High resolution fMRI of hippocampal subfields and medial temporal cortex during working memory
MetadataShow full item record
Computational models combined with electrophysiological studies have informed our understanding about the role of hippocampal subfields (dentate gyrus, DG; CA subfields, subiculum) and Medial Temporal Lobe (MTL) cortex (entorhinal, perirhinal, parahippocampal cortices) during working memory (WM) tasks. Only recently have functional neuroimaging studies begun to examine under which conditions the MTL are recruited for WM processing in humans, but subfield contributions have not been examined in the WM context. High-resolution fMRI is well suited to test hypotheses regarding the recruitment of MTL subregions and hippocampal subfields. This dissertation describes three experiments using high-resolution fMRI to examine the role of hippocampal subfields and MTL structures in humans during WM. Experiment 1 investigated MTL activity when participants performed a task that required encoding and maintaining overlapping and non-overlapping stimulus pairs during WM. During encoding, activity in CA3/DG and CA1 was greater for stimulus pairs with overlapping features. During delay, activity in CA1 and entorhinal cortex was greater for overlapping stimuli. These results indicate that CA3/DG and CA1 support disambiguating overlapping representations while CA1 and entorhinal cortex maintain these overlapping items. Experiment 2 investigated MTL activity when participants performed a WM task that required encoding and maintaining either low or high WM loads. The results show a load effect in entorhinal and perirhinal cortex during the delay period and suggest that these regions act as a buffer for WM by actively maintaining novel information in a capacity-dependent manner. Experiment 3 investigated MTL activity when participants performed a WM task that required maintaining similar and dissimilar items at different loads. Analysis of a load by similarity interaction effect revealed areas of activity localized to the CA1 subfield. CA1 showed greater activity for higher WM loads for dissimilar, but not similar stimuli. Our findings help identify hippocampal and MTL regions that contribute to disambiguation in a WM context and regions that are active in a capacity-dependent manner which may support long-term memory formation. These results help inform our understanding of the contributions of hippocampal subfields and MTL subregions during WM and help translate findings from animal work to the cognitive domain of WM in humans.