Abstract
Voices, in addition to faces, enable person identification. In particular, voice recognition has been shown to evoke a distributed network of brain regions that includes, amongst others, the fusiform face area (FFA) and the anterior temporal lobe. In the present fMRI study, we examined fMRI brain activation in a prosopagnosic patient M.S. who, despite extensive late-onset damage to these two areas, nevertheless retains the ability to recognize people based only on their voice. Relative to speech-modulated noise, familiar and unfamiliar voices elicited enhanced hemodynamic activity in the posterior portion of the left superior temporal gyrus, as well as the left planum temporale (pT) and posterior cingulate gyrus (pCG). Region-of-interest analyses on those areas revealed that pT and pCG activity was greater for voices rated as familiar as opposed to unfamiliar. Results are consistent with the notion that an intact FFA is not essential for voice recognition. Instead, the activation seen in this face-sensitive region during voice recognition tasks in healthy subjects may reflect some form of implicit face imagery.
This work was supported by Canadian Institutes of Health Research grants to S.R.A. and M.A.G.