Abstract

Cross-modal processing and multisensory integration (MSI) can be observed at early stages of sensory processing in the cortex. However, the neurophysiological mechanisms underlying these processes and how they vary across sensory systems remain elusive. The aim of this study was to investigate how cross-modal processing and MSI are reflected in power and phase of oscillatory neuronal activity at different temporal scales in different sensory cortices. To this goal, we recorded stereo-electroencephalographic (SEEG) responses from early visual (calcarine and pericalcarine) and auditory (Heschl’s gyrus and planum temporale) regions in patients with drug-resistant epilepsy while performing an audio-visual oddball task. To Investigate crossmodal processing and MSI in the power domain of oscillatory activity, we explored a wide range of frequency bands (theta/alpha band: 5-13Hz; beta band: 13-30 Hz; gamma band: 30-80 Hz; high-gamma band: 80-200 Hz) during the first 150 ms post-stimulus onset. Differently, to investigate crossmodal processing and MSI in the phase domain of oscillatory activity, we explored a narrow range of frequency bands (theta/alpha band: 5-13Hz; beta band: 13-30 Hz; gamma band: 30-80 Hz) during the first 300 ms post-stimulus onset. In the power domain, we showed that cross-modal processing occurs mainly in the high-gamma band (80-200Hz) in both cortices. However, we evidenced that the way MSI is expressed across modalities differs considerably: in the visual cortex, MSI relies mainly on the beta band, however it is also evident, to a lesser extent, in the gamma and high-gamma band, while the auditory cortex reveals widespread MSI in the high-gamma band and, to a lesser extent, across the gamma band and the other investigated frequency bands. In the phase domain, we showed that cross-modal processing is differently expressed across modalities: in the auditory cortex it induces an increased phase concentration index (PCI) in ongoing oscillatory activity across all the investigated frequency bands, while, in the visual cortex, it induces an increased PCI particularly evident in the theta/alpha band with few or no effect respectively in the gamma and beta band. Importantly in both cortices, the most part of the COIs showing increased PCI, were not accompanied by a concomitant increase in power. These results indicate that in both auditory and visual cortex, cross-modal processing induces a pure phase resetting of the oscillatory activity. During MSI processing we observed, in both cortices, a stronger increase in PCI, in comparison to the intramodal processing, in the theta/alpha band and in the gamma band. Our results confirm the presence of cross-modal information representations at neuronal populations level and conform to a model where the cross-modal input induces phase-locked modulation of the ongoing oscillations. Importantly, our data showed that the way MSI is expressed in power modulations differs between the investigated sensory cortices suggesting the presence of different types of neurophysiological interactions during this process. These results are discussed in the framework of the current literature.