Communication of perceptual predictions from the hippocampus to the deep layers of the parahippocampal cortex

Current evidence points to the hippocampus as an essential region coordinating learning and exploiting predictive relationships in the service of perception. However, it remains unclear whether the hippocampus drives the communication of predictions to the sensory cortex or acts as a recipient of predictions from else-where. Here, we collected sub-millimetre 7T fMRI data to investigate neural signals in the medial temporal lobe (MTL). We used layer-specific fMRI to infer the direction of communication between the hippocam-pus and cortex. Specifically, superficial layers of the MTL cortex project to the hippocampus, while deep MTL layers receive feedback projections. Participants performed a task in which auditory cues predicted abstract shapes. Crucially, we omitted the expected shape on 25% of trials, thus isolating the prediction signal from bottom-up input and allowing us to ask: In which direction are predictions communicated between the hippocampus and neocortex? Neural patterns in CA23, pre/parasubiculum and the parahippocampal cortex (PHC) reflected shape-specific predictions. Layer-specific informational connectivity analyses revealed that communication between CA23 and PHC was specific to the deep layers of PHC. These findings are in line with the hippocampus generating predictions through pattern completion in CA23 and feeding these pre-dictions back to the neocortex.