Abstracts

Rationale: Explicit memory is defined as the intentional recollection of previously stored information, while implicit memory describes the system that allows for the more automatic retrieval of information. We previously demonstrated a dissociation of implicit and explicit memory procedures by showing left temporal lobe epilepsy (LTLE) patients with impaired explicit but intact implicit verbal memory (Delvecchio, 2004). To unravel the neural correlates of poorly understood implicit memory neural pathways, and to better localize this intact skill in our medial temporal (hippocampal system deprived) epilepsy patients, we employed functional magnetic resonance imaging (fMRI) during a process dissociation memory paradigm (PDP) (Jacoby, 1998). Methods: A total of 16 matched normal controls (NC) and 12 LTLE patients were asked to memorize 40 five letter words presented visually for 3 seconds prior to scanning. During the scan session (1.5 tesla), subjects were shown three letter word stems along with two types of instructional cues. Subjects were to recall a word from the original memory list when given the cue “OLD”, and to improvise a novel, unseen word from the stem when given the cue “NEW”. Subjects responded orally followed by a delay period to allow the whole-brain blood oxygen level dependent (BOLD) signal to return to baseline. Functional images were post-processed and the key statistical contrasts analyzed using SPM ‘05. Results: Successful response of a new word during the “NEW” condition isolates the explicit memory system while the unintentional recitation of a word from the original list during the “NEW” condition reflects implicit memory (formulas to account for contamination effects from explicit memory were utilized). Implicit and explicit memory scores were calculated demonstrating LTLE patients had deficient explicit recollection (p<.05), but preserved implicit memory compared to NC. When highlighting explicit recollection differences in NC compared to the LTLE patients, the BOLD response appeared strongest in left superior parietal, middle cingulate, and precuneus regions (fig 1). When highlighting implicit memory processing in the LTLE patients relative to NC, the BOLD response in posterior medial and more inferior regions of left cingulate and left occipital cortex appeared strongest (fig 2). In contrast, implicit memory processing in NCs appeared mediated by left superior frontal, left insula regions. Conclusions: The PDP paradigm isolates implicit from explicit memory function. This study provides a unique opportunity to identify implicit memory pathways by examining this process in LTLE patients with a degraded left mesial temporal lobe system. Our data suggests that LTLE patients are able to maintain intact implicit memory function by recruiting unique neural pathways that are not dedicated to that function in healthy normals.