Voice representations in the brain

What is it about?

We investigated neural mechanisms that support voice recognition in a training paradigm with fMRI. The same listeners were trained on different weeks to categorize the mid-regions of voice-morph continua as an individual's voice. Stimuli implicitly defined a voice-acoustics space, and training explicitly defined a voice identity space. Cortical sensitivity to voice similarity appeared over different time-scales and at different representational stages. These findings are interpreted as effects of neural sharpening of long-term stored typical acoustic and category-internal values. The analyses also reveal anatomically separable voice representations: one in a voice-acoustics space and one in a voice-identity space. Voice-identity representations flexibly followed the trained identity shift, and listeners with a greater identity effect were more accurate at recognizing familiar voices.

Featured Image

Why is it important?

Our results are in line with the proposal that voice recognition is supported by a categorical level of processing that is anatomically separable from voice structural processing (Belin et al., 2004). Our findings also confirm that there exist dissociable neural mechanisms for short-interval versus long-interval fMRI repetition suppression (Epstein et al., 2008). More specifically, we have argued for the existence of dynamic, long-lasting ‘mean voice’ representations at both voice-acoustic and voice-identity stages of processing. In accordance with recent findings in behavioural studies of voice processing (Papcun et al., 1989; Mullennix et al., 2009, Bruckert et al., 2010) and with those in the face processing domain (Loffler et al., 2005), our demonstrations of neural ‘mean voice’ representations constitute the first neuroimaging evidence that voice representations are centered around prototypes in long-term memory.