Neuropsychologia 41 (2003) 1254–1261

Preserved use of spatial cues for sound segregation
in a case of spatial deafness
Anne Bellmann Thiran a,b , Stephanie Clarke a,∗
a Division de Neuropsychologie, Centre Hospitalier Universitaire Vaudois, 1011 Lausanne, Switzerland
b Unité de Neuropsychologie, Clinique Romande de Réadaptation, Sion, Switzerland

Received 21 May 2002; received in revised form 2 September 2002; accepted 3 January 2003

Abstract
Auditory spatial cues contribute to sound localisation and to sound object segregation. We have investigated these capacities in a patient
(NM) who complained having difficulties to localise sounds in everyday life after a right temporo-parieto-frontal ischemic lesion. Two
groups of tasks were used, in which spatial dimension was simulated by interaural time differences (ITD): (i) active localisation of stationary
or moving sound targets, and (ii) sound segregation on the basis of spatial cues. This latter included a spatial release from masking paradigm
and two ITD diotic tasks. NM failed to localise stationary and moving sounds: she perceived all the stimuli at the centre of the head, and
could not differentiate stationary from moving targets. In contrast, NM was able to use ITD cues to segregate simultaneous sound sources
in the spatial-release-from-masking paradigm and in ITD diotic tasks.
These results suggest that sound localisation and sound object segregation based on spatial cues do not rely on the same mechanisms.
© 2003 Elsevier Science Ltd. All rights reserved.
Keywords: Sound localisation; Auditory spatial cues; Spatial-release-from-masking; Auditory attention; Brain damage; Human

1. Introduction
Auditory spatial information is conveyed by interaural
time differences (ITD) and interaural intensity differences
(IID), and attenuation patterns by the outer ear [7]. Psychophysical studies in normal subjects suggest two different
roles for auditory spatial cues. One is auditory localisation, i.e. the ability to attribute a precise egocentric spatial
co-ordinate to a sound, involving overt perception of sound
location. Testing paradigms include pointing to real or
apparent sound sources [29] or discrimination of two successive sound locations [30]. Auditory spatial information
can also be used as a cue for sound object segregation, also
referred to as ‘auditory streaming’ [8,45] or ‘cocktail party
effect’ [13]. Spatial cues facilitate the grouping of sound
components that belong to the same sound object and the
distinction from other sound objects. Segregation of sound
objects contributes decisively to sound recognition in noisy
environment.
Overt sound localisation was reported to be disturbed by
brainstem [1,20,33] and by hemispheric lesions. Deficits
Abbreviations: IID, interaural intensity differences; ITD, interaural
temporal differences
∗ Corresponding author. Tel.: +41-21-314-1310; fax: +41-21-314-1319.
E-mail address: stephanie.clarke@chuv.hospvd.ch (S. Clarke).

after right-hemispheric lesions were reported either in a
free field setting [25,32,35,46], or by means of stereophonic simulation manipulating ITD [3,16,23,40] or IID
[6,37]. Sound localisation impairments were also found
after left-hemispheric damage (free field [31,36]; ITD simulations [15]). Complete inability to perceive the spatial
dimension of sounds, which can be called ‘spatial deafness’,
has been described in a few studies [1,20,33,40].
Brain lesions can also affect sound objects segregation
[12,14,18]. It is not known, however, if sound localisation
and spatial segregation of sound objects rely on the same
mechanisms and structures. We addressed this question by
investigating these two aspects of auditory spatial processing
in a patient who complained having difficulties localising
sounds.

2. Case description
NM was a 37-year-old, right-handed woman, with a short
formal education, but good general intelligence since she
worked as a cashier in a supermarket before and after her
illness. She had sustained an infarction in the territory of
the right middle cerebral artery from cardio-embolic origin
(permeable foramen). The CT-scan, performed before the

0028-3932/03/$ – see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0028-3932(03)00014-9

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261

1255

Fig. 1. NM’s lesions acquired through CT-scan and normalised in the Talairach space [39].

hospitalisation, showed right-hemispheric lesions located in
the inferior and middle frontal gyri, inferior parietal lobule,
insula and superior temporal region, including Heschl’s
gyrus (Fig. 1). Neuropsychological examination in postacute stage revealed mild signs of right-hemispheric damage,
characterised by left-ear extinction on a verbal dichotic
listening task, constructional apraxia and topographical
disorientation on geographical maps.
NM was referred to our division, 9 years later, for neuropsychological evaluation. The neurological examination
showed normal and symmetrical strength and sensitivity, the
EEG examination was normal. At that time, the patient lived
independently, took care of her four children, and worked at
50% as a cashier. She complained, however, of difficulties
to localise sounds. She was afraid to cross a road, because
she could not trust auditory cues and had to look to the right,
to the left, and again to the right before being confident to
cross. The neuropsychological examination revealed significant cognitive impairment characterised by emotional dullness, constructional apraxia, severe verbal and visuo-spatial
memory impairment, and mild executive disorder (slowness,
failure of inhibition, attentional deficits). A mild tendency
to visuo-spatial neglect was also suggested by the initiation
of the bells cancellation task from the right [34], but the patient did not omit any targets, nor did she show neglect on
line bisection, copying, reading and writing.
Tonal audiometry showed a moderate left-ear loss for the
frequencies above 4000 Hz, and an ear asymmetry of 12.5 dB
in favour of the right ear, as averaged over all frequencies.
The Weber test was symmetrical. Non-verbal sound recognition was assessed with a test of identification of 50 environmental sounds with multiple choice device (see [14–16]).
NM totalised 48 good responses out of 50, a score which is
0.5 standard deviations above the mean of normal subjects.
There was a complete left-ear extinction on dichotic listening with words (right ear: 30/30, left ear 0/30) as well as
environmental sounds (right ear: 24/24, left ear 0/24).

3. Methods
We presented two types of auditory spatial tasks. The
first type of tasks assessed the ability to localise stationary or moving sound targets. In the second type of tasks,
we assessed the ability to use spatial cues to segregate simultaneous sound sources. In this second group of tasks,
no explicit processing of the spatial location of sounds
was required. In all auditory tests of this study, spatial
lateralisations were simulated by means of interaural time
differences.
3.1. Sound localisation
Two overt localisation tasks were proposed, one involving
stationary, and one involving moving sounds.
3.1.1. Localisation of stationary sounds
This test has already been described elsewhere [5,15].
It consisted of sixty 2 s broadband “bumblebee” sounds,
shaped with 100 ms rising and falling times, and presented through earphones at the intensity level judged
most comfortable by the patient. Five different azimuthal
(intra-cranial) positions were simulated by varying interaural time difference. One central (no ITD) and four lateral
positions, two in each hemi-space, were created. For the
lateral positions, the ITD was 300 ␮s or 1 ms, with either
the left or the right ear leading. Subjects were instructed
to indicate the perceived position on their head with their
ipsilesional hand (same procedure as [3,6]). A graduated
half-circle fixed on the headphones was used to determine
the angular value of the position (from 0◦ at the vertex
to 90◦ at each ear). Sixty normal subjects confirmed that
they heard single images. The average angular perceived
positions were 60.1◦ (S.D. = 13.0◦ ) on the left and 62.9◦
(S.D. = 12.5◦ ) on the right for 1 ms ITD, 37.8◦ (S.D. =
13.8◦ ) on the left and 40.5◦ (S.D. = 14.2◦ ) on the right for

1256

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261

Fig. 2. Mean angular value (in degrees, minus for the left) attributed to
each stimulus position. LL and RR: extreme left and right positions, with
1 ms ITD. L and R: intermediate left and right positions, with 0.3 ms ITD.
Ce: central position, with no ITD. Filled circles: average performance of
the 60 control subjects with inter-individual variability. Crosses: NM’s
performance.

a 300 ␮s ITD and 0.09◦ (S.D. = 4.5◦ ) on the left for 0 ITD
(Fig. 2). As a measure of overall performance, the relative
position attributed to two consecutive stimuli was compared;
a response was counted as correct when a stimulus was
correctly placed to the left or the right of the previous stimulus in correspondence with the difference in ITD or within
±10◦ of the previous location for identical ITD. The average
relative score obtained by the control population was 57.15
(S.D. = 1.79).
3.1.2. Auditory motion
An illusion of sound motion in the horizontal plane was
created by progressive changing of interaural differences.
Six different motions were simulated, each presented 10
times: left to right and reverse; left to midsagittal plane
and reverse; and right to midsagittal plane and reverse.
Motion from one extreme lateral position was created by
interaural time difference of 1 ms at the onset and progressive changes by 100 ␮s for each 50 ms of stimulus duration
until both channels were in phase (for motion stopping at
midsagittal plane) or until the reverse balance was reached
(for motion crossing the midsagittal plane). Subjects were
instructed to indicate the perceived motion on their head
with the ipsilesional hand. Responses were classified as
correct or incorrect, a margin of ±20◦ being accepted
for the midline departures/arrivals. The average score obtained by the control population (N = 60) was 52.4/60
(S.D. = 8.93).

3.2.1. Spatial-release-from-masking
The masking phenomenon refers to the fact that a soft
noise might ‘disappear’ if a louder noise of similar frequency range is presented simultaneously. Several studies
have shown that speech (or other sound sources) becomes
more intelligible with increasing spatial separation between
the speech source and the masking noise [26]. This release
of the masking effect has even been demonstrated when the
spatial removal of the masking noise was simulated by ITD
(and thus, each ear continued to receive the same frequencies at a same intensity level [11]). Our test counted 66
items (plus 10 which were not included in the analysis). On
each item, a masking noise, consisting of a 2.5 s broadband
helicopter sound (20–5500 Hz, with dominant pitch around
700 Hz), was presented at one of 11 possible spatial positions simulated with ITD (400, 320, 240, 160, 80 ␮s favouring either the left or right ear, and 0 ITD). The to-be-detected
target was a 800 ms noise of a tawny owl (20–5000 Hz, essentially centred between 350 and 900 Hz) appearing always
at the centre of the head (no ITD). The pressure level of the
helicopter sound, measured at ear entrance, was 79 dB SPL,
whereas the owl sound was 44 dB SPL (50 times lower). The
target was present in 44 out of 66 trials and began 1 s after
the masking noise. In order to avoid expectation of the target
at a constant interval after the onset of the helicopter sound,
10 other trials were added to the test but not included in the
result analysis; in 5 of them, the target began 500 ms after the
masking sound, and in 5 others, 1500 ms after. Subjects were
instructed to tell whether the target was present or not. The
average number of target detections for the 60 control subjects (mean age: 41.8 years, S.D.: 15.9 years) as a function
of the masking sound lateralisation is illustrated in Fig. 3.

3.2. Sound object segregation
Three tasks were presented. All were characterised by
the fact that the spatial dimension was only indirectly assessed, and involved the simultaneous processing of two
sound sources located in different places by means of ITD.

Fig. 3. Mean number of target detections (maximum four) as a function
of the laterality of the simultaneous masking sound. The target was
always presented at the central position (0 ITD). Filled circles: average
performance of the 60 control subjects with inter-individual variability.
Crosses: NM’s performance.

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261

3.2.2. ITD diotic listening, verbal version
This test has already been described elsewhere [5]. In
contrast to the dichotic listening task, it avoids any difference
of content and intensity between the two ears. Thirty pairs
of words were presented. Both ears received both items of
each pair at a same intensity level, but one of them was
lateralised in the left hemi-space and the other one in the
right hemi-space. The spatial lateralisations were simulated
by interaural time differences of 1 ms. The instruction was
to report both items. We have assessed the performance as
number of correctly repeated words on the right or left side,
and by the lateralisation index ([(right side − left side)/(right
side + left side)] × 100). The right-side average score for
the whole normal population (N = 60, mean age: 42.5 years,
S.D.: 14.3 years) was 26.15 out of 30 (S.D. = 4.632) and the
left-side average score was 24.867 out of 30 (S.D. = 5.02).
Paired t-tests between right- and left-side scores revealed a
statistically significant right side advantage (P ≤ 0.0001).
The average lateralisation index was 3.521 (S.D. = 5.96).
3.2.3. ITD diotic task, non-verbal version
Twenty-four pairs of 3 s environmental sounds were
presented following the same procedure of crossed ITD
lateralisation. The test was split into two parts of 12 pairs
each (each sound was lateralised once to the left, once to
the right associated to another sound). Subjects were first
trained to recognise individually each environmental sound,
before initiating the diotic part. Responses were given by
pointing the two pictures corresponding to the pair, among
12 vertically aligned pictures. The right-side average score
for the whole normal population (N = 60) was 22.5 out
of 24 (S.D. = 1.58) and the left-side average score was
22.8 out of 24 (S.D. = 1.68). Paired t-tests between rightand left-side scores revealed a non-significant left-side advantage (P = 0.095). The average lateralisation index was
−0.33 (S.D. = 3.21).
4. Results
4.1. Sound localisation
4.1.1. Localisation of stationary sounds
NM was completely unable to do the task. She reported
hearing all the stimuli at the same position, in the centre
of the head, and accordingly, pointed repeatedly to the 0◦
position (Fig. 2). Additional investigation, not detailed here,
showed that processing of IID cues was severely deficient
as well.
4.1.2. Sound motion perception
Complete break down was observed at the auditory motion task. NM fell no sense of motion, and reported that all
the stimuli were identical, at a same stationary position at
the centre of the head. Asked to guess, she was unable to
tell the direction of motion. Additional investigation, not de-

1257

tailed here, showed identical results in an IID version of the
test.
4.2. Sound object segregation
4.2.1. Spatial release from masking
The test comprised 44 trials with the target and 22 without
it. NM never reported false positives. She never perceived
the target when the masking sound was at the same central
position (0 ITD), or near the centre (80 ␮s ITD favouring
either the left or right ear, 160 ␮s ITD favouring the right
ear). By contrast, she perceived the targets on two out of four
presentations when the masking noise was lateralised with a
400 or 320 ␮s ITDs favouring the right or left ear, and 240 ␮s
ITD favouring the right ear, which reveals an effect of release
from masking. Intermediate performance was obtained for
the 160 and 240 ␮s ITDs favouring the left ear (Fig. 3).
Once the release from masking test was achieved, we
asked the patient to hear the stimuli a second time, but now
to indicate the spatial origin of the helicopter sound on the
head (explicit localisation task). NM was unable to do the
task, and answered that she heard it always at the same place
at the centre of the head, superimposed to the owl-target. She
could not even guess the hemi-space of origin (right versus
left) for the ±400, ±320, ±240, ±160, ±80 ␮s ITD stimuli.
4.2.2. Verbal ITD diotic test
NM’s performance on the ITD diotic tasks demonstrated
that she was able to make use of spatial cues. First, she segregated well the two words and fusion occurred only in 1
of the 30 pairs presented, which is better than the average
number of fusion in the control population (2.55). In the
ITD diotic task both words are presented to both ears, which
may lead to fusion if the words are not properly segregated.
The most powerful segregation cue in this task is the simulation of different spatial origins for each word by using
different ITD. NM’s performance indicates that she could
segregate the dual message using the ITD information. Second, her performance at that task was highly asymmetric:
she reported 17 out of 30 words from the right side, and
none from the left side. This gave an asymmetry index of
100, 16.2 S.D. below the average asymmetry index of normal subjects (Table 1 and Fig. 4). Performance was normal
Table 1
NM’s percentage of reports of right and left stimuli in the verbal and
non-verbal ITD diotic tasks

Left side
Right side

Verbal ITD diotic (%)

Non-verbal ITD
diotic (%)

Repeat left
+ right

Repeat left
only

Repeat left
+ right

0
56.7

6.7
50 (alloacusis)

12.5
75

Instructions were to report both stimuli in the ‘repeat left + right’ condition, and to repeat left-sided and ignore right-sided stimuli in the ‘repeat
left only’ condition.

1258

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261

Fig. 4. Lateralization index for the verbal and non-verbal ITD diotic tasks: 100 × (Rtot − Ltot )/(Rtot + Ltot ). The maximal asymmetry corresponds to 100.
The mean value from the control population (CTRL) and NM’s results are represented. The dashed lines indicate the limit of normal performance set at
2 S.D. outside the mean lateralisation index of the control subjects.

in the unilateral condition (left side as well as right side).
After the verbal test, we asked her to do it a second time, but
to report only the left-sided words and ignore the right-sided
stimuli. She found the instruction strange, as she perceived
nothing from the left side, but asked to guess, reported 15
words from the right-side and only 2 originating actually
from the left side (Table 1).
4.2.3. Non-verbal ITD diotic test
NM’s performance was significantly asymmetric: 18 out
of 24 stimuli from the right hemi-space were recognised,
against 3 out of 24 from the left hemi-space. The asymmetry index of 71.4 contrasted significantly with the mean
asymmetry index of −0.33 obtained in the control population (Table 1 and Fig. 4).
After the passation of each ITD diotic task, we asked NM
to tell whether if she heard one or two words/sounds, and to
locate them to the right or left. She reported hearing sometimes one sometimes two items; forced to chose between
right and left, she located all the stimuli (one or two items
of the pair) to the right.

5. Discussion
NM had a very severe impairment of auditory localisation, experienced in everyday life and objectivised with
our tasks assessing active auditory localisation by means
of ITD cues. Despite a complete inability to interpret
ITDs for a localisation purpose, this patient proved to be

able to use ITD cues to segregate simultaneous sound
sources.
5.1. Impairment in auditory localisation
NM was unable to derive a precise spatial information
from binaural cues: the ability to localise fused images lateralised by ITD was completely abolished. Even the most
lateral stimuli were perceived at the centre. Auditory motion
(simulated by progressive change in interaural time) was not
differentiated from static stimuli. These results cannot be
attributed to pointing impairment or perseverations, as NM
confirmed verbally that she heard all stimuli at the same
place. We have reported here only the localisation performance at tasks simulating the spatial dimension with ITD
to allow direct comparison with the segregation tasks using
this same ITD cue. In everyday life, however, auditory localisation depends on ITD cues as well as IID (interaural
intensity differences) and spectral monaural cues [7]. Other
investigations, not detailed here, ascertained that localisation and motion perception were also disrupted using IID
cues. We did not test free field localisation but NM complained having difficulties localising sound sources in everyday life. We are thus confident that NM was impaired
in auditory localisation and not simply in spatial processing
of ITD cues. Sound localisation impairments have been described after brain lesions, but in all reports of left-damaged
patients and most reports of right-damaged patients, deficits
consisted of imprecision of localisation responses rather than
complete spatial deafness as shown by our patient. Cases

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261

of spatial deafness, revealed by a testing paradigm similar to ours, have been described after brainstem lesions at
the level of trapezoid body [1,20,33]. With another testing
paradigm (discrimination of lateralisation thresholds), some
right-damaged patients were found unable to perceive lateral
shifts simulated by ITD [40] (four patients), or had largely
abnormal discrimination thresholds [23] (one patient). These
patients might have a deficit similar to NM. Their lesions involved the right parietal lobe and auditory pathway [40], or
the right inferior parietal, superior temporal and insular cortices [23]. NM’s cerebral damage also involved the inferior
parietal lobule, insula and superior temporal gyrus (including Heschl’s gyrus), but unlike pre-cited cases, it extended to
the inferior and middle frontal gyri. This additional site may
explain the extreme deficit of sound localisation found in our
patient. Functional studies on sound localisation revealed a
bilateral network of activations in the parietal, temporal and
frontal areas [10,22,24,28,42]. It is possible that the complete absence of sound location perception in NM was due to
the disruption of the whole network in the right hemisphere.
The frontal part of the network may underlie the component
of conscious awareness of sounds’ spatial position. In line
with this proposition, bilateral increase of cerebral blood
flow in the lateral frontal cortex was recently found to be
associated with conscious awareness of sounds [19].
5.2. Preserved sound segregation on the basis
of spatial cues
We present two lines of evidence that NM used ITD spatial
cues for sound object segregation: (i) she had a positive
spatial-release-from-masking effect; and (ii) she displayed a
spatial bias in the verbal and non-verbal ITD diotic tasks.
To our knowledge, the spatial-release-from-masking
paradigm has never been applied to brain-damaged patients.
NM’s ability to detect a central sound target proved to depend on the spatial distance between this target and a masking sound, the position of which was simulated by ITD. This
is a demonstration that the ITD cues could be used by the
patient for sound segregation purpose. The effect occurred
even though she reported perceiving no lateralisation of
the masking sound, which suggests that spatial segregation
does not rely on a conscious analysis of the spatial location of sounds. The release effect was of similar amplitude
than in the control population, but the global number of
target detections was reduced. This raising of the detection
threshold may be explained by NM’s slight hearing loss.
Cherry [13] and Broadbent [9] have used the dichotic listening task to study the ‘cocktail party effect’—i.e. the use
of auditory spatial information to segregate simultaneous
sounds—in normal subjects. However, in the classical dichotic listening paradigm, the spatial segregation is coupled
with a sensory segregation (each ear receives a different information). In our task, the two opposed spatial lateralisations are provided by the sole spatial parameter—ITD (see
[5]). NM fused only one pair of words out of 30, which is

1259

better than the average 2.55 fusions found in the 60 control
subjects in this task. It suggests that she could use the ITD
cue to differentiate the two simultaneous messages. Moreover, she displayed a marked asymmetry in the report of ITD
diotic pairs of words and environmental sounds, favouring
the right-sided stimuli. This asymmetry is coherent with the
left-ear extinction observed on the dichotic listening tasks,
and denotes auditory spatial neglect [5]. The only difference between the two simultaneous items was the ITD. The
present results not only show that the patient could use ITD
to segregate between two sound objects (already shown with
the release-from-masking effect), but also demonstrate that
the spatial value attributed to the ITD cues was analysed, at
least in terms of right versus left: NM neglected consistently
the items from the left hemi-space, even though she did not
experience the sensation of spatial lateralisation. It is of interest that left-sided stimuli were neglected in the ITD diotic tasks, whereas the release from masking effect occurred
independently of the side of lateralisation of the masking
sound. Alain and Arnott [2] have argued that the ability to
focus attention selectively on a particular sound source depends on a preliminary analysis that partitions the auditory
input into distinct perceptual objects. Previous electrophysiological studies on sound object segregation have shown
that these mechanisms occur at a pre-attentive level [38,44].
Accordingly, we propose that NM had normal pre-attentive
mechanisms of sound object segregation (even if they involved left sided stimuli). Left hemi-neglect became apparent when she had to attend to lateralised objects once they
were formed. Following this explanation, auditory neglect is
a high-level process which operates on preliminary spatial
processing.
5.3. Two different uses of spatial information
NM’s contradictory spatial competencies could be explained in terms of dissociation between conscious and
unconscious processing of auditory spatial cues. Indeed, our
tasks of auditory localisation imply overt spatial processing, whereas the release from masking and ITD diotic tasks
assess the spatial dimension only indirectly. In the visual
modality, some patients with cortical blindness retain the
ability to detect and localise targets of which they have no
awareness [43]. Cowey and Stoerig [17] have hypothesised
the existence of an analogue to this blind sight phenomenon
in the auditory modality, which they have called ‘deaf
hearing’. In line with this proposition, some authors have
emphasised the paradoxical ability of some patients with
cortical deafness to react to sounds they deny hearing [41].
Recently, Garde and Cowey [21] have described a patient
with cerebral deafness, who had retained some ability to
respond reflexively to sounds; she was trained to become
aware of her appropriate orienting head movements, and
could subsequently detect and localise sounds successfully
in a forced-choice paradigm, although remaining agnosic
to their meaning. The links between consciousness and

1260

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261

auditory perception are further documented in the study of
another cortically deaf patient who was found to be conscious of the onset and offset of sounds only when selectively
attending to the auditory modality. H2 15 O-PET showed
that conscious awareness of sounds was associated with increased cerebral blood flow bilaterally in the lateral frontal,
middle temporal cortices, and cerebellum [19]. Our patient
did not suffer from cortical deafness. She was perfectly
aware of verbal and non-verbal sounds and could identify
them normally. She displayed ‘deaf spatial hearing’ rather
than ‘deaf hearing’. The dichotomy between conscious
and unconscious spatial processing may account partly for
NM’s results. It should be noticed, however, that when
asked to guess the origin of one sound, the patient did not
improve her performance, which contrasts with Garde and
Cowey’s patients, as well as some patients with blindsight.
Another way to explain the pattern of auditory spatial
processing displayed by our patient is to consider the differential cognitive context associated with the impaired versus
preserved spatial abilities. NM’s deficit affects the ability to
give sense to a spatial cue, tell if it is stationary or moving,
attribute spatial co-ordinates to it and select an action accordingly (i.e. point to the perceived position in our localisation task, or cross the road in everyday life). The preserved
spatial competencies of NM (spatial-release-from-masking
effect, spatial bias at ITD diotic tasks; absence of complaints
of difficulties following conversations with many people)
concern situations where more than one sound object are
simultaneously presented and the sorting is done on the
basis of spatial analysis. Binaural spatial cues (here ITD) in
sound localisation and sound segregation serve thus different goals, one being spatial orientation and action, the other
being facilitation of perception. There is now increasing
evidence that the auditory system, like the visual system,
is organised in at least two main cortical networks, one,
dorsal, dedicated to the spatial ‘where’ aspects, another,
ventral, to the content (‘what’) of sounds (clinical data
[15,16]; fMRI data [28]; evoked potentials [4]). Lesions responsible for the localisation deficit in NM match perfectly
the temporo-fronto-parietal network associated with active
auditory spatial processing in functional data [28]. The dissociation observed in our patient demonstrates that sound
segregation does not rely on the ability to localise sounds
(completely abolished in NM). Sound source segregation
may rely on different mechanisms being part of the ‘what’
system (see also [27] for a similar proposal). There is convincing evidence that NM’s ‘what’ system is intact: she can
recognise verbal and environmental sounds perfectly, which
is in agreement with the sparing of the ventral part of her
temporal lobes. In line with this interpretation, Efron et al.
[18] have proposed the existence of temporal enhancement
mechanisms (one in each temporal lobe) which function
would be to facilitate perception of sound sources located
in the opposite side of space when other sounds are present
throughout the auditory field. They found that this ‘cocktail
party’ competency was impaired in the hemi-space con-

tralateral to a temporal lobectomy. Another evidence that auditory streaming depends on high-level mechanisms and can
be disrupted by brain lesions comes from Carlyon et al. [12].
NM’s preserved spatial segregation competencies may
also reflect residual spatial processing by the brainstem, instead involvement of a mechanism qualitatively different
from auditory localisation. The finding of a dissociation inverse to that of NM (impaired spatial segregation with normal sound localisation) could reject this hypothesis. Patients
with traumatic brain injury, who often complain of difficulty
following conversations involving many persons, may fulfil
this condition. Further research is, however, needed to clarify this point.

Acknowledgements
We are very grateful to Dr. C. Bindschaedler for data on
the neuropsychological evaluation of NM, to D. Cereghetti,
V. Mathey Croquelois and C. Morandi for collecting control data for the ‘spatial-release-from-masking’ test, and to
Prof. M. Rossi from the Electromagnetics and Acoustics
Laboratory of the Swiss Federal Institute of Technology
for analysis of sound pressure level for the ‘spatial-releasefrom-masking’ test. This work was supported by the Swiss
National Science Foundation grant 3100-064085.00/1 and
Biomed grant from UNIL-CHUV-EPFL-UNIGE-HCUG.
References
[1] Aharonson V, Furst M, Levine RA, Chaigrecht M, Korczyn AD.
Lateralization and binaural discrimination of patients with pontine
lesions. Journal of the Acoustical Society of America 1998;103:
2624–33.
[2] Alain C, Arnott SR. Selectively attending to auditory objects.
Frontiers in Bioscience 2000;5:D202–12.
[3] Altman J, Balonov L, Delgin V. Effects of unilateral disorder
of the brain hemispheric function in man on directional hearing.
Neuropsychologia 1979;17:295–301.
[4] Anourova I, Nikouline VV, Ilmoniemi RJ, Hotta J, Aronen HJ,
Carlson S. Evidence for dissociation of spatial and nonspatial auditory
information processing. NeuroImage 2001;14:1268–77.
[5] Bellmann Thiran A, Meuli R, Clarke S. Two types of auditory
neglect. Brain 2001;124:676–87.
[6] Bisiach E, Cornacchia L, Sterzi R, Vallar G. Disorders of perceived
auditory lateralization after lesions of the right hemisphere. Brain
1984;107:37–52.
[7] Blauert J. Spatial hearing. The psychophysics of human sound
localization. Cambridge (MA): MIT Press; 1997.
[8] Bregman AS. Auditory scene analysis. The perceptual organization
of Sound. Cambridge (MA): The MIT Press; 1990.
[9] Broadbent DE. The role of auditory localization in attention and
memory span. Journal of Experimental Psychology 1954;47:191–6.
[10] Bushara KO, Weeks RA, Ishii K, Catalan M-J, Tian B, Rauschecker
JP, et al. Modality-specific frontal and parietal areas for auditory
and visual spatial localization in humans. Nature Neuroscience
1999;2:759–66.
[11] Carhart R, Tillman TW, Johnson KR. Release of masking for speech
through interaural time delay. Journal of the Acoustical Society of
America 1967;42:124–37.

A.Bellmann Thiran, S. Clarke / Neuropsychologia 41 (2003) 1254–1261
[12] Carlyon RP, Cusack R, Foxton JM, Robertson IH. Effects of attention
and unilateral neglect on auditory stream segregation. Journal
of Experimental Psychology: Human Perception and Performance
2001;27:115–27.
[13] Cherry EC. Some experiments on the recognition of speech with
one and two ears. Journal of the Acoustical Society of America
1953;25:975–9.
[14] Clarke S, Bellmann Thiran A, de Ribaupierre F, Assal G. Non-verbal
auditory recognition in normal subjects and brain-damaged
patients: evidence for parallel processing. Neuropsychologia 1996;34:
587–603.
[15] Clarke S, Bellmann Thiran A, Meuli RA, Assal G, Steck AJ. Auditory
agnosia and auditory spatial deficits following left hemispheric
lesions: evidence for distinct processing pathways. Neuropsychologia
2000;38:797–807.
[16] Clarke S, Bellmann Thiran A, Maeder P, Adriani M, Vernet O,
Regli L, et al. What versus where in the auditory domain: distinct
processing pathways in human auditory cortex. Experimental Brain
Research 2002;147:8–15.
[17] Cowey A, Stoerig P. Reflections in blindsight. In: Milner D, Rugg M,
editors. The neuropsychology of consciousness. New York: Academic
Press; 1991. p. 11–37.
[18] Efron R, Crandall PH, Koss B, Divenyi PIL, Yund EW. Central
auditory processing. III. The ‘cocktail party’ effect and anterior
temporal lobectomy. Brain and Language 1983;19:254–63.
[19] Engelien A, Huber W, Silbersweig D, Stern E, Frith CD, Döring W, et
al. The neural correlates of ‘deaf-hearing’ in man. Conscious sensory
awareness enabled by attentional modulation. Brain 2000;123:
532–45.
[20] Furst M, Levine RA, Korczyn AD, Fullerton BC, Tadmor R, Algom
D. Brainstem lesions and click lateralization in patients with multiple
sclerosis. Hearing Research 1995;82:109–24.
[21] Garde MM, Cowey A. ‘Deaf hearing’: unacknowledged detection
of auditory stimuli in a patient with cerebral deafness. Cortex
2000;36:71–80.
[22] Griffiths TD, Green GGR, Rees A, Rees G. Human brain areas
involved in the analysis of auditory movement. Human Brain
Mapping 2000;9:72–80.
[23] Griffiths TD, Rees A, Witton C, Shakir RA, Henning GB, Green
GGR. Evidence for a sound movement area in the human cerebral
cortex. Nature 1996;383:425–7.
[24] Griffiths TD, Rees G, Rees A, Green GGR, Witton C, Rowe D,
et al. Right parietal cortex is involved in the perception of sound
movement in humans. Nature Neuroscience 1998;1:74–9.
[25] Haeske-Dewick H, Canavan GN, Hömberg V. Sound localization in
egocentric space following hemispheric lesions. Neuropsychologia
1996;34:937–42.
[26] Hirsh IJ. The relation between localization and intelligibility. Journal
of the Acoustical Society of America 1950;22:196–200.
[27] Kubovy M, Van Valkenburg D. Auditory and visual objects.
Cognition 2001;80:97–126.
[28] Maeder P, Meuli R, Adriani M, Bellmann Thiran A, Fornari E,
Thiran JPH, et al. Distinct pathways involved in sound recognition
and localization: A human fMRI study. NeuroImage 2001;14:802–16.

1261

[29] Makous JC, Middlebrooks JC. Two-dimensional sound localization
by human listeners. Journal of the Acoustical Society of America
1990;87:2188–200.
[30] Mills AW. On the minimum audible angle. Journal of the Acoustical
Society of America 1958;30:237–47.
[31] Pinek B, Duhamel J-R, Cavé Ch, Brouchon M. Audio-spatial deficits
in human: differential effects associated with left versus right
hemisphere parietal damage. Cortex 1989;25:175–86.
[32] Poirier P, Lassonde M, Villemure J-G, Geoffroy G, Lepore F. Sound
localization in hemispherectomized patients. Neuropsychologia
1994;32:542–53.
[33] Pratt H, Polyakov A, Aharonson V, Korczyn AD, Tadmor R,
Fullerton BC, et al. Effects of localized pontine lesions on
auditory brain-stem evoked potentials and binaural processing
in humans. Electroencephalography and clinical Neurophysiology
1998;108:511–20.
[34] Rousseaux M, Beis JM, Pradat-Diehl P, Martin Y, Bartolomeo P,
Bernati T, et al. Présentation d’une batterie de dépistage de la
négligence spatiale. Normes et effets de l’âge, du niveau d’éducation,
du sexe, de la main et de la latéralité. Revue Neurologique
2001;157:1385–400.
[35] Ruff RM, Hersch NA, Pribram KH. Auditory spatial deficits in
the personal and extra-personal frames of reference due to cortical
lesions. Neuropsychologia 1981;19:435–43.
[36] Sanchez-Longo LP, Forster FM. Clinical significance of impairment
of sound localization. Neurology 1958;8:119–25.
[37] Sterzi R, Piacentini S, Polimeni M, Liverani F, Bisiach E. Perceptual
and premotor components of unilateral auditory neglect. Journal of
the International Neuropsychological Society 1996;2:419–25.
[38] Sussman E, Ritter W, Vaughan HG. An investigation of the
auditory streaming effect using event-related brain potentials.
Psychophysiology 1999;36:22–34.
[39] Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human
brain. Stuttgart: Georg Thieme; 1988.
[40] Tanaka H, Hachisuka K, Ogata H. Sound lateralization in patients
with left or right cerebral hemispheric lesions: relation with
unilateral visuospatial neglect. Journal of Neurology, Neurosurgery,
and Psychiatry 1999;67:481–6.
[41] Tanaka Y, Kamo T, Yoshida M, Yamadori A. ‘So-called’ cortical
deafness. Clinical, neurological and radiological observations. Brain
1991;114:2385–401.
[42] Weeks RA, Aziz-Sultan A, Bushara KO, Tian B, Wessinger CM,
Dang N, et al. A PET study of human auditory spatial processing.
Neuroscience Letters 1999;262:155–8.
[43] Weiskrantz L. Blindsight: a case study and implications. Oxford:
Oxford University Press; 1998.
[44] Yabe H, Winkler I, Czigler I, Koyama S, Ryusuke K, Sutoh T, et
al. Organizing sound sequences in the human brain: the interplay
of auditory streaming and temporal integration. Brain Research
2001;897:222–7.
[45] Yost WA. Auditory image perception and analysis: the basis for
hearing. Hearing Research 1991;56:8–18.
[46] Zatorre RJ, Penhune VB. Spatial localization after excision of human
auditory cortex. The Journal of Neuroscience 2001;21:6321–8.