Abstracts

Rationale: Evaluation of the functional integrity of medial and lateral temporal lobe structures is clinically relevant in patients with refractory epilepsy and candidates for surgical treatment. There is no consensus on a MEG procedure (including activation task and signal-s of interest) that is suited for the purpose of assessing the functional integrity of cortico-hyppocampal networks and its contribution to memory. Here we describe our ongoing experience using MEG recordings collected in the context of a bimodal association memory task. Methods: Data from 6 healthy controls and 8 epileptic patients is presented. For all the epileptic patients, the results from phase I evaluation were available. MEG recordings were conducted while the participants performed a bimodal association memory task. This task has an encoding stage and a cued recall stage. During the encoding stage participants are asked to learn the association between a pictogram (visual input) and a word (auditory input) with overlapping presentation. A total of 6 pairs (pictogram-word) are presented during each block of the encoding phase. Following each block of the encoding phase, they are presented with each of the pictograms (visual cue) and they are asked to recall the associated word (cued recall stage). Number of correctly recalled items were recorded. Signal analysis was centered in the characterization of changes during encoding versus recall in two frequencies of interest (Gamma and Theta) and two regions of interest (MTL vs lateral temporal). Using Brainstorm toolboxes (Tadel et al., 2011) and following the averaging of the MEG evoked fields derived during the encoding (visual and bimodal) and cued recall (visual) conditions, a time-frequency decomposition (TF) of the MEG data was conducted with source localization of Theta and Gamma generators. Results: Behaviorally, the performance during cued recall was significantly lower in the group of patients with epilepsy. Healthy volunteers demonstrated a sharp and positive learning curve with >90% retention of newly established audio-visual associations. Patients showed difficulties to establish bimodal associations and lower ability to retain information. In healthy controls, there was a significant (p > 0.05) increase in Theta power bilaterally during the encoding phase and a marginal increment of gamma power (p > 0.08) during successful recall. Sources of Theta activity were primarily distributed in the MTL (p > 0.05) while sources of Gamma activity did show a distinct (lateral Vs medial) distribution (p > 0.01). In patients, poor performance was associated to inexistent (or unilateral) increase of Theta power during the encoding phase and lack of lateral temporal Gamma activity during the recall stage. Conclusions: Our results suggest that the differential functionality of MTL and lateral temporal structures during the encoding and recall of multimodal information can be characterized by using MEG. Successful bimodal associative encoding is associated with changes in MTL functionality (increase in theta power) and successful cued recall is associated with increase in gamma power in lateral temporal lobe structures. Funding: This study was supported by the Brain Initiative (Florida Hospital for Children).