Fig. 4: Learning of stimulus-reward associations.
From: Perirhinal cortex learns a predictive map of the task environment
a Example of cross-session Hit vs. non-Hit trial decoder performance trained on activity during the reporting period for the session and tested on all other sessions for one animal across training. b Cross-session performance for Hit trial decoders trained on sessionN and tested on sessionN+1 for report activity (left) or pre-stimulus activity (right) across training stages. c Cross-session decoder performance across training stages for report activity (left) or pre-stimulus activity (right). d Example of cross-temporal (CT) decoder for reward conditions trained on report activity across each training session for one animal. The first decodable time point above chance is shown (white dot). e Decodable onset timepoint for the cross-temporal decoder of report activity for decoders trained on hit, false alarm, or correct rejection trials (P < 0.002, F4,282 = 4.44, one-way ANOVA with post-hoc multiple comparison test. Error bars = SEM. Red lines indicate the 95th percentile of shuffled performance (b). n = 70 T1 sessions, 75 T2 sessions, 30 T3 sessions, 79 T4 sessions, 48 T5 sessions from 7 animals.
