Abstract
Functional neuroimaging studies of the human superior temporal sulcus (STS) have established this brain region to be involved in the perception of biological motion (Grossman et al., 2000; Thompson et al., 2005), suggesting that this area is the human homologue of macaque STPa, a region with neurons tuned to specific body actions (Oram & Perrett, 1994). Integrated within these action-tuned neurons in monkey STS are individual cells that generalize across multiple viewpoints for the same action, forming a class of viewpoint-invariant, or ‘action-invariant’ neurons (e.g. Jellema & Perrett, 2006). In this study, we use the rapid fMR-adaptation paradigm to probe the tuning specificity of the human STS, specifically measuring the viewpoint-dependence of neural populations within this region. Methods. Subjects viewed pairs of point-light animations of human actors depicting the identical action, two different actions, or the same action mirror-reversed. Tuning for specific actions should be reflected by a reduced peak response for the repeated animations relative to the different actions (adaptation), while viewpoint invariance would be revealed by a reduced response to the left-right reversed animations compared to the different actions pairs. Results. Whole-brain general linear model analysis in individual subjects reveals clusters of action tuned responses (i.e. adaptation to repeated) and viewpoint-invariant responses (adaptation to left-right reversed pairs) on the STS, in and adjacent to the STS brain region localized with point-light animations. Conclusions. Regions of the human STS contain clusters of neurons that are viewpoint-invariant, much like those identified on monkey STS. These results are consistent with models of STS function as building abstracted representations of actions, which likely form the basis of some higher social functions such as perceived animacy and social understanding (e.g. Saxe et al., 2001; Martin & Weisberg, 2003).