Figure 6
Time-resolved cross classification for false alarms, forgotten items, and subjective old and new, but incorrect answers in the object recognition dataset. These analyses required the relabeling of specific trials in the training dataset (A) False alarms for familiar objects. familiar: new item, response: familiar. unfamiliar: new item, response: new. (B) False alarms for remembered objects. familiar: new item, response: remembered. unfamiliar: new item, response: new (C) Forgotten objects. familiar: old item, response: new. unfamiliar: new item, response: new. (D) Subjective old/new answers in trials with incorrect responses. familiar: new item, response: old. unfamiliar: old item, response: new. Results of the leave-one-subject-out classification for personally familiar and unfamiliar faces is presented in black for comparison (two-sided cluster permutation tests, p < 0.05). Spatio-temporal searchlight results are shown as scalp maps, with classification accuracy scores averaged in 100 ms steps. Sensors and time points forming part of significant clusters are marked (Two-sided spatio-temporal cluster permutation tests, p < 0.05). RA/LA: right/left anterior, RC/LC: right/left central, RP/LP: right/left posterior.
