Figure 6.
Single units also respond to written words based on phonemes in pronunciation. (A) PSTHs and raster plots for 2 units (24b and 28d) for the presentation of written (top) or spoken (bottom) words. (B) Firing rates (between 0 and 1000 ms) to words containing a given phoneme are significantly correlated between visual and auditory words (Spearman, unit 24b: ρ = 0.54, unit 28d: ρ = 0.56, P < 0.01). Solid line represents best linear fit. (C) Response latency to written words is delayed by an average of 170 ms compared with auditory words (dashed line). Solid line represents no delay between modalities. (D) Correlation of firing rates between auditory words containing given phonemes between 0 and 1000 ms and written words containing the same phonemes in 300-ms sliding time windows. Windows for both visual and auditory words begin at the time postword onset indicated on the x-axis. The onset of significant correlation for unit 24b was in the 175- to 475-ms time window, and peak correlation was from 330 to 630 ms. For unit 28d, significant correlation started at 325–625 and peaked from 450 to 750 ms.