Scene-sensitive Medial Temporal Lobe Subregions Are Recruited for the Integration of Non-scene Stimuli

A hallmark feature of episodic memory is the ability to flexibly recombine information across episodes to form new associations and guide behavior. This process, termed associative inference, relies on the hippocampus and surrounding medial temporal lobe (MTL) subregions. We previously found that cross-episode binding was improved when episodes were linked by scenes rather than by faces or objects. Here, we tested whether differential recruitment of category-sensitive MTL subregions underlies these behavioral differences. Participants completed study-test phases of the Associative Inference in Memory task while undergoing fMRI scanning. During the study phase, they encoded overlapping AB and BC pairs. A and C items were always objects. The linking B item was either a face or a scene. At test, memory for the direct (AB, BC) and indirect associations (inferred AC) was tested. Category sensitivity in MTL subregions was tested using an independent functional localizer and the low integration (AB) trials from the study phase of the Associative Inference in Memory task. Within the MTL, no subregions exhibited face sensitivity. The anterior hippocampal head, anterolateral and posteromedial entorhinal cortices, and parahippocampal cortex were identified as scene sensitive. Although accuracy of the indirect inferences did not differ between pairs linked by faces and scenes, MTL subregion recruitment differed across categories. Scene-sensitive subregions in MTL cortex (anterolateral entorhinal cortex, posteromedial entorhinal cortex, and parahippocampal cortex), but not the hippocampus (anterior hippocampal head), were recruited to support associative inference for faces during encoding. These findings suggest that regions in MTL cortex identified as scene sensitive here may be involved in integrating disparate elements of episodes into coherent representations, and may be recruited for non-scene stimuli when integration demands during encoding are high (e.g., during associative inference).