Skip to main content
. Author manuscript; available in PMC: 2018 May 2.
Published in final edited form as: Neural Comput. 2018 Mar 22;30(5):1209–1257. doi: 10.1162/NECO_a_01072

Figure 8.

Figure 8

Dependence of input variance on occlusion levels. Model responses across occlusion levels when (A) the variances σ1, σ2 and σ3 increase with added occlusion at the same rate (σ1 = σ2 = σ3 = 1 + 5 · c), and (B) the variances for the shape-selective V4 units σ1 and σ2 increase at the same rate as occlusion increases, while σ3 remains unchanged (σ1 = σ2 = 1 + 5 · c; σ3 = 1), (C) the variances for the shape-selective V4 units σ1 and σ2 both increase, but σ2 at a slower rate; here σ3 again remains unchanged (σ1 = 1 + 5 · c; σ2 = 1 + 2 · c; σ3 = 1), (D) the variances σ1, σ2, and σ3 decrease with added occlusion at the same rate (σ1 = σ2 = σ3 = 1 − c)