Skip to main content
. 2010 Dec 14;5(12):e15555. doi: 10.1371/journal.pone.0015555

Figure 2. Conditional dependencies in perceptual models of associative learning.

Figure 2

Left: cascade of events leading to the sensory outcomes. A Gaussian prior (with variance Inline graphic) is defined at the level of the cue-outcome association Inline graphic. Passed through a sigmoid mapping, this determines the probability of getting a house (Inline graphic) or a face (Inline graphic). Finally, this determines the visual outcome Inline graphic within the natural range of variation (Inline graphic) of house/face images. Right: Equivalent graphical model.