Bayesian priors are encoded independently from likelihoods in human multisensory perception.

By Ulrik R Beierholm, Steven R. Quartz, Ladan Shams. 2009

108 citations recorded by the source corpus.

It has been shown that human combination of crossmodal information is highly consistent with an optimal Bayesian model performing causal inference. These findings have shed light on the computational principles governing crossmodal integration/segregation. Intuitively, in a Bayesian framework priors represent a priori information about the environment, i.e., information available prior to encountering the given stimuli, and are thus not dependent on the current stimuli. While this interpretation is considered as a defining characteristic of Bayesian computation by many, the Bayes rule per se does not require that priors remain constant despite significant changes in the stimulus, and therefore, the demonstration of Bayes-optimality of a task does not imply the invariance of priors to varying likelihoods. This issue has not been addressed before, but here we empirically investigated the independence of the priors from the likelihoods by strongly manipulating the presumed likelihoods (by using two drastically different sets of stimuli) and examining whether the estimated priors change or remain the same. The results suggest that the estimated prior probabilities are indeed independent of the immediate input and hence, likelihood.

Subfields: Experimental and Cognitive Psychology, Nutrition and Dietetics, Sensory Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Experimental and Cognitive Psychology

In Nutrition and Dietetics

In Sensory Systems

Citations in the corpus

Assessing the Role of the ‘Unity Assumption’ on Multisensory Integration: A Review (2017, background): “So, for example, when the sensory representations change, such as the reliability of the stimulus being reduced by decreasing the luminance contrast of the visual stimulus, the prior remains constant (Beierholm et al., 2009).”
The COGs (context, object, and goals) in multisensory processing (2016, background): “…both within and across the senses (reviewed in Frost et al. 2015), with effects transpiring across multiple temporal scales (e.g. Baier et al. 2006; Beierholm et al. 2009; Chandrasekaran et al. 2009; Barakat et al. 2013; Barenholtz et al. 2014; Altieri et al. 2015; Sarmiento et al. 2012, 2016).”
Contextual factors multiplex to control multisensory processes (2016, background): “The importance of statistical learning has likewise been demonstrated for multisensory processing [Altieri et al., 2015; Baier et al., 2006; Barakat et al., 2013; Barenholtz et al., 2014; Beierholm et al., 2009; Chandrasekaran et al., 2009].”
Sensory reliability shapes perceptual inference via two mechanisms. (2015, methodology): “Moving beyond previous modeling efforts (Beierholm et al., 2009; Koerding et al., 2007; Sato et al., 2007; Wozny et al., 2010), we validated Bayesian CI models on two psychophysics data sets that included two (see Supplementary Experiment 1) or four levels of visual reliabilities.”