Abstracts

Rationale: Temporal lobe epilepsy (TLE) is the most common type of partial epilepsy and is the leading cause of refractory epilepsy [Téllez-Zenteno J. F. et al. Epilepsy Res. Treat. 2012]. Cognitive deficits in several domains have been described in TLE patients, including working memory (WM) [Hermann et al., J Int Neuropsychol Soc 2007]. Several white matter regions that interconnect the temporal lobe to the rest of the brain are known to be affected in TLE [Otte, W. M. et al. Epilepsia 2012], which explains why several cognitive functions (such as WM) are affected, despite not having their neural substrate in the affected temporal lobe. Functional MRI (fMRI) studies have shown that the normal bilateral frontoparietal network in WM is impaired in TLE patients [Winston, G. P. et al. Epilepsia 2013]. Here, we investigate cortical activity in response to a WM task, and how this activity correlates with a thorough neuropsychological evaluation in TLE patients.Methods: We recruited 35 patients with diagnosis of TLE and 23 healthy controls. Patients were assessed with WAIS-IV and WMS-IV neuropsychological tests. Structural and functional images were obtained using a 3.0 T Philips MRI scanner. During fMRI scanning participants completed Sternberg’s task, which evaluates WM and divides it into phases of codification, retention and recall. First, we identified the brain regions (cluster p<0.05, corrected) that were active during the retention phase in the controls. Next, we queried the signal change within those regions for each patient. Finally, we correlated these signal changes with neuropsychological evaluations. For statistical analysis, we used Student's t-test for reaction times, Mann-Whitney test for number of correct answers and Pearson's correlation for correlations between BOLD signal and psychometric tests scores.Results: There was a significant difference on reaction times of the task (p=0.001) between controls (mean 1670±324 ms) and patients (mean 2002±407 ms) as well on number of correct answers (p=0.03) between controls (mean 10.9±2.3) and patients (mean 9.3±3.4). Both groups showed cortical activity during retention phase that involved the bilateral frontoparietal network as previously described [Stretton, J. et al. NeuroImage 2012]. There were no significant differences between groups in fMRI brain activity during the task. Correlations are listed in Figure 1: visual working memory index showed a positive correlation with activity of the right opercular gyrus (r=0.47; p=0.01) (A), as well with the right visual cortex (r=0.48; p=0.01) (B). Left premotor cortex, showed a positive correlation with the visual memory index (r=0.51; p=0.006) (C) and with the visual working memory index (r=0.49; p=0.009) (D).Conclusions: Our results show that while patients have relatively small deficits in WM, the activity of the cortical networks that support such a function are predictive of their performance.