Abstracts

Rationale: The relationships between the perception that stress is a factor in seizure control, the physiological and neural responses to stress and seizure frequency are poorly understood in patients with epilepsy. Our objective was to assess the physiological/neural responses to psychosocial stress during functional magnetic resonance imaging (fMRI) in patients with left temporal lobe epilepsy (LTLE) who do and do not believe that stress affects their seizure control. Methods: 35 LTLE patients and 36 healthy control subjects underwent fMRI while performing control (CMT) and stress math (SMT) tasks in which they received auditory positive (non-stressful) and negative (stressful) feedback, respectively. 27 LTLE patients believed stress affected their seizures (LTLE+S); 8 did not (LTLE-S). HC were enrolled as high- (HC+S; n=22) or low-stress (HC-S; n=14) based on the 4-item Perceived Stress Scale. Measures of stress reactivity included fMRI response to SMT negative feedback, % change in salivary cortisol from post-stress to recovery baseline (dCORT), and change in CMT to SMT heart rate (dHR). Subjects were administered the Beck Depression Inventory (BDI-II) prior to MRI. Processing and statistical analysis of fMRI data were performed using AFNI. Regression analysis examined differences in fMRI response to positive and negative feedback between the following: LTLE+S and LTLE-S; HC+S and HC-S. We also performed regression analysis to compare fMRI response to feedback between LTLE patients with poorly-controlled (LTLE+sz; n=18) versus well-controlled seizures (LTLE-sz; n=17). Results: All groups showed significant increase in CMT to SMT heart rate (all p<.001). LTLE+S and LTLE-S did not differ in age, age of epilepsy onset, illness duration, number of antiepileptic drugs, dHR and dCORT (all p>.05), but LTLE+S had more seizures in the past 3 months (p=.005) and higher BDI-II scores (p=.002) than LTLE-S. LTLE+S exhibited widespread increased fMRI activation to negative feedback compared to LTLE-S, but no difference in response to positive feedback (Fig. 1). HC+S and HC-S did not differ in age, dHR and dCORT (all p>.05), but HC+S scored higher on BDI-II than HC-S (p<.001). No differences were observed in activation to positive or negative feedback between HC+S and HC-S. LTLE+sz and LTLE-sz did not differ in demographic, clinical and response variables except LTLE+sz had more seizures in the past 3 months (p=.003) and higher BDI-II scores (p=.002) than LTLE-sz. LTLE+sz also exhibited increased activation to negative feedback compared to LTLE-sz in a number of brain regions, but no difference in response to positive feedback (Fig. 2). Conclusions: Increased CMT to SMT heart rate indicated a physiological response to psychosocial stress, although the increase was similar in all groups. Increased fMRI activation to negative feedback in LTLE+S and in LTLE+sz suggest that the neural response to psychosocial stress can be modulated by both stress perception and seizure status in patients with LTLE.