Fax: +36 1 381 21 80Running title: Voice-sensitive regions in the dog and human brain
SummaryDuring the approximately 18-32 thousand years of domestication [1], dogs and humans have shared a similar social environment [2]. Dog and human vocalizations are thus familiar and relevant to both species [3], although they belong to evolutionarily distant taxa, as their lineages split approximately 90-100 million years ago [4]. In this first comparative neuroimaging study of a non-primate and a primate species, we made use of this special combination of shared environment and evolutionary distance. We presented dogs and humans with the same set of vocal and non-vocal stimuli to search for functionally analogous voice-sensitive cortical regions. We demonstrate that voice areas exist in dogs, and that they show a similar pattern to anterior temporal voice areas in humans. Our findings also reveal that sensitivity to vocal emotional valence cues engages similarly located non-primary auditory regions in dogs and humans. Although parallel evolution cannot be excluded, our findings suggest that voice areas may have a more ancient evolutionary origin than previously known.
4
Results and DiscussionAn important social function of the auditory system is to process the vocalizer's identity and emotional state. Non-primary auditory brain regions preferring conspecific vocalizations were found in both humans [5, 6] and non-human primates, suggesting that 'voice areas' evolved at least 30 million years ago [7][8][9][10]. In humans, auditory regions sensitive to vocal emotional cues have also been identified [11][12][13][14] Behavioural field research has revealed that the efficient processing of conspecificity and emotional information in vocalizations is important in both primate [18][19][20] and non-primate species [21][22][23][24]. Indeed, both the acoustic recognition of conspecifics and their emotional state is fundamental for making decisions in behaviour contexts like mate choice, territory disputes or hierarchy-related challenges [25]. Nevertheless, little is known about the underlying neural mechanisms of vocalization processing in non-primates.To reveal possible functional analogies between human and non-primate auditory brain regions, this study describes a comparative investigation of dogs and humans. We investigated (1) whether in dogs, similarly to humans, certain auditory regions ('voice areas') would respond stronger to conspecific vocalizations than to either heterospecific vocalizations or non-vocal sounds, and (2) whether dogs are similar to humans in the cortical processing of emotional cues in vocal signals.To address these questions, we used a non-invasive functional magnetic resonance imaging (fMRI) procedure with awake dogs (N=11) and humans (N=22). We built on our group's first small sample-sized attempts of awake dog fMRI [26], providing a procedure different from others' [27]. All participants were unrestrained and instructed to lay motionless in an fMRI 5 scanner for three 6-min runs (see Figure 1 and...