TY - JOUR T1 - Distinct representational structure and spatial distribution for visual encoding and recall JF - bioRxiv DO - 10.1101/842120 SP - 842120 AU - Wilma A. Bainbridge AU - Elizabeth H. Hall AU - Chris I. Baker Y1 - 2019/01/01 UR - http://biorxiv.org/content/early/2019/11/14/842120.abstract N2 - During memory recall, reinstatement is thought to occur as an echoing of the neural patterns during encoding. However, the precise information represented in these recall traces is relatively unknown, with previous work investigating broad distinctions (e.g. scenes versus objects) or individual images, rarely bridging these levels of information. Further, prior work has primarily used cued recall tasks, where this memory trace may reflect a combination of a cue, its paired stimulus, and their association. Using ultra-high-field (7T) fMRI with an item-based recall task, we conducted an in-depth comparison of encoding and recall along a spectrum of granularity, from broad stimulus class (scenes, objects) to object or scene type (e.g., natural, manmade) to individual categories (e.g. living room, cupcake). In the scanner, human participants viewed a trial-unique item, followed by a distractor task, and then visually recalled the initial item. During encoding, we observed decodable information at all levels of granularity in category-selective visual cortex. Conversely, during recall, only stimulus class was decodable in these cortical regions, with the exception of the medial place area. In hippocampus, information was only decodable during perception and only for stimulus class. A closer look within category-selective cortex revealed a segregation between voxels showing the strongest effects during encoding and during recall. Finally, in a whole-brain analysis, we find the strongest evidence for encoding-recall similarity in regions anterior to category-selective cortex. Collectively, these results suggest recall is not merely a reactivation of encoding patterns, displaying a different granularity of information and spatial distribution from encoding. ER -