Sound localization in hemispherectomized subjects: the contribution of crossed and uncrossed cortical afferents.

The aim of the present study was to evaluate how hemispherectomized subjects localize sounds in free field using residual auditory structures under monaural testing conditions. The main objective of using a monaural condition with these subjects, who lack the terminal fields of auditory projections on one side, was to evaluate how the crossed and uncrossed pathways compare, with the aim of resolving this biologically critical function. In this model, crossed and uncrossed inputs refer to auditory stimulation presented to the unobstructed ear on the contralateral and the ipsilateral side of the intact hemisphere, respectively. Three hemispherectomized subjects (Hs) and ten control subjects (Cs) were tested for their accuracy to localize broad band noise bursts (BBNBs) of fixed intensity presented on the horizontal plane. BBNBs were delivered randomly through 16 loudspeakers mounted at 10 degrees intervals on a calibrated perimeter frame located inside an anechoic chamber. Subjects had to report the apparent stimulus location by pointing to its perceived position on the perimeter. Hs were less accurate than Cs in the baseline binaural condition, confirming the finding that with a single hemisphere and/or residual (subcortical) structures they cannot analyze binaural cues to sound localization as efficiently as with two fully functional hemispheres. In the monaural condition, Hs localized poorly when they had to depend on the uncrossed input, but performed as well or even better than the Cs with the crossed input. These findings suggest that monaural spectral cues, which constitute the only residual cue to localization under the monaural testing condition, are treated more efficiently, that is, they lead to better localization performance when relayed to the cortex via crossed pathways than through uncrossed pathways.