Background: A pre-surgical evaluation of cognitive functions in patients with mesial temporal lobe epilepsy (mTLE) is critical. The limitations of the usual brain analysis model were resolved by the spatial Bayesian variable selection (SBVS) method. An Ising and Dirichlet Process (Ising-DP) model considers SBVS and the grouping of a large number of voxels. The present study aimed to identify brain areas involved in episodic memory in patients with right mTLE and controls via the Ising-DP model. The model was extended to include between-subject factors (BSFs), and the results were compared with other classical methods.
Methods: The present cross-sectional study was conducted on 15 patients with right mTLE and 20 controls in Tehran, Iran, in 2018. During functional magnetic resonance imaging, the subjects were tested with the face-encoding memory task, followed by a recognition memory test. The participants demographic factors such as age, sex, marital status, area of residence, and years of schooling were considered to comprise BSFs. The independent t test, the chi-square test, and the correlation test were conducted using the SPSS software (version 20.0). The image processing was carried out using SPM (version 12.0) and MATLAB (version R2014a).
Results: The Ising-DP model appropriately (R2=0.642) detected activated hippocampal areas. The model adjusted for BSFs indicated a better fit by the significant effect of age (P[γ]>0.91), sex (P[γ]>0.87), and years of schooling (P[γ]>0.89). The heat maps exhibited decreased activation in the right hippocampal region in the patients compared with the controls (p <0.0001). Right hippocampal activity had a significant positive correlation with the recognition memory test in the mTLE group (r=0.665) and the control group (r=0.593).
Conclusion: The Ising-DP model was sufficiently sensitive to detect activated areas in our patients with right mTLE during the face-encoding memory task. Since the model adjusted for BSFs improved sensitivity, we recommend the use of more detailed BSFs such as seizure history in future research.