Abstract
Navigating through environment requires an accurate estimation of self-motion. Such an accurate estimation becomes especially challenging when the environment itself moves and generates retinal signals which can be interpreted as environmental or self-motion. Here, based on that the synchronicity depends on the causal relation, we explore whether synchronicity of environmental and self-motion signals modulates the causal inference in heading direction estimation. In the experiment, observers reported self-motion direction after being moved to ±15° or ±45° from vertically upward for 2000ms. Self-motion stimuli were paired with a rightward or leftward visual motion that was presented through an aperture. Crucially, visual stimuli changed its velocity at one of 10 time points ranging from 1600ms before to 1600ms after the onset of self-motion and then returned to its initial velocity after 1600ms. In addition, there were two different visual motion conditions. In the acceleration condition, visual motion velocity was initially zero and at the change-point, accelerated to ±10°/s. In the deceleration condition, visual motion velocity was initially ±10°/s and at the change-point, decelerated to 0°/s. In the acceleration condition, heading bias increases as asynchrony between visual and self-motions decreases (F(9,60)=7.31, p<0.001). By contrast, in the deceleration condition heading bias decreases as asynchrony between visual and self-motions decreases (F(9,50)=2.89, p=0.008). Results suggest that when visual and self-motions were synchronized, our system assumes that visual and self-motion signals came from the same source and attributes retinal motion to self-motion. Furthermore, in an additional experiment in which visual velocity stayed constant after the change-point, we found synchronicity still matters even when the visual motion during the self-motion is the same across change-point conditions. Taken together, our findings identified synchrony between two sensory signals as a critical factor determining multisensory causal inference.