Figure 2.
Adaptation reduces the theoretical detectability of punctate sensory stimuli. A, Baseline and feature-evoked population spike rate distributions for one example RS unit (see Materials and Methods). For this unit (qualitatively representative of the average), adaptation decreased the mean population response rate, increasing the overlap of the baseline and feature-evoked distributions (right). B, ROC curves for example unit, for the control (black line) and adapted (gray line) conditions, and associated AUROC values. C, Grand average mean (±SEM) theoretical detectability (AUROC) versus stimulus condition for all significantly responsive RS units (***p < 0.0005, Wilcoxon signed-rank test); 300°/s control: mean ± SEM AUROC = 0.81 ± 0.02, adapted: 0.66 ± 0.02, hybrid: 0.66 ± 0.02, control versus adapted: W = 1936.0, p = 2.09 × 10−11; control versus hybrid: W = 1154.0, p = 1.91 × 10−16; adapted versus hybrid: W = 5243.0, p = 0.81, Wilcoxon signed-rank test, N = 119 units from 19 recording sessions; 900°/s control: mean ± SEM AUROC = 0.89 ± 0.02, adapted: 0.73 ± 0.03, hybrid: 0.71 ± 0.03; control versus adapted: W = 166.0, p = 3.18 × 10−6; control versus hybrid: W = 106, p = 1.78 × 10−7; adapted versus hybrid: W = 528.0, p = 0.21, N = 49 units from 8 recording sessions. D, Grand mean (± SEM) time-resolved theoretical detectability (AUROC) for all significantly-responsive RS units (see Materials and Methods), for the control (empty dots) and adapted (filled dots) conditions.