Authors :
Presenting Author: Behnaz Akbarian, M.S. – Vanderbilt University
Lucas Sainburg, B.S – Vanderbilt University; Baxter Rogers, PhD – Vanderbilt University; Dario Englot, PhD, MD – Vanderbilt University; Victoria Morgan, PhD – Vanderbilt University
Rationale:
Surgery is an effective therapy for selected individuals with medically refractory epilepsy. Surgical outcomes are variable and can change over time after surgery, even in patients with homogeneous presurgical clinical characteristics. This study aims to explore the relationship between postsurgical functional connectivity (FC) and surgical outcomes at different time points within and between patients.
Methods:
We studied 32 patients with mesial temporal lobe epilepsy (11 left, age 43.1±11.9 yrs.) who underwent selective amygdalohippocampectomy and 85 healthy controls (age 37.7±13.5 yrs.). T1-weighted MRI (1x1x1 mm
3) for brain segmentation into 113 brain regions (nodes) and resting-state functional MRI (TR=2 s, 3x3x4 mm
3) for FC calculation were collected. Patients underwent a presurgical and at least one postsurgical MRI sessions (n=48; follow-up: 24.4±15.4 months). Controls had one MRI session. Outcomes were classified as seizure-free (n=30, Engel Ia-Ib) or non-seizure-free (n=18, Engel Ic+) at each acquisition.
Two patients had one seizure-free followed by non-seizure-free acquisitions.
To identify postsurgical nodes related to surgical outcomes, we compared node strength, measured as the total FC from that node to all other nodes, between seizure-free and non-seizure-free acquisitions. We then investigated the FC of the nodes identified above to each other node to identify FC patterns related to outcome. To do this we decomposed the FC patterns from all patients into a linear combination of latent patterns which describe each subject’s patterns with unique loading. Then we compared pattern loadings between the two outcome groups. All comparisons were performed using linear mixed models.
Results:
Higher postsurgical node strength in the ipsilateral central operculum (I-CO), contralateral central operculum (C-CO), and ipsilateral posterior insula (I-PIns) was related to seizure freedom, adjusting for time after surgery and duration of disease prior to surgery (Fig 1). There were four latent FC patterns related to outcome in two of the three nodes (Fig 2a, d). Two patterns had higher pattern loadings for non-seizure-free outcomes (Fig 2b, e), while two had higher pattern loadings for seizure-free outcomes (Fig 2c, f), adjusting for time after surgery and duration of disease prior to surgery. As we only had two patients with acquisitions in both outcomes, we could not compare differences in different outcomes in same patient statistically. Conclusions:
We found that the FC node strength of two bilateral nodes in the inferior frontal cortex and one subcortical node is higher in seizure-free post-surgical acquisitions compared to non-seizure-free. From the inferior frontal nodes, we identified latent patterns of FC related to these different outcomes. These results indicate that post-surgical FC outside the temporal lobes may contribute to post-surgical seizure recurrence after mesial temporal resection.
Funding:
Funded by NIH R01 NS075270, R01 NS108445, R01 NS110130, R00 NS097618, and T32 EB021937