Participants viewed short trials of audiovisual speech. For each subject, all trials were sorted by the amount of time fixating the mouth region of the talker’s face; trials with greater than the median amount of time spent fixating the mouth were classified as mouth trials and the remained were classified as eye trials. The contrast of brain activations between mouth vs. eye trials was used to define mouth and eye selective regions of the pSTS. The response of these mouth and eye ROIs to a separate experimental condition consisting of blocks of auditory-only (A), visual-only (V) and audiovisual (AV) speech were calculated for each participant and entered into a linear mixed-effects model. The linear mixed-effects model had fixed factors of ROI (mouth, eye) and stimulus (A, V, AV) with participant as a random factor and participant × stimulus as a random interaction. The first three rows show the mean response to A, V and AV conditions in mouth and eye ROIs and the contrast between mouth and eye ROIs using reduced linear mixed-effects (tested with chi square tests). The next rows show the parameter estimates for each factor in the model (the baseline condition was the response to A speech in the eye ROI). The final rows show the results of chi square tests of the models.
