Fig. 3
From: Causal inference shapes crossmodal postdiction in multisensory integration
Model predictions demonstrate the BCI model’s superiority over three competing models in explaining the Invisible and Illusory AV Rabbit Illusions. The group-averaged observed data (orange in the left and blue in the right panel) and the predicted illusion rates from the fitted BCI (green), forced-fusion (FF, black), forced-segregation (FS, purple), and non-postdictive BCI (BCI-NP, pink) models are shown for the Invisible (right) and Illusory (left) AV Rabbits. Predicted illusion rates are shown for unisensory and multisensory control conditions, the illusion condition, and for asynchronous AV conditions. The BCI model shows the best overall data fit for both illusions, as demonstrated by Bayesian model comparison (Table 1). Uni Ctrl = unisensory control; Multi Ctrl = multisensory control; A-Lead = auditory lead asynchronous condition; V-Lead = visual lead asynchronous condition.
