Here we brought together these distinct lines of research by examining properties of the STS in terms of selective response to social stimuli. Normal adult volunteers participated in an ‘audiovisual localiser’ scan during which they were stimulated with auditory, visual, or audiovisual stimuli of people or objects. We proposed, given that face-selective,

voice-selective and integrative regions are found within the STS, that in addition to areas preferring both faces and voices (i.e., ‘people-selective’ regions) there could also be audiovisual regions that are more sensitive to social stimuli, as compared to information from non-social categories, such as objects. We found check details that a restricted portion of the right pSTS was characterised by a conjunction of (1) an ‘integrative’ response, i.e., stronger response to audiovisual stimuli compared to visual and compared C59 wnt to auditory stimuli and (2) ‘people-selectivity’, i.e., preference for social stimuli irrespective of the modality (voice > objects; face > objects). Furthermore, a large region further extending down the trunk of the right STS was observed to be heteromodal: that is, this region was activated

by both faces and voices, but did not necessarily show integrative properties. Forty English-speaking participants (15 males and 25 females; mean age: 25 years ± 5 years) took part in the scan. All had self-reported normal or corrected vision and hearing. The ethical committee from the University of Glasgow approved the study.

All volunteers provided informed written consent before, and received payment for, participation. 24 people (12 males and 12 females) were video-recorded producing a variety of vocal expressions, both speech and non-speech (e.g., saying the word ‘had’, humming, SPTLC1 yawning). Recordings took place in the television studio at the Learning and Teaching Centre, Glasgow University, and participants were paid at the rate of £6 per hour. The participants were filmed under standard studio lighting conditions (standard tungsten light), and sat directly facing the camera, at a distance so that the whole face was in frame. Videos were recorded with 25 frames per second (40 msec per frame) using a Panasonic DVC Pro AJD 610 camera, fitted with a Fujiform A17 × 7.8 BERM-M28 lens, and transferred and edited using Adobe Premier Elements. Within the video recording, vocalisations were recorded with 16-bit resolution at a sampling frequency of 44,100 Hz. Under the same conditions, 24 moving objects producing sound were also filmed (e.g.