An FMRI study to investigate auditory attention: a model of the cocktail party phenomenon

In human life, discrimination of a target voice from other voices or sounds is indispensable, and inability for such discrimination results in sensory aphasia. To investigate the neuronal basis of the attentional system for human voices, we evaluated brain activity during listening comprehension tasks using functional magnetic resonance imaging (fMRI) at 3T. Diotic listening comprehension tasks, in which a narration was superimposed by another given by the same speaker (SV experiment) or by a different speaker (DV experiment), were presented to normal volunteers. The story indicated in the baseline task blocks, in which only one narration was presented, was intensively followed during the superimposed task blocks. In each experiment, 6 task blocks, 3 blocks for each condition, and 7 rest blocks were alternatively repeated, and the contrast of the superimposed condition to the baseline condition in each session was obtained. In the DV experiment, compared with the control condition, activation in Wernicke's area (BA22) was increased. In the SV experiment, activation in the frontal association cortex (BA6, BA9/ 46, BA32, BA13/47) was additionally increased. These results suggested that difficulty in phonological processing to discriminate human voices calls for further semantic, syntactic, and prosodic processing, as well as augmented selective attention.