Search Papers | Poster Sessions | All Posters
Poster B17 in Poster Session B - Thursday, August 8, 2024, 1:30 – 3:30 pm, Johnson Ice Rink
Evidence accumulation across the senses in the face of causal uncertainty
Jochem Beurskens1, Steffen Buergers2,4, Agata Wlaszczyk3, Uta Noppeney1,2; 1Radboud University, 2University of Birmingham, 3Technical University of Denmark, 4Royal HaskoningDHV
Humans need to make accurate and timely decisions based on a constant influx of noisy sensory signals from common and separate causes. Current Bayesian Causal Inference (BCI) models have shed light on how the brain arbitrates between sensory integration and segregation when uncertain about the signals’ causal structure. Yet, these static models made predictions only for response choices and ignored the dynamic nature of perceptual decision making. Using psychophysics, we show that crossmodal biases (CMB) decline for longer response times. We then develop a dynamic Bayesian causal inference (BCI) model that accumulates evidence jointly about an event’s location and the signals’ causal structure over time until a heuristic decisional threshold is reached. This dynamic BCI model accounts for the decline in crossmodal influences over time by the progressive resolution of spatial and causal uncertainty.
Keywords: Bayesian causal inference Evidence accumulation Behavioural modelling Ventriloquist effect