for any significant differences in performance

between the groups across stimulus categories. An ANOVA with

Group as a between-subject factor and Stimulus Category as a

within-subject factor was performed on the lateralization index

values obtained in the pre-test. There was no significant effect

of Group [F(1, 10) = 0.009, P40.92] or Stimulus Category

[F(3, 30) = 2.11, P40.12], nor was there an interaction.

Although the total number of fusion errors was low, the two

groups were also compared for the raw number of fusion errors

obtained during the pre-test. A marginally significant difference

was found [t(10) =�2.25; P40.048], caused by the ‘Prism’

group (0.83) producing an average of two fewer fusion errors

than the ‘Control’ group (2.83).

Effect of prism adaptation on dichoticlistening performanceThe following analyses compared the two groups of patients

(‘Prism’ and ‘Control’) for their dichotic listening performance

across the three evaluation sessions (pre-test, post-test 0 h and

post-test 2 h, i.e. before and after prism adaptation).

First, a comparison of the two groups’ global performance (total

number of correct responses) was performed over the three eval-

uation sessions with a two-way ANOVA (Fig. 4). No group effect

was observed [F(1, 10) = 1.28; P40.28], but a significant Session

effect was obtained [F(2, 20) = 14.41; P50.001]. Planned compar-

isons showed a significant difference between pre-test and 0 h

post-test for both groups [F(1, 10) = 27.94; P50.0005], as well

as between pre-test and 2 h post-test [F(1, 10) = 16.00;

P50.005]. No difference was found between 0 h and 2 h post-test

[F(1, 10) = 0.21; P40.6]. There was no Group� Session interaction

[F(2, 20) = 1.43; P40.26]. Hence, both groups improved their per-

formance across repeated task sessions, presumably owing to task

practice effects. Although the improvement in the ‘Prism’

group was twice that of the ‘Control’ group (‘Prism’:

pre-test = 79.7� 9.2, 0 h post-test = 93� 9.6, 2 h post-test = 93

� 9.3; ‘Control’: pre-test = 69.2� 9.2, 0 h post-test = 75.3� 9.6,

2 h post-test = 77.2� 9.3), this difference was not significant.

The absence of an interaction suggests that prism adaptation

does not significantly improve global, non-lateralized aspects of

auditory performance.

To assess our primary hypothesis, that prism adaptation would

improve lateralized aspects of task performance (i.e. auditory

extinction), we analysed the lateralization index data. An

ANOVA with Group as the between-subject factor and Session

(pre-test, 0 h post-test and 2 h post-test) and Stimulus Category

as within-subject factors revealed a dramatic improvement of left

auditory extinction after prism adaptation in the ‘Prism’ group

compared to the ‘Control’ group (Fig. 5). There was no main

effect of Group [F(1, 10) = 0.39; P40.54] or Session

[F(2, 20) = 1.09; P40.35]. A main effect of Stimulus Category

[F(3, 30) = 3.41; P50.05] was observed, caused by a smaller later-

alization index for phonologically similar words than for short

groups of words. There was also a significant Category� Session

interaction [F(6, 60) = 2.49; P50.05]. The three-way interaction

was not significant [F(6, 60) = 0.19; P40.97]. Critically, however,

there was a significant Group� Session interaction [F(2, 20) = 3.64;

P50.05]. This interaction was caused by a significant reduction of

the lateralization index for the ‘Prism’ group immediately after the

test (0 h post-test) [planned comparison: F(1, 10) = 7.03; P50.05].

In addition, 2 h later, a trend towards a reduction of the lateraliza-

tion index (that is towards improved performance) could still be

detected (2 h post-test) [F(1, 10) = 2.40; P = 0.076: see Fig. 5].

To investigate whether this change in lateralization index was a

robust effect measurable across all the different stimulus cate-

gories, we plotted data separately for Group, Session and

Figure 3 Auditory extinction level in the two groups during the

pre-test for the different categories of stimulus pairs

(mean� SE). The asymmetry score (lateralization index; LI) of

patients in the pre-test session was similar in the two groups.

gr. words = groups of words; p.s. words = phonologically similar

words.

Figure 4 Global task performance of the two groups across the

three evaluation sessions (mean� SE). No Group effect was

observed, but a significant Session effect was obtained. This

suggests that prism adaptation does not significantly improve

global, non-lateralized aspects of task performance.

Auditory extinction: glasses to hear better? Brain 2010: 133; 895–908 | 901

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

Category (Fig. 6). This confirmed that there was a systematic

improvement in performance for the ‘Prism’ versus ‘Control’

groups across all four categories of stimulus (‘Prism’ group percent

mean improvement at 0 h post-test for words =�21%, groups of

words =�20%, phonologically similar words. =�13%, num-

bers =�36%; ‘Control’ group % mean improvement at post-test

0 h for words = 1%, groups of words = 8%, phonologically similar

words = 35%, numbers =�10%; for all individual patient effect

sizes see Table 3).

The clinical relevance of the present results depends on obser-

ving not only a significant group effect, but also an effect size that

leads to concrete individual benefit. To determine treatment effi-

cacy at the individual level, we counted the number of patients in

each group who exhibited a benefit across test sessions of at least

20%. These results are shown in Table 3 {0 h post-test/pre-test

comparison, post-pre(%) at 0 h i.e. [(LI 0 h post-test)� (LI

pre-test)/LI pre-test]}. In the ‘Prism’ group, all but one patient

improved and three showed an improvement of 420%. In the

‘Control’ group, only one patient showed an improvement

(16%). In the 2 h post-test/pre-test comparison [post-pre(%) at

2 h] [i.e. (LI 2 h post-test)� (LI pre-test)/LI pre] in the ‘Prism’

group, four of the six patients showed an improvement of at

least 10%, with two of those improving by 420%.

As a supplementary analysis, we performed a repeated measure

ANOVA on the number of fusion errors. There was no main effect

of Group [F(1, 10) = 1.44; P40.25] or Session [F(2, 20) = 0.26;

P40.77] but the interaction was marginally significant

[F(2, 20) = 3.42; P = 0.052]. The planned comparison (between

Group and pre-test versus post-test, with the post-test data

pooled over 0 h and 2 h post-test) showed a significant increase

in the number of fusion errors after prism adaptation

[F(1, 10) = 5.87; P50.04] (Fig. 7).

This secondary finding of an increase in the number of fusion

responses, taken together with the improvement in lateralization

index after prism adaptation, indicates that prism adaptation spe-

cifically benefits a lateralized component of patient deficit in audi-

tory extinction. In the context of no baseline difference, and no

global change over time in dichotic listening performance between

the ‘Prism’ and ‘Control’ groups, the present results suggest that

prism adaptation does not induce a general improvement in atten-

tional capacity, but rather specifically redresses an imbalance in

left–right attentional allocation.

As described in the ‘Subjects’ section, the time since stroke

onset tended to be higher in the ‘Prism’ group than in the sham

‘Control’ group. This difference was not significant. There was also

no difference in baseline task performance between the two

groups (Fig. 5). In any case, such a difference in time since

stroke onset would likely predict a more favourable spontaneous

recovery rate in the ‘Control’ group than the ‘Prism’ group, so it

cannot explain the specific improvement in the ‘Prism’ group that

was observed. Alternatively, it could be argued that lesion–age

difference might have contributed to the performance improve-

ment in the ‘Prism’ group. The lesion–age difference was caused

mainly by two patients (B.E. and L.A.) who suffered from their

lesion over 500 days prior to the test. One of these patients (L.A.)

was an outlier in the ‘Prism’ group, but in a direction that wea-

kened the effect. This argues against the possibility that lesion-age

difference between the groups caused the observed effect.

It is conceivable that some form of unconscious bias on the part

of the experimenter might have contributed to the results

observed. That is, this study was not blinded, so the potential

for sources of bias need to be considered. To determine whether

any such effect was likely to have contaminated the data, we

performed an additional experiment. The responses of three new

neglect patients were recorded in the same dichotic listening task,

and the recording was analysed by ten independent hearers.

Results showed remarkably low between-hearer variability in the

assessments. The average standard error of the mean obtained

across the three patients was 0.16 units and the greatest individual

hearer’s deviation from the mean was 1.54 units. These values

Figure 5 Performance asymmetry in the two groups (laterali-

zation index) through the three evaluation sessions (mean� SE).

There was a significant reduction of the lateralization index for

the ‘Prism’ group immediately after prism adaptation (0 h

post-test) compared to the ‘Control’ group, with a trend-level

persistence of the effect two hours after prism adaptation

(2 h post-test).

Figure 6 Lateralization index values for the two groups across

the three evaluation sessions for each Stimulus Category

(mean� SE). gr. words = groups of words; p.s. words =

phonologically similar words.

902 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

are �110 times smaller than the average prism effect.

These additional measurements confirm the reliability of the

assessment procedures used in this study.

DiscussionThe present experiment aimed to test the effect of a visuo-manual

adaptation procedure (prism adaptation) on an auditory task per-

formed by neglect patients. The results clearly show that prism

adaptation can improve the left-sided deficit typically found in

dichotic listening tasks in neglect patients. The findings suggest

that adaptation of visuo-motor coordination can affect perfor-

mance in a sensory modality (audition) that is not directly impli-

cated in the visuo-motor adaptation process per se. This

generalization further underscores the potential of prism adapta-

tion as a clinically effective intervention for post-stroke rehabilita-

tion of neglect.

Prism adaptation improves leftauditory extinctionAuditory extinction was examined in 12 patients with right hemi-

sphere lesions and left visuo-spatial neglect, using bilateral simul-

taneous presentation of verbal stimuli (dichotic listening). The

principal data analysed in this study were patients’ verbal repro-

duction of the auditory stimuli presented to each ear. Left omis-

sions were observed in all patients during bilateral trials. None of

the patients showed omissions in the preliminary monaural test

session, which rules out an explanation of the auditory deficit on

bilateral trials in terms of a peripheral sensory deficit. All

12 patients initially exhibited a marked left ear disadvantage

with significant asymmetry. At the baseline pre-test stage, the

two randomized patient groups were comparable regarding clinical

and paraclinical data and initial auditory extinction level.

After prism adaptation, the reliable left auditory extinction

found in the dichotic listening task was reduced. In contrast,

there was no such improvement in the ‘Control’ group in which

patients performed the same task but wore neutral goggles that

did not induce an optical shift. The improvement in the ‘Prism’

group was measurable as a significant reduction in the lateraliza-

tion index, i.e. an amelioration of the detection asymmetry

between left and right stimuli. In contrast, the global measure of

task performance did not evolve differentially over time between

the two groups. Hence, the lateralization index was more specif-

ically and reliably affected by adaptation than was global task

performance. The robustness of the prism effect was evidenced

by the consistent pattern of results obtained across all stimulus

categories (phonologically similar words, numbers, single words

and short groups of words) and in all but one patient. Therefore

the effects of adaptation generalized across semantic categories.

Moreover, the reliability measures obtained from 10 independent

hearers indicate that these effects are not a simple artefact of

experimenter bias.Tab

le3

Audit

ory

exti

nct

ion

resu

lts

inth

etw

ogro

ups

(‘Pri

sm’an

d‘C

ontr

ol’

)w

ith

abso

lute

late

rali

zati

on

index

score

asa

funct

ion

of

eval

uat

ion

sess

ions

and

stim

ulu

sca

tegory

Gro

up

Pat

ient

Word

sG

roups

of

word

sPbonolo

gic

alty

sim

ilar

word

sN

um

ber

sG

lobal

LIPost

–pre

(%)

Pre

-tes

t0

hpost

-te

st

2h

post

-te

st

Pre

-tes

t0

hpost

-te

st

2h

post

-te

st

Pre

-tes

t0

hpost

-te

st

2h

post

-te

st

Pre

-tes

t0

hpost

-te

st

2h

post

-te

st

Pre

-tes

t0

hpost

-te

st

2h

post

-te

st

At

0h

At

2h

Prism

D.M

.76

50

81

94

76

100

78

50

60

43

33

54

77

57

81

�26

5Prism

B.E

.93

73

93

79

53

54

75

60

75

80

11

18

84

52

60

�38

�29

Prism

A.E

.64

67

33

100

100

100

100

71

67

100

100

78

90

87

77

�3

�14

Prism

G.C

.66

30

23

30

712

40

17

38

29

65

43

16

18

�63

�58

Prism

L.A

.20

27

21

49

45

49

�9

33

33

25

11

11

29

32

30

10

3Prism

M.J

.86

72

70

100

81

76

60

67

100

82

67

54

89

75

72

�16

�19

Mea

n(S

EM)

67

(10.5

)53

(8.5

)53

(12.8

)75

(11.9

)60(1

3.4

)65

(13.9

)57(1

5.

0)

50

(8.6

)62

(10)

60

(12.9

)38

(15.5

)37(1

19)

68

(10.7

)53

(10.

8)

56

(10.

8)

Post

–pre

(%)

�21

�21

�20

�13

�13

8�

36

�39

�23%

(11)�

18%

(10)

Contr

ol

B.B

.73

54

31

60

53

37

28

40

27

18

55

50

42

28

�16

�44

Contr

ol

N.G

.56

50

60

56

62

47

33

67

71

45

60

54

50

59

56

18

12

Contr

ol

B.C

.22

60

24

100

100

93

043

100

80

20

561

66

49

8�

20

Contr

ol

D.J

.100

100

100

100

100

100

100

100

100

100

100

100

100

100

100

00

Contr

ol

G.M

.58

50

64

20

54

76

40

56

56

33

54

67

38

53

68

39

79

Contr

ol

F.J.

Y.

100

100

85

100

100

81

100

100

100

100

100

100

100

100

87

0�

13

Mea

n(S

EM)

68

(12.2

)69

(9.9

)61

(12.1

)73

(13.

5)

78

(9.8

)72

(10.3

)50

(16.

7)

68

(11)

76

(12.3

)63

(14.

5)

57

(16.1

)55

(17.5

)66

(11)

70(1

0.1

)65

(10.7

)Post

–pre

(%)

1�

11

80

35

51

�10

�12

8(8

)2

(17)

Posi

tive

and

neg

ativ

ech

arge

inla

tera

lizat

ion

index

valu

esco

rres

pond

toan

aggra

vation

or

anim

pro

vem

ent

of

extinct

ion,

resp

ective

ly.

Although

task

per

form

ance

by

the

contr

olgro

up

chan

ges

afte

rth

esh

amin

terv

ention

asa

funct

ion

of

Stim

ulu

sC

ateg

ory

,th

e‘P

rism

’gro

up

mai

nta

ins

its

adva

nta

ge

afte

rprism

adap

tation,

irre

spec

tive

of

Stim

ulu

sC

ateg

ory

.

Auditory extinction: glasses to hear better? Brain 2010: 133; 895–908 | 903

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

Prism adaptation effect on dichoticlistening excludes a simple left earsensory deficitWhen using dichotic stimuli, one issue to address is whether

extinction for left stimuli reflects poor or abolished processing of

auditory information entering the contra-lesional ear, or higher

level attentional disturbance involving contra-lesional space

(Bellmann et al., 2001; Beaton and McCarthy, 1995). Several stu-

dies have aimed at specifying the characteristics of extinguished

stimuli in order to localize the level at which the auditory process-

ing deficit arises (Deouell and Soroker, 2000; Deouell et al., 2000;

Shisler et al., 2004). The findings parallel previous studies of visual

extinction, suggesting that auditory extinction may be due, in part,

to a failure to bind together information about stimulus identity

and stimulus location.

In assessing auditory attention deficits (as in extinction), it is

important to rule out potential peripheral hearing deficits. This

can be achieved by presenting auditory stimuli diotically. Diotic

presentation involves that each stimulus reaches both ears, with

an interaural time difference serving as the only lateralization cue

(see Bellman et al., 2001; Spierer et al., 2007). In such cases, there

is no difference of content because both ears receive both items of

each pair at the same intensity level, allowing control for sensory

deficits. In dichotic presentation two different stimuli are pre-

sented, one to the right ear and one to the left one, implying

that different content reaches each ear. In the present study, we

used dichotic presentation. However, the key finding in our study

was a change in dichotic listening performance after prism adap-

tation. It does not seem plausible that prism adaptation could

cause such a change in patients suffering from unilateral sensory

hearing loss. The fact that five out of six patients in the ‘Prism’

group consistently showed an improvement after the adaptation

procedure suggests that the auditory extinction deficit in these

patients is a consequence of higher level deficits in contra-lesional

spatial processing. The finding that prism adaptation improves

auditory extinction suggests a locus of effect at a higher level

of sensory integration; whereby adaptation shifts the spatial

alignment between vision and proprioception, with apparent

knock-on benefits for location processing in other sensory domains

and for spatial cognition more generally.

Cross-modal effects of prism adaptationChanges in visuo-motor (target pointing) performance following

prism adaptation are believed to arise through changes in the

mapping between visual, proprioceptive and motor processing

(Redding et al., 2005). After-effects of these changes in

visuo-motor performance have, however, been demonstrated to

generalize over a range of tasks beyond the pointing procedure

that is directly adapted. Benefits for unilateral neglect patients

after prism adaptation have been shown on line bisection and

cancellation tasks, postural imbalance, spatial judgement tasks,

reading tasks and even wheel-chair driving (Rossetti et al.,

1998; Tilikete et al., 2001; Farne et al., 2002; MacIntosh et al.,

2002; Pisella et al., 2002; Jacquin-Courtois et al., 2008). Since

each of these tasks involve some form of visual or motor process-

ing, it is possible that performance improvements may arise

directly from changes to visuo-motor processes that are to a

greater or lesser extent directly altered by the prism adaptation

procedure itself. Mental imagery (Rode et al., 1998, 2001, 2007)

has also been found to be improved following adaptation, but

visuo-spatial imagery tasks are known to recruit the visual

system (Kosslyn et al., 1999). The amelioration of number bisec-

tion in neglect patients (Rossetti et al., 2004) may also result from

an activation of visuo-spatial representations via spatial-numerical

association (Hubbard et al., 2005; Lacour et al., 2005; Rode et al.,

2007). A few reports have described an amelioration of

somato-sensory processing in neglect patients following prism

adaptation. First, the haptic circle centring task described by

MacIntosh et al. (2002) involves a strong proprioceptive compo-

nent, which might have been directly altered by prism adaptation.

Second, Maravita et al. (2003) tested tactile extinction in four

neglect patients and found an improvement of contra-lesional tac-

tile perception in all patients after adaptation to a rightward pris-

matic shift. Third, Dijkerman et al. (2004) described a patient who,

despite having recovered from visual neglect, still exhibited a

somato-sensory deficit on the left side. After prism adaptation,

the patient’s tactile and proprioceptive thresholds were signifi-

cantly lowered. These last two studies are interesting because

they demonstrate a beneficial effect of visuo-manual adaptation

on tactile sensitivity, a modality that appears not to be directly

modified by prism adaptation. However, it can be hypothesized

that cross-talk between proprioceptive and tactile processes may

explain these effects.

In contrast, in the present study, beneficial effects of prism

adaptation were observed on auditory extinction, a modality

that is independent from visual, somato-sensory and manual com-

ponents. In addition, the semantic nature of the task (verbal

reproduction of words) clearly distinguishes it from spatial tasks.

These new results show that the effects of prism adaptation are

not restricted to visuo-motor tasks but can also affect perception

in a non-adapted sensory modality. Prism adaptation is believed to

alter the mapping between proprioceptive and visual modalities

directly, and both visual and somato-sensory processing are

Figure 7 Quantitative evaluation of the two groups’ fusion

errors across the three evaluation sessions (mean� SE).

904 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

intrinsically bound to the action system. This is clearly not the case

for auditory verbal recognition. There is no obvious functional link

between the adaptation of a visuo-manual pointing task to a lat-

eral visual shift and the ability to identify verbal stimuli from the

left ear. Therefore the present results demonstrate a striking

cross-modal transfer of prismatic after-effects to the auditory

domain.

The present auditory effect suggests that the lateralized remap-

ping of visuo-motor information induced by prism adaptation may

subsequently alter the orientation of attention in other sensory

modalities. Eye movement modulation could be one potential

mechanism for such effects. However, it has been shown that

the visuo-motor and cognitive effects of prism adaptation can

be double-dissociated (Dijkerman et al., 2003; Ferber et al.,

2003, but see Serino et al., 2006). Dijkerman et al. (2003)

showed that prism adaptation can reduce ocular exploration asym-

metry without affecting size under-estimation in a neglect patient.

In addition, Ferber et al. (2003) reported on a neglect patient

exposed to a prismatic deviation, who exhibited a shift in explor-

atory eye movements towards the left, but without a concomitant

improvement of the deficit in stimulus identification on the left

side. So far, no clear link has been demonstrated between mod-

ification of oculomotor patterns and clinical effects induced by

prism adaptation.

A crucial feature of the present study is that the amelioration

found in the ‘Prism’ group was specific to the lateralization index.

No significant changes were obtained for the total number of

responses. This suggests that prism adaptation did not act through

a general arousal enhancement, but rather through a specific later-

alized mechanism. However, several lines of argument suggest

that the after-effects of prism adaptation produce widespread

changes in spatial cognition that are not purely restricted to the

lateralized spatial processing deficits observed in unilateral neglect.

First, non-visuo-motor, cognitive effects of prism adaptation have

been demonstrated in healthy control subjects (Colent et al.,

2000; Michel et al., 2003a, b, 2006; Loftus et al., 2008).

Second, beneficial effects of prism adaptation have been reported

for non-neglect symptoms and non-neglect patients (Tilikete

et al., 2001; Striemer and Danckert, 2007; Sumitani et al.,

2007). Tilikete et al. (2001) tested the effect of prism adaptation

on postural imbalance in left-hemiparetic patients who did not

have neglect at the time of testing. Postural imbalance was

reduced following adaptation to the right (but not the left),

suggesting a recalibration of the representational distortion in

brain-damaged patients. Striemer and Danckert (2007) demon-

strated that rightward prism adaptation can reduce both the right-

ward attentional bias and the disengage deficit in right

brain-damaged patients, irrespective of the presence of neglect.

Finally, Sumitani et al. (2007) showed that adaptation to a pris-

matic displacement of the visual field toward the unaffected side

can alleviate pathological pain in patients with complex regional

pain syndrome, possibly through an attentional effect. The fact

that prism adaptation after-effects generalize across such a wide

range of functions suggests that adaptation may modify a

common level of spatial representation important for multisensory

integration (Rossetti et al., 1998). That adaptation can also

improve higher order aspects of spatial cognition (e.g. mental

imagery, constructional deficits; Rode et al., 1998), both by redu-

cing lateralized bias and enlarging the represented space, further

suggests that prism adaptation might be used to treat not only

neglect, but a range of spatial cognition disorders and visuo-spatial

dysfunctions (Rode et al., 2006a, 2007). The question arises about

the possible neural substrates of this higher order representation.

The current model of prism adaptation suggests that

prism-induced cerebellar effects interact with contra-lateral poste-

rior parietal cortex (Pisella et al., 2006). The implication of such a

lateralized cerebello-cerebral network has been recently confirmed

by a neuroimaging study in neglect patients (Luaute et al.,

2006b). Interestingly, posterior parietal cortex is not only involved

in visual attention and spatial domains, but could also mediate

auditory attention (Shomstein and Yantis, 2006).

The present study focused on ameliorating neglect patients’

inability to report contra-lesional stimuli when a concurrent

target is presented in ipsilesional space (auditory extinction).

However, neglect patients have also been shown to have deficits

for single auditory targets in terms of correctly localizing them in

space (e.g. Bisiach et al., 1984; Pavani et al., 2001, 2005). Such

deficits have been characterized as reflecting increased uncertainty

for the spatial location of contra-lesional sounds, rather than mere

mislocalization along the horizontal dimension (Pavani et al.,

2002). Future studies should examine whether prism adaptation

can also improve these neglect-related deficits for sound

localization.

ConclusionIn this randomized controlled study, a beneficial effect of prism

adaptation was observed on auditory extinction in a group of

chronic stroke patients. The lateralization index (detection asym-

metry on bilateral trials) was the specific variable improved after

adaptation. The positive effect was significant immediately after

prism adaptation and tended to be maintained over at least 2 h, as

has been previously reported for several other tasks (Rossetti et

al., 1998; Tilikete et al., 2001; Jacquin-Courtois et al., 2008;

review in Rode et al., 2003, 2006a; Luaute et al., 2006a; Pisella

et al., 2006). It is interesting, and largely unexpected, that prism

adaptation produced such a marked and lasting beneficial effect

in the auditory domain, and further indicates the potential value

of prism adaptation for the amelioration of a variety of

neglect-related deficits (not restricted to the visuo-motor

domain). In addition, the beneficial effects of prisms for neglect

and neglect-related symptoms appear to last longer than the pos-

itive consequences observed after other interventions such as ves-

tibular stimulation or trunk rotation. Clinical efficacy could be

further increased by use of repeated prism interventions, which

might result in even longer lasting benefits (Shiraishi et al.,

2008; Serino et al., 2009).

The fact that prism adaptation after-effects generalize across

tasks of various levels of complexity and/or several sensory mod-

alities implies a restructuring of high-level spatial representations.

This conclusion should open a new orientation of rehabilitation

strategies and widen the potential scope of application of

‘bottom-up’ therapeutic approaches. Rather than to rehabilitate

an injured process directly, prism adaptation suggests that using

Auditory extinction: glasses to hear better? Brain 2010: 133; 895–908 | 905

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

plasticity may provide a new processing route for affected sensory

information. It may activate brain functions related to multisensory

integration necessary for spatial representations and gating the

access of sensory information to the higher levels of spatial

integration.

AcknowledgementsThe authors would like to thank an anonymous referee for raising

the issue that the study was not blinded and therefore potential

for sources of bias needed to be considered.

FundingThis work was supported by funds from Hospices Civils de Lyon

(Thematique Prioritaire), from Inserm and from Universite

de Lyon.

ReferencesAlbert M. A simple test of visual neglect. Neurology 1973; 23: 658–64.

Beaton A, McCarthy M. "Auditory neglect after right frontal lobe and

right pulvinar thalamic lesions": comments on Hugdahl, Wester, and

Asbjornsen (1991) and some preliminary findings. Brain Lang 1993;

44: 121–6.

Beaton A, McCarthy M. On the nature of auditory neglect: a reply to

Hugdahl and Wester. Brain Lang 1995; 48: 351–8.

Bellmann A, Meuli R, Clarke S. Two types of auditory neglect. Brain

2001; 124: 676–87.

Bender MB, Diamond SP. An analyzis of auditory perceptual defects with

observations on the localization of the dysfunction. Brain 1965; 88:

675–86.Bisiach E, Luzzatti C. Unilateral neglect of representational space. Cortex

1978; 14: 129–33.

Bisiach E, Cornacchia L, Sterzi R, Vallar G. Disorders of perceived audi-

tory lateralization after lesions of the right hemisphere. Brain 1984;

107: 37–52.

Brozzoli C, Dematte ML, Pavani F, Frassinetti F, Farne A. Neglect and

extinction: within and between sensory modalities. Restor Neurol

Neurosci 2006; 24: 217–32.

Calabria M, Michel C, Honore J, Guillaume A, Pisella L, Luaute J, et al.

Prism adaptation and the lack of awareness in spatial neglect. In.

Proceedings of the First Congress of the European

Neuropsychological Societies. Modena: Italy; 2004. p. 222.

Calamaro N, Soroker N, Myslobodsky MS. False recovery from auditory

hemineglect produce by source misattribution of auditory stimuli (the

ventriquolist effect). Restor Neurol Neurosci 1995; 7: 151–6.Cappa S, Sterzi R, Vallar G, Bisiach E. Remission of hemineglect and

anosognosia during vestibular stimulation. Neuropsychologia 1987;

25: 775–82.

Colent C, Pisella L, Bernieri C, Rode G, Rossetti Y. Cognitive bias induced

by visuo-motor adaptation to prisms: a simulation of unilateral neglect

in normal individuals? Neuroreport 2000; 11: 1899–902.Cusack R, Carlyon RP, Robertson IH. Neglect between but not within

auditory objects. J Cogn Neurosci 2000; 12: 1056–65.

Damasio H, Damasio AR. Lesion analysis in neuropsychology. New York:

Oxford University Press; 1989.

Denes G, Semenza C, Stoppa E, Lis A. Unilateral spatial neglect and

recovery from hemiplegia. A follow-up study. Brain 1982; 105:

543–52.

Deouell LY, Soroker N. What is extinguished in auditory extinction?

Neuroreport 2000; 11: 3059–62.

Deouell LY, Bentin S, Soroker N. Electrophysiological evidence for an

early (pre-attentive) information processing deficit in patients with

right hemisphere damage and unilateral neglect. Brain 2000; 123:

353–65.

De Renzi E, Gentilini M, Barbieri C. Auditory neglect. J Neurol Neurosurg

Psychiatry 1989; 52: 613–7.

De Renzi E, Gentilini M, Pattacini F. Auditory extinction following hemi-

sphere damage. Neuropsychologia 1984; 22: 733–44.

Dijkerman HC, Webeling M, ter Wal JM, Groet E, van Zandvoort MJ. A

long-lasting improvement of somatosensory function after prism adap-

tation, a case study. Neuropsychologia 2004; 42: 1697–702.

Dijkerman HC, MacIntosh RD, Milner AD, Rossetti Y, Tilikete C,

Roberts RC. Ocular scanning and perceptual size distortion in hemi-

spatial neglect: effects of prism adaptation and sequential stimulus

presentation. Exp Brain Res 2003; 153: 220–30.

Driver J, Mattingley JB. Parietal neglect and visual awareness. Nat

Neurosci 1998; 1: 17–22.

Farne A, Rossetti Y, Toniolo S, Ladavas E. Ameliorating neglect with

prism adaptation: visuo-manual and visuo-verbal measures.

Neuropsychologia 2002; 40: 718–29.Ferber S, Danckert J, Joanisse M, Goltz HC, Goodale MA. Eye move-

ments tell only half the story. Neurology 2003; 60: 1826–9.Frassinetti F, Angeli V, Meneghello F, Avanzi S, Ladavas E. Long-lasting

amelioration of visuospatial neglect by prism adaptation. Brain 2002;

125: 608–23.

Gainotti G, Messerli P, Tissot R. Qualitative analysis of unilateral spatial

neglect in relation to laterality of cerebral lesions. J Neurol Neurosurg

Psychiatry 1972; 35: 545–50.

Guariglia C, Piccardi L, Iaria G, Nico D, Pizzamiglio L. Representational

neglect and navigation in real space. Neuropsychologia 2005; 43:

1138–43.

Halligan PW, Marshall JC. Spatial compression in visual neglect: a case

study. Cortex 1991; 27: 623–9.

Heilman KM, Valenstein E. Auditory neglect in man. Arch Neurol 1972;

26: 32–5.

Heilman KM, Bowers D, Coslett HB, Whelan H, Watson RT. Directional

hypokinesia: prolonged reaction times for leftward movements in

patients with right hemisphere lesions and neglect. Neurology 1985;

35: 855–9.

Hubbard EM, Piazza M, Pinel P, Dehaene S. Interactions between

number and space in parietal cortex. Nat Rev Neurosci 2005; 6:

435–48.

Hugdahl K, Wester K, Asbjornsen A. Auditory neglect after right frontal

lobe and right pulvinar thalamic lesions. Brain Lang 1991; 41: 465–73.

Humphreys GW, Watelet A, Riddoch MA. Long-term effects of prism

adaptation in chronic visual neglect: a single case study. Cognitive

Neuropsychol 2006; 23: 463–78.Husain M, Kennard C. Visual neglect associated with frontal lobe infarc-

tion. J Neurol 1996; 243: 652–7.Jacquin-Courtois S, Rode G, Boisson D, Rossetti Y. Wheel-chair driving

improvement following visuo-manual prism adaptation. Cortex 2008;

44: 90–6.

Jehkonen M, Ahonen JP, Dastidar P, Koivisto AM, Laippala P, Vilkki J,

et al. Visual neglect as a predictor of functional outcome one year

after stroke. Acta Neurol Scand 2000; 101: 195–201.Karnath HO, Ferber S, Himmelbach M. Spatial awareness is a function of

the temporal not the posterior parietal lobe. Nature 2001; 411: 950–3.Kerkhoff G, Artinger F, Ziegler W. Contrasting spatial hearing deficits in

hemianopia and spatial neglect. Neuroreport 1999; 10: 3555–60.Kosslyn SM, Pascual-Leone A, Felician O, Camposano S, Keenan JP,

Thompson WL, et al. The role of area 17 in visual imagery: convergent

evidence from PET and rTMS. Science 1999; 284: 167–70.

Lacour S, Jacquin-Courtois S, Rossetti Y. Liens entre representations spa-

tiales et representations numeriques: apport de la negligence spatiale

unilaterale et de la plasticite visuo-motrice. In: Coello Y, Casalis S,

906 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

Moroni C, editors. Vision, espace et cognition: fonctionnement normal

et pathologique. Septentrion: Presses Universitaires, Psychologie; 2005.

p. 61–72.

Loftus AM, Nicholls ME, Mattingley JB, Bradshaw JL. Left to right: repre-

sentational biases for numbers and the effect of visuomotor adapta-

tion. Cognition 2008; 107: 1048–58.

Luaute J, Halligan P, Rode G, Jacquin-Courtois S, Boisson D. Prism adap-

tation first among equals in alleviating left neglect: a review. Restor

Neurol Neurosci 2006a; 24: 409–18.

Luaute J, Michel C, Rode G, Pisella L, Jacquin-Courtois S, Costes N, et al.

Functional anatomy of the therapeutic effects of prism adaptation on

left neglect. Neurology 2006b; 66: 1859–67.

MacIntosh RD, Rossetti Y, Milner AD. Prism adaptation improves chronic

visual and haptic neglect: a single case study. Cortex 2002; 38:

309–20.

Maravita A, McNeil J, Malhotra P, Greenwood R, Husain M, Driver J.

Prism adaptation can improve contralesional tactile perception in

neglect. Neurology 2003; 60: 1829–31.

Marshall JC, Halligan PW. Blindsight and insight in visuo-spatial neglect.

Nature 1988; 336: 766–7.Mattingley JB, Bradshaw JL, Phillips JG. Impairments of movement initia-

tion and execution in unilateral neglect. Directional hyokinesia and

bradykinesia. Brain 1992; 115: 1849–74.Mattingley JB, Husain M, Rorden C, Kennard C, Driver J. Motor role of

human inferior parietal lobe revealed in unilateral neglect patients.

Nature 1998; 392: 179–82.Michel C. Simulating unilateral neglect in normals: myth or reality?

Restor Neurol Neurosci 2006; 24: 419–30.Michel F, Mauguiere F, Duquesnel J. Listening with one or two hemi-

spheres: verbal dichotic testing after intracarotid barbiturate injection.

Neuropsychologia 1986; 24: 271–6.Michel C, Rossetti Y, Rode G, Tilikete C. After-effects of visuo-manual

adaptation to prisms on body posture in normal subjects. Exp Brain Res

2003a; 148: 219–26.Michel C, Pisella L, Prablanc C, Rode G, Rossetti Y. Enhancing visuomo-

tor adaptation by reducing error signals: single-step (aware) versus

multiple-step (unaware) exposure to wedge prisms. J Cogn Neurosci

2007; 19: 341–50.Michel C, Pisella L, Halligan PW, Luaute J, Rode G, Boisson D, et al.

Simulating unilateral neglect in normals using prism adaptation: impli-

cations for theory. Neuropsychologia 2003b; 41: 25–39.Milner AD, MacIntosh RD. The neurological basis of visual neglect.

Curr Opin Neurol 2005; 18: 748–53.

Milner AD, Harvey M, Roberts RC, Forster SV. Line bisection errors in

visual neglect: misguided action or size distorsion? Neuropsychologia

1993; 31: 39–49.

Pavani F, Farne A, Ladavas E. Poor hand-pointing to sounds in right

brain-damaged patients: not just a problem of spatial-hearing. Brain

Cogn 2005; 59: 215–24.

Pavani F, Ladavas E, Driver J. Selective deficit of auditory localisation in

patients with visuospatial neglect. Neuropsychologia 2002; 40:

291–301.

Pavani F, Ladavas E, Driver J. Auditory and multisensory aspects of

visuospatial neglect. Trends Cogn Sci 2003; 7: 407–14.

Pavani F, Meneghello F, Ladavas E. Deficit of auditory space perception

in patients with visuospatial neglect. Neuropsychologia 2001; 39:

1401–9.

Pavani F, Husain M, Ladavas E, Driver J. Auditory deficits in visuospatial

neglect patients. Cortex 2004; 40: 347–65.

Pisella L, Rode G, Farne A, Boisson D, Rossetti Y. Dissociated long

lasting improvements of straight-ahead pointing and line bisection

tasks in two hemineglect patients. Neuropsychologia 2002; 40:

327–34.

Pisella L, Rode G, Farne A, Tilikete C, Rossetti Y. Prism adaptation in the

rehabilitation of patients with visuo-spatial cognitive disorders. Curr

Opin Neurol 2006; 19: 534–42.

Redding GM, Rossetti Y, Wallace B. Applications of prism adaptation: a

tutorial in theory and method. Neurosci Biobehav Rev 2005; 29:

431–44.

Revol P, Pisella L, Luaute J, Jacquin-Courtois S, Farne A, Ota H, et al.

Prism adaptation therapy for unilateral neglect: from the facts to a

bottom-up explanatory model. In: Wu JL, Ito K, Tobimatsu S,

Nishida T, Fukuyama H, editors. Complex medical engineering. New

York: Springer Science; 2007. p. 413–22.

Robertson IH, Manly T, Beschin N, Daini R, Haeske-Dewick H,

Homberg V, et al. Auditory sustained attention is a marker of unilat-

eral spatial neglect. Neuropsychologia 1997; 35: 1527–32.

Rode G, Perenin MT, Boisson D. Negligence de l’espace represente mise

en evidence par l’evocation mentale de la carte de France. Rev Neurol

1995; 151: 161–64.

Rode G, Rossetti Y, Boisson D. Prism adaptation improves representa-

tional neglect. Neuropsychologia 2001; 39: 1250–4.

Rode G, Rossetti Y, Li L, Boisson D. Improvement of mental imagery

after prism exposure in neglect: a case study. Behav Neurol 1998; 11:

251–58.

Rode G, Klos T, Courtois-Jacquin S, Rossetti Y, Pisella L. Neglect and

prism adaptation: a new therapeutic tool for spatial cognition disor-

ders. Restor Neurol Neurosci 2006a; 24: 347–56.Rode G, Pisella L, Rossetti Y, Farne A, Boisson D. Bottom-up transfer

of sensory-motor plasticity to recovery of spatial cognition:

visuomotor adaptation and spatial neglect. Prog Brain Res 2003;

142: 273–87.

Rode G, Pisella L, Marsal L, Mercier S, Rossetti Y, Boisson D. Prism

adaptation improves spatial dysgraphia following right brain damage.

Neuropsychologia 2006b; 44: 2487–93.

Rode G, Jacquin-Courtois S, Revol P, Pisella L, Sacri AS, Boisson D, et al.

Bottom-up effects of sensory conflict and adaptation on mental ima-

gery: sensorimotor grounds for high level cognition? In: Mast FW,

Jancke L, editors Spatial processing in navigation, imagery and percep-

tion. New York: Springer Science; 2007. p. 369–87.

Rossetti Y, Rode G. Reducing spatial neglect by visual and other sensory

manipulations: non-cognitive (physiological) routes to the rehabilita-

tion of a cognitive disorder. In: Karnath HO, Milner AD, Vallar G,

editors. The cognitive and neural bases of spatial neglect. New York:

Oxford University Press; 2002. p. 375–96.

Rossetti Y, Jacquin-Courtois S, Rode G, Ota H, Michel C, Boisson D.

Does action make the link between number and space representation?

Visuo-manual adaptation improves number bisection in unilateral

neglect. Psychol Sci 2004; 15: 426–30.

Rossetti Y, Rode G, Pisella L, Farne A, Li L, Boisson D, et al. Prism

adaptation to a rightward optical deviation rehabilitates left hemispa-

tial neglect. Nature 1998; 395: 166–9.

Ruff RM, Hersh NA, Pribram KH. Auditory spatial deficits in the personal

and extrapersonal frames of reference due to cortical lesions.

Neuropsychologia 1981; 19: 435–43.

Schenkenberg T, Bradford DC, Ajax ET. Line bisection and unilateral

visual neglect in patients with neurologic impairment. Neurology

1980; 30: 509–17.

Shomstein S, Yantis S. Parietal cortex mediates voluntary control of spa-

tial and nonspatial auditory attention. J Neurosci 2006; 26: 435–9.

Serino A, Angeli V, Frassinetti F, Ladavas E. Mechanisms underlying

neglect recovery after prism adaptation. Neuropsychologia 2006; 44:

1068–78.

Serino A, Barbiani M, Rinaldesi ML, Ladavas E. Effectiveness of prism

adaptation in neglect rehabilitation: a controlled trial study. Stroke

2009; 40: 1392–8.Shiraishi H, Yamakawa Y, Itou A, Muraki T, Asada T. Long-term effects

of prism adaptation on chronic neglect after stroke.

NeuroRehabilitation 2008; 23: 137–51.Shisler RJ, Gore CL, Baylis GC. Auditory extinction: the effect of

stimulus similarity and task requirements. Neuropsychologia 2004;

42: 836–46.

Auditory extinction: glasses to hear better? Brain 2010: 133; 895–908 | 907

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from

Soroker N, Calamaro N, Myslobodsky M. ‘‘McGurk illusion’’ to bilateraladministration of sensory stimuli in patients with hemispatial neglect.

Neuropsychologia 1995; 33: 461–70.

Soroker N, Calamaro N, Glicksohn J, Myslobodsky MS. Auditory inatten-

tion in right-hemisphere-damaged patients with and without visualneglect. Neuropsychologia 1997; 35: 249–56.

Spierer L, Meuli R, Clarke S. Extinction of auditory stimuli in hemineglect:

Space versus ear. Neuropsychologia 2007; 45: 540–51.

Striemer C, Danckert J. Prism adaptation reduces the disengagedeficit in right brain damage patients. Neuroreport 2007; 18:

99–103.

Sumitani M, Rossetti Y, Shibata M, Matsuda Y, Sakaue G, Inoue T, et al.Prism adaptation to optical deviation alleviates pathologic pain.

Neurology 2007; 68: 128–33.

Tanaka H, Hachisuka K, Ogata H. Sound lateralisation in patients with

left or right cerebral hemispheric lesions: relation with unilateral visuos-patial neglect. J Neurol Neurosurg Psychiatry 1999; 67: 481–6.

Tilikete C, Rode G, Rossetti Y, Pichon J, Li L, Boisson D. Prism adaptation

to rightward optical deviation improves postural imbalance in left-

hemiparetic patients. Curr Biol 2001; 11: 524–8.Vallar G, Guariglia C, Nico D, Bisiach E. Spatial hemineglect in back

space. Brain 1995; 118: 467–72.

908 | Brain 2010: 133; 895–908 S. Jacquin-Courtois et al.

at INS

ER

M on A

pril 5, 2010 http://brain.oxfordjournals.org

Dow

nloaded from