Skip to main content
. Author manuscript; available in PMC: 2023 Aug 11.
Published in final edited form as: Nat Hum Behav. 2023 Mar 2;7(5):740–753. doi: 10.1038/s41562-023-01520-0

Fig. 3 |. Temporal profile of linguistic feature encoding.

Fig. 3 |

a, Mean (lines) and standard deviation (shading) of the first eigenvector of the regression weights assigned to each linguistic feature computed over all neural sites that showed significant encoding for that feature (t > 19; sample sizes: A1/2, 242; P, 174; T, 51; F, 73; L1/2, 55; S, 32 electrodes). The statistics were calculated using a 1,000-sample bootstrap. b, Mean and standard deviation of the peak latencies of the first PCs computed in a (n = 1,000 bootstraps).