Fig. 2.

(A) Schematic of item-level reinstatement of overlapping, yet incorrect contextual details. A) For each anatomically defined ROI (i.e., bilateral anterior hippocampus – depicted above, L. ITG, bilateral subcallosal gyrus), the pattern of neural activity was extracted for every pre-exposure and detail retrieval trial. Patterns from the pre-exposure phase were averaged across all repetitions of the unique image. (B) Item-level reinstatement of overlapping, yet incorrect contextual details was measured by calculating the similarity between neural patterns during detail retrieval trials (e.g., AB₁ detail retrieval) and neural patterns when participants viewed the overlapping, yet incorrect event context during the pre-exposure phase (e.g., BC₁ pre-exposure; r_match) relative to neural patterns from the pre-exposure phase from other unrelated contexts coded in the same triad bin (e.g., successful inference vs. unsuccessful inference, before vs. after; r_mismatch). That is, if event ABC₁ were a successful inference triad from after the test of directly learned/associative inference trials, the neural patterns associated with AB₁ detail retrieval trials would be correlated with all other ‘BC’ pre-exposure patterns associated with successful inference triads whose detail retrieval questions also occurred after the test of directly learned/associative inference trials. r_match relative to r_mismatch represents the item-specific reinstatement of overlapping, yet incorrect contextual details.