Both hemispheres are activated in auditory spatial tasks (Maeder et al., 2000), and unilateral right or left lesions can lead to auditory
localisation impairments. Moreover, there is evidence for a processing specificity within the auditory modality: each hemisphere
contains a dorsal cortical network dedicated to auditory spatial information, and a ventral cortical network for auditory recognition
(Clarke et al., 2OOOa,b). We report here on two hemispherectomized patients. Their performance in auditory localisation, matched
to the cerebral activation generated by auditory tasks (fMRI), are relevant to our understanding of the compensatory strategies
elaborated to carry out this normally bi-hemispheric function.
Auditory localisation was tested by simulating different azimuthal positions of sound sources by varying differences in interaural
time (ITD) or intensity (IID). Normal performance was established in 60 controls. Brain activations associated with sound
recognition or localisation was investigated in 18 normal subjects using fMRI (1.5 Tesla Siemens Vision; Maeder et al. 2000).
Three conditions were used: i) comparison of spatial stimuli simulated with ITD; ii) identification of environmental sounds; and
iii) rest. Activation patterns were analysed using SPM99, for individual control subjects and the whole group. In normal subjects,
specific regions of activation were found for recognition and localisation (Maeder et al., 2000).
MA was a 36 year old right-handed man, who developed normally until the age of 21, when he began to suffer from epilepsy
(Rasmussen encephalitis). Aged 34 years, he underwent a functional hemispherectomy. JG was a 34 year old right-handed woman,
who had the first manifestations of Rasmussen encephalitis at the age of 4. Right peri-insular hemispherectomy was performed
when she was 32. Both patients were tested 2 years later. Their neuropsychological examination was only mildly abnormal.
Both patients were deficient in auditory localisation as assessed by ITD and IID, one only mildly (MA), and the other severely
(JG). Thus, one patient had developed quite successful1 compensatory strategies while the other had not. MA's performance was
characterised by a leftwards bias; JG's performance by a marked imprecision on the whole field, with several alloacusis.
Activation patterns were investigated with fMRI. The mild impairment (MA) was associated with an activation pattern very
similar to the control data in the remaining hemisphere: the segregation between a fronto-parietal network for the spatial task, and
a temporo-frontal network for the recognition task was found. By contrast, the severe impairment (JG) was associated with an
absence of specificity for the processing of auditory spatial and non-spatial information: the same cerebral structures were used for
localisation and recognition. These two different patterns of activation might illustrate different strategies to overcome the loss of
one hemisphere. The confrontation with performance emphasises the fact that the preservation of a specificity of processing
(anatomical segregation between spatial and content information) is associated with better performance in localisation.
References
Clarke et al. (2000a). Neuropsychologia, 38: 797-807
Clarke et al. (2OOOb). European Journal of Neuroscience Suppl, 11: 129
Maeder et al. (2000). NeuroImage, 11, S52.
S861