Heterogeneous auditory–visual integration: Effects of pitch, band-width and visual eccentricity

The identification of monosynaptic connections between primary cortices in non-human primates has recently been complemented by observations of early-latency and low-level non-linear interactions in brain responses in humans as well as observations of facilitative effects of multisensory stimuli on behavior/performance in both humans and monkeys. While there is some evidence in favor of causal links between early–latency interactions within low-level cortices and behavioral facilitation, it remains unknown if such effects are subserved by direct anatomical connections between primary cortices. In non-human primates, the above monosynaptic projections from primary auditory cortex terminate within peripheral visual field representations within primary visual cortex, suggestive of there being a potential bias for the integration of eccentric visual stimuli and pure tone (vs. broad-band) sounds. To date, behavioral effects in humans (and monkeys) have been observed after presenting (para)foveal stimuli with any of a range of auditory stimuli from pure tones to noise bursts. The present study aimed to identify any heterogeneity in the integration of auditory–visual stimuli. To this end, we employed a 3 × 3 within subject design that varied the visual eccentricity of an annulus (2.5°, 5.7°, 8.9°) and auditory pitch (250, 1000, 4000 Hz) of multisensory stimuli while subjects completed a simple detection task. We also varied the auditory bandwidth (pure tone vs. pink noise) across blocks of trials that a subject completed. To ensure attention to both modalities, multisensory stimuli were equi-probable with both unisensory visual and unisensory auditory trials that themselves varied along the abovementioned dimensions. Median reaction times for each stimulus condition as well as the percentage gain/loss of each multisensory condition vs. the best constituent unisensory condition were measured. The preliminary results reveal that multisensory interactions (as measured from simple reaction times) are indeed heterogeneous across the tested dimensions and may provide a means for delimiting the anatomo-functional substrates of behaviorally-relevant early–latency neural response interactions. Interestingly, preliminary results suggest selective interactions for visual stimuli when presented with broadband stimuli but not when presented with pure tones. More precisely, centrally presented visual stimuli show the greatest index of multisensory facilitation when coupled to a high pitch tone embedded in pink noise, while visual stimuli presented at approximately 5.7° of visual angle show the greatest slowing of reaction times.