Network dysfunction in pre and postsurgical epilepsy: connectomics as a tool and not a destination

Abstract

Purpose of reviewPatients with focal drug-resistant epilepsy (DRE) sometimes continue to have seizures after surgery. Recently, there is increasing interest in using advanced network analyses (connectomics) to better understand this problem. Connectomics has changed the way researchers and clinicians view DRE, but it must be applied carefully in a hypothesis-driven manner to avoid spurious results. This review will focus on studies published in the last 18 months that have thoughtfully used connectomics to advance our fundamental understanding of network dysfunction in DRE -hopefully for the eventual direct benefit to patient care. Recent findingsImpactful recent findings have centered on using patient-specific differences in network dysfunction to predict surgical outcome. These works span functional and structural connectivity and include the modalities of functional and diffusion magnetic resonance imaging (MRI) and electrophysiology. Using functional MRI, many groups have described an increased functional segregation outside of the surgical resection zone in patients who fail surgery. Using electrophysiology, groups have reported network characteristics of resected tissue that suggest whether a patient will respond favorably to surgery.

The Interictal Suppression Hypothesis in Focal Epilepsy: Electrographic and Structural Evaluation

“…The SOZ and PZ power was not significantly altered in the alpha band [8][9][10][11][12] Hz] (Figure 2B). The SOZ and PZ power was elevated compared to baseline in both beta [13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30] and gamma bands , but post-hoc multiple comparisons revealed that the SOZ elevation in power was only significantly different from non-involved regions in the beta band and only from PZs in the gamma band (Figure 2C-D, beta: SOZ single-population t-test p-value=9.00e-3, one-way ANOVA p-value=2.98e-2, post-hoc pairwise t-test comparison SOZ-Non p-value=2.24e-2; gamma: SOZ single-population t-test p-value=6.70e-3, one-way ANOVA p-value=1.11e-2, post-hoc pairwise t-test comparison SOZ-PZ p-value=1.33e-2).…”

“…75,[87][88][89][90][91] Overall, when presumably healthy regions were stimulated, the SOZ theta [4-8 Hz] relative band power was markedly decreased from pre-stimulation baseline. Conversely, SOZ beta [13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30] and gamma relative band power increased when healthy regions were stimulated. Increased inward evoked gamma band power in SOZs is in alignment with previous low-frequency stimulation findings.…”

“…To quantify the overall direction of a node's connectivity, reciprocal connectivity was calculated as the difference between each node's inward and outward connectivity. All connectivity was analyzed in the theta [4][5][6][7], alpha [8][9][10][11][12], beta [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30], and gamma oscillatory bands.…”

“…a network). [21][22][23][24][25][26] Specifically, recent work has shown that SOZs exhibit increased inward interictal connectivity from other nodes of the brain. 20,27 This work has led us to propose the Interictal Suppression Hypothesis (ISH), which posits that the SOZ is tonically suppressed by other areas of the brain during the resting state to prevent seizure initiation (Figure 1).…”

Deep learning distinguishes focal epilepsy groups using connectomes: Feasibility and clinical implications

“…Moreover, there is growing interest in multimodal image integration pipelines suitably designed for the clinical workflow [34]. The value of connectomics in unveiling network alterations in drug-resistant focal epilepsy was recently underscored [35]. Leveraging deep learning in combination with network neuro-science presents opportunities to identify salient features in high-dimensional imaging datasets [36].…”