Fig. 4 |. Encoding of semantic features.

a. Auditory task semantic feature kernels. The shaded areas indicate the 95% confidence regions of the surrogate (chance-level) distribution. The perceptual category feature is significantly encoded during a time window centered around 400 ms. b. Euclidean distances in the PC space between the auditory task feature kernels. The perceptual semantic feature kernels are separated in the PC space slightly before the categorical semantic feature kernels (around 400 and 450 ms, respectively). c. Natural speech perception semantic kernels. Word class is processed around 150 ms, and lexical semantics is processed in a wide window around 400 ms. d. Euclidean distances in the PC space between the semantic features during natural speech perception are maximal around 150 ms and 400 ms respectively.