Abstract

Background: Traumatic brain injury (TBI) is one of the most common causes of epilepsy in humans, with approximately 1 in 10 TBI patients developing post-traumatic epilepsy. The incidence of the initial seizure often occurs within the first couple of weeks after the TBI incident and continues to affect the patient for weeks and even months later. Understanding the pathways affected in those TBI patients who have an increased likelihood of developing post-traumatic epilepsy (PTE) is crucial for therapy and intervention. Neuroimaging prior to the initial seizure onset may contain information for predictive models of epilepsy risk. This study focuses on utilizing Magnetic Resonance Imaging (MRI) and functional MRI (fMRI) to investigate the location and severity of lesions and their impact on underlying functional connectivity networks in the brain. Prior studies offer mixed results on post-traumatic lesion localization and epilepsy risk; however, no study has evaluated the effect of post-traumatic lesions on remote neural networks on post-traumatic epilepsy risk. 

Objective: To assess epilepsy risk in a cohort of patients with post-traumatic contusions identified on MRI two-weeks post-injury using functional MRI to map the remote network effects of lesions. 

Design/Methods: We performed manual lesion (contusion) segmentation on two-week post-injury MRI of participants enrolled in phase one of the Transforming Research and Clinical Knowledge in TBI (TRACK-TBI) study. Lesion core and edema were segmented using co-registered T1 and FLAIR contrasts. Each participant’s T1 image and lesion masks were then normalized to standard space where the lesion will be used as a seed in a seed-based connectivity analysis. For this analysis, we will use normative functional connectomes from the Human Connectome Project and compare lesion network maps of participants who later develop seizures to those with TBI who do not develop seizures. Group analysis will be carried out using the generalized linear model and threshold free cluster enhancement as implemented in FSL.

Results Of 606 TBI participants enrolled in phase one of TRACK-TBI, we identified 100 with at least one contusion identified on two-week post-injury MRI. Average core lesion volume was 2.2 mL and average edema volume was 128 mLs. Further results of the connectivity analyses are forthcoming.

Conclusions: We demonstrate the feasibility of mapping lesions in TBI patients on two-week post-injury MRI. Lesion connectivity analyses are underway.