Abstracts

Rationale: Temporal lobe epilepsy (TLE) is associated with widespread effects that are not fully explained by focal temporal lobe dysfunction. Sources of variance include changes in alertness, fatigue, and neurocognitive deficits that are unrelated to temporal lobe functions.¹ Recent work by our group suggests that interictal abnormalities in arousal structures may be related to neurocognitive disturbances in TLE.^1,2 Characterizing whole-brain fMRI vigilance signatures could improve the detection of TLE neurocognitive effects in fMRI, and reveal vigilance-related signals as potential biomarkers for epilepsy.

Methods: Simultaneous eyes-closed, resting-state fMRI-EEG data was collected with a 3T MRI and a 32-channel MR-compatible EEG system for nine temporal lobe epilepsy patients (6 males; 44.4yo+/-13) and nine healthy subjects (3 males; 29.8yo+/-10.2). MRI/EEG acquisition and processing was conducted by methods described in Goodale et al.³ An EEG index of vigilance was formed (ratio of power in the alpha (8-12Hz) to theta (3-7Hz) frequency bands, averaged across occipital and parietal electrodes) and temporally aligned to the fMRI data. fMRI patterns linked with vigilance were mapped by correlating each subject’s EEG vigilance index with each voxel’s fMRI signal (Figure 1). A t-test across the absolute value of each group was calculated with FDR correction._x000D_
Since EEG is typically unavailable during fMRI, we previously developed a template approach for extracting moment-to-moment vigilance fluctuations from fMRI data alone.³ Previously, we have demonstrated that this existing vigilance template (derived from a separate control dataset) indicated significant declines in vigilance levels across back-to-back fMRI scans in TLE patients compared to healthy controls, but concurrent EEG data were not available for validation. Here, we evaluate how well this template predicts vigilance fluctuations in TLE patients and a new cohort of healthy controls, evaluating the fMRI-derived index via correlation with the (gold-standard) EEG vigilance index.

Results: Correlations between the EEG vigilance index and fMRI showed magnitude differences between TLE and controls; however, few voxels remained significant after FDR correction (Figure 1). The fMRI-based estimates of vigilance in TLE patients vs controls did not show any significant differences, suggesting that our previously defined template translates well to an epilepsy cohort (Figure 2).

Conclusions: Overall, these results demonstrate that fMRI patterns linked with an EEG measure of vigilance are similar across TLE patients and healthy subjects and our previously established vigilance template, based on healthy controls could be used in epilepsy populations that do not have simultaneously collected EEG. However, a larger sample size is necessary to validate these preliminary findings.

References:_x000D_ 1.  Englot DJ et al. Impaired vigilance networks in temporal lobe epilepsy: Mechanisms and clinical implications. Epilepsia. 2020._x000D_ 2. Gonzalez HFJ et al. Role of the Nucleus Basalis as a Key Network Node in Temporal Lobe Epilepsy. Neurology. 2021._x000D_ 3. Goodale SE et al. fMRI-based detection of alertness predicts behavioral response variability. Elife. 2021.

Funding: NIH R01NS112252 and NSF-GFRP