Fig. 7: Omission responses localize to different areas than repetition suppression.

A We estimated the degree of response suppression to repeated stimulation and prediction errors using two additional stimulus sequences. The oddball sequence (top) was structured identically to the omission sequence, but omissions were replaced with deviant tones (orange). The many standards control sequence (MSC, bottom) leaves the oddball sounds in identical positions (green), but randomizes the frequency of the standards. As in previous studies, the prediction error (PE) was the difference between the responses to identical tones when they occurred as deviant or inside the MSC. Repetition suppression (RS) was the difference in response between the tone when it was inside the MSC and when it was a standard in the oddball sequence. To prevent frequency-specific confounds, responses to 4 kHz and 32 kHz were averaged for each condition. We also present the corresponding indices of PE and RS in Supplementary Fig. 6. B Responses to the same tone when occurring as a deviant, standard, or in the MSC for the primary region (top) and in the omission-responsive region (bottom). Single mouse average (m193). C Responses map to the tone inside the MSC are most dominant in A1, A2, and DP (same animal, response average), omission-responsive area shown as a red line. D Response map to the tone when it is the standard in the oddball sequence. AAF shows the least adaptation (same animal, response average). E Response map to the tone when it is an oddball/deviant inside the oddball sequence shows high responses widely, with the highest responses in A1, A2, and DP. F Population summary (N = 8 areas) of responses across areas. A1 and A2 have the highest oddball responses, compared to smaller responses in other areas. Across-area significance was assessed using Kruskal–Wallis ANOVA for the different measures (p-values on the right), and significance stars per area indicate p < 0.05 (*), <0.01 (**), and <0.001 (***) (Wilcoxon signed-rank test, with false discovery rate correction via Benjamini–Hochberg); same for (I). G Repetition suppression of the standard tone is focused on A1 and A2 (red = more suppression), while the omission-responsive area shows comparatively little suppression, i.e., overall negative iRS. H The prediction-error index was positive throughout the auditory cortex. I Population summary of iRS and iPE across subareas of auditory cortex (N = 8 areas), where iRS significantly differed between areas and was strongest in A1/A2, while near 0 in the subareas of TeA. iPE was overall positive, with the highest values in the core areas A1/2/AAF and vTeA. Source data are provided with this paper.