Visual perception of incomplete stimulus information and its modulation by auditory noise

Abstract

Bistable perception is an intriguing phenomenon whereby the interpretation of ambiguous visual signals alternates spontaneously between two possible states. Often, the alternation of states is observed to be fairly periodic with a consistent phase duration of each state. This specific temporal features of bistable perception may provide insight into the dynamics of perceptual processing for incomplete sensory information, so for this reason, it has been extensively studied. However, the exact mechanism of bistable perception is still elusive. So far, it has been unknown whether the frequency of phase alternation is simply determined by given visual stimuli or is independent of stimuli and reflects internal parameters instead. Also, it has been unclear how this frequency can be altered by other sensory modalities such as auditory noise. To address this issue, we performed a series of psychophysics experiments on bistable perception with various visual stimuli, and with and without acoustic noise. First, we measured the phase duration of bistable perception for three types of visual stimuli. We observed an identical log-normal distribution of phase duration time within each individual, which suggests that the frequency of phase alternation is fairly consistent across stimuli. Second, we statistically showed that acoustic tone can alter perception, which shows possibility that other sensory stimuli can alter temporal dynamics. Lastly, when continuous acoustic noise was introduced during the experiments, the phase duration of most subjects was systematically modulated in a way that variation across subjects was significantly reduced, indicating that visual perception under ambiguous conditions can be significantly altered by multi-modal noise. To explain these psychophysical experiment results, we built a conceptual model that could explain the bistable perception mechanism and multi-modal integration process. Our simulation suggests that our observation can be achieved from the combined contribution of external noise and the internal modulation of neural activity. Our results reveal that bistable perception may be a multimodal integrated process involving both upstream and downstream neural activity.