Temporal lobe epilepsy (TLE) represents 60% of all epilepsy cases and involves the hippocampus resulting in memory and cognitive deficits. Within the hippocampus is the dentate gyrus (DG), a selectivity filter which generates unique representations of contextually similar inputs, a process known as pattern separation. Pattern separation relies on the coordinated activation of multiple types of interneurons (INs) which, in TLE, are susceptible to cell death and reorganize. Without proper inhibition, granule cells (GCs), the main projection neuron, fire imprecisely leading to failure in pattern separation. Two important IN subtypes in the DG are the parvalbumin (PV) INs, which modulates GC firing by delivering reliable perisomatic inhibition thus affecting output signals, and the somatostatin (SOM) INs, which modulates incoming signals by synapsing onto the distal dendrites. However, their individual contributions to pattern separation computation have yet to be determined. Recently, the semilunar granule cells (SGCs), an excitatory neuron identified by their wide dendrites, has been hypothesized to aid in maintaining suppression of the non-firing GCs. How SGCs and GCs differ in molecular and connectivity profiles is currently unknown. Interestingly, SGCs have been shown to be the primary source of perisomatic excitation onto PV-INs, potentially enhancing feedback inhibition onto local GCs. However, how SGCs affect network activity and their contribution to pattern separation in TLE is unknown. I hypothesize that SGCs will show reduced intrinsic pattern separation compared to GCs and that SGC driven PV-IN activity more robustly supports pattern separation than feedback dendritic inhibition by SOM-INs. Furthermore, experimental TLE will disrupt the precision of SGC to PV/SOM-IN mediated inhibition resulting in pattern separation deficits. This proposal will investigate the unique connectome of SGCs and, using an ex vivo temporal pattern separation paradigm as well as a in silico DG network model, to elucidate the contributions of PV-INs and SOM-INs to pattern separation in SGCs and GCs in healthy and epileptic circuits. Together, identification of the local circuit mechanisms underlying dentate pattern separation and how it is impaired during epileptogenesis will pave the way for novel strategies to manage memory related co-morbidities in epilepsy.