Hemispheric specialization for dynamic face processing: A comparative fMRI study between human and non-human primates

Most face processing studies in humans show stronger activation in the right compared to the left hemisphere. It is unclear whether this property is unique for humans or also exists in non-human primates. Furthermore, evidence in humans is largely based on studies with static stimuli focusing on the fusiform face area (FFA) whereas the pattern of lateralization for dynamic faces is less clear. We aimed to address both issues by studying lateralization for dynamic facial expressions in monkeys and humans. We conducted an identical event-related fMRI experiment in twenty-one right handed humans and three macaque monkeys. During the fMRI experiment, we presented, to both species, human and monkey dynamic facial expressions (chewing and fear) as well as their spatially and temporally scrambled versions (same stimuli as in Zhu et al. 2012) while a passive fixation task was performed. To study lateralization in unbiased regions-of-interest (ROI), we used half of the runs to create these ROIs and calculated the degree of lateralization using a bootstrapping method within the LI-toolbox in SPM8 using the other half of the data (Wilke et al.2006). This analysis yielded a weighted lateralization index (LI). Additionally, we conducted a localizer experiment to examine lateralization in face-selective areas of both species. Finally we added a third language-related expression, i.e. lip reading, in a separate fMRI experiment in 10 human volunteers. Within face-responsive regions, we found consistent lateralization for dynamic faces (chewing and fear) versus scrambled versions in the right human posterior superior temporal sulcus (pSTS), but not in ventral occipito-temporal cortex. Similar results were obtained in face-selective areas, showing right-hemispheric lateralization for dynamic faces in pSTS but not in FFA. In monkeys, on the other hand, we did not find a consistent pattern of lateralization for dynamic facial expressions in occipito-temporal cortex. Finally, LI’s based on the contrast between different types of dynamic facial expressions (relative to scrambles) within human pSTS revealed left-hemispheric lateralization for language-related expressions (lip reading) compared to neutral and emotional ones. To conclude, we found consistent laterality effects in human posterior STS but no comparable effects in monkeys. Taken together, it is tempting to speculate that lateralization for dynamic face processing in humans may be driven by left-hemispheric language specialization which may not have been present yet in the common ancestor of human and macaque monkeys.status: publishe