Abstract
OBJECTIVE: Epilepsy and depression share neurobiological origins, and evidence suggests a possible bidirectional relationship that remains poorly understood. This exploratory, cross-sectional study aimed to investigate this relationship by employing magnetic resonance spectroscopic imaging (MRSI) and thermometry (MRSI-t) in patients with temporal lobe epilepsy (TLE) with comorbid depressive symptoms and control participants. This is the first study to combine MRSI and MRSI-t to examine brain temperature and choline abnormalities in regions implicated in seizure onset and depression. METHODS: Twenty-six patients with TLE and 26 controls completed questionnaires and underwent imaging at 3T. Volumetric echo-planar MRSI/MRSI-t data were processed within the Metabolite Imaging and Data Analysis System (MIDAS). Choline (CHO) was quantified as a ratio over creatine (CRE; CHO/CRE). Brain temperature (T(CRE) ) was calculated based on the chemical shift difference of H(2) O relative to CRE's stable location on the ppm spectrum. The Hospital Anxiety and Depression Scale measured anxiety and depressive symptoms. The Chalfont Seizure Severity Scale measured seizure severity in patients with TLE. Two sets of voxelwise independent sample t tests examined group differences in CHO/CRE and T(CRE) maps. Voxel-based multimodal canonical correlation analysis (mCCA) linked both datasets to investigate if, how, and where CHO/CRE and T(CRE) abnormalities were correlated in TLE participants and controls. RESULTS: Compared to controls, patients with TLE reported more depressive symptoms (P = 0.04) and showed CHO/CRE and T(CRE) elevations in left temporal and bilateral frontal regions implicated in seizure onset and depressive disorders (p(FWE) < 0.05). For the TLE group, CHO/CRE levels in temporal and frontal cortices were associated with elevated T(CRE) in bilateral frontal and temporal gyri (r = 0.96), and decreased T(CRE) in bilateral fronto-parietal regions (r = -0.95). SIGNIFICANCE: Abnormalities in T(CRE) and CHO/CRE were observed in seizure-producing areas and in regions implicated in depression. These preliminary findings suggest that common metabolic changes may underlie TLE and depression. Our results suggest further investigations into the proposed bidirectional mechanisms underlying this relationship are warranted.