hide
Free keywords:
-
Abstract:
The brain is thought to generate selective and efficient representations of important sensory events
such as communicative signals, yet the various sensory systems might instantiate such selective
representations in different ways. Since the 1980s the processing of facial information by ‘face’ cells
has been repeatedly studied. Although auditory ‘voice’ regions showing a strong fMRI activity
preference for the voice of conspecific individuals have now been identified in humans and
monkeys, the fMRI signal cannot specify the encoding properties of the underlying neurons or
whether fMRI voice-preferring clusters contain ‘voice cells’. We investigated the responses of
neurons in an fMRI-identified voice cluster in awake macaque monkeys and provide the first
systematic evidence for voice cells, defined, in analogy to face cells, as neurons responding at least
two-fold stronger to conspecific voices than to heterospecific animal voices or natural/environmental
sounds. Surprisingly, whereas face clusters contain high proportions of face-preferring cells that
respond broadly to many faces, we found a considerable yet, by comparison, moderate proportion of
voice-preferring cells that exhibited a sparse-coding strategy for voices. The observed selective
representation for individual voices might stem from the different evolutionary pressures that would
have affected how the auditory system has specialized relative to the visual. In all cases, our results
highlight the interesting processing strategies used by the primate brain to encode auditory and visual components of communication signals.
