Fig. 5
Schematic of different models, using temporal estimation as an example. For all the plots in the left column, the likelihoods (dashed lines) for audio (in green) and vision (in cyan) are modeled by a Gaussian distribution centered on the current stimulus to estimate and with a width corresponding to the sensory precision of each modality. In the specific case of temporal performance, the precision is higher for auditory stimuli (in green) than visual stimuli (in cyan). The prior is represented by the Gaussian distribution based on the assumption of each model. In the audio segregation model (A), the prior is centered on the average of the auditory stimuli only. In the vision segregation model (B), the prior is centered on the average visual stimuli only. In the central tendency effect model (C), the prior is centered on the average of the stimuli in the session, i.e., averaging both auditory and visual stimuli. In the weighted central tendency effect model (D), the prior is centered on the average of the stimuli in the session, taking into account the weight given to each sensory modality, i.e., averaging both auditory and visual stimuli. The posteriors are represented on the right column by the green and cyan continuous lines, respectively, for audio and vision, as a combination of prior and likelihood.
