Integrating auditory and visual inputs plays an important role in language processing and social perception. We presently investigated mechanisms of this audio-visual (A-V) integration by analyzing temporal characteristics of multisensory neurons in the human brain. Specifically, an inter-trial coherence (ITC, neural index for phase resetting) of scalp electroencephalography (EEG) were measured while participants made a temporal-order judgment between a beep (A) and a flash (V). The posterior temporal region showed higher ITC (phase resetting) to bimodal (A+V) than unimodal (A or V) stimuli, showing an involvement of multisensory neurons selectively reacting to the bimodal inputs. They responded more vigorously as a beep-flash SOA (stimulus-onset asynchrony) came closer to 0 ms. Finally, this increase in ITC was most clearly seen in beta band (13-30 Hz). These results indicate that multisensory neurons with a basic rhythm of 13- 30 Hz play a critical role in human A-V integration.