Change in Resting-State functional MRI Connectivity to Measure Individual Response to Epilepsy Surgery
Preprint
Abstract
Objective: This study evaluates fractional power spectrum contribution (fPSC), a biomarker derived from resting-state functional MRI (rs-fMRI), as an indicator of epileptogenic network activity in drug-resistant epilepsy (DRE) patients undergoing surgery. We aimed to assess pre-to post-operative changes in fPSC and their correlation with seizure outcomes. Methods: A retrospective cohort of 56 pediatric DRE patients with pre- and post-operative rs-fMRI were evaluated. Independent component analysis (ICA) was applied to identify resting-state networks (RSNs). Each ICA RSN’s power spectrum within the range of 0.06–0.25 Hz was quantified by fPSC. The change in this fPSC was compared pre- and post-operatively using paired t-tests. Multivariate analyses including correlations with clinical outcomes were evaluated by linear mixed effects models and ANOVA. Results: Among the 56 patients, 80.4% demonstrated greater than 50% seizure reduction post-surgery, with 64.3% achieving seizure freedom. fPSC significantly decreased after surgery (t=3.0, p=0.005), indicating a reduction in epileptogenic network activity. The mixed effects model, controlling for covariates, also showed a significant effect of post-surgical scan on fPSC reduction (χ ² =8.4, df=1, p=0.004). However, there was insufficient evidence to establish an association between changes in fPSC and clinical improvement score (p=0.16) or seizure frequency (p=0.49). Conclusion: The observed reduction in fPSC post-surgery highlights its potential as a biomarker of atypical network activity in epilepsy, offering a network-specific, whole-brain approach independent of anatomical coordinates. However, its lack of correlation with clinical outcomes underscores the need for further refinement and validation to establish fPSC as a reliable measure of epileptogenic burden. HIGHLIGHTS: fPSC analysis using rs-fMRI detects a moderate treatment effect of epilepsy surgery, independent of relative network volume changes. fPSC analysis provides a method to quantify whole-brain network dysfunction in epilepsy without the need for a seizure onset zone identification-hypothesis. A trend towards direct correlation is observed between fPSC post-operative change and seizure outcomes in epilepsy surgery.Keywords
This publication has 56 references indexed in Scilit:
- Short and long term prediction of seizure freedom in drug-resistant focal epilepsy surgeryClinical Neurology and Neurosurgery, 2023
- Reduction in anti-seizure medications use in pediatric patients with pharmacoresistant epilepsy submitted to surgical treatmentChild's Nervous System, 2022
- Five weeks of intermittent transcutaneous vagus nerve stimulation shape neural networks: a machine learning approachBrain Imaging and Behavior, 2021
- Functional connectivity of hippocampus in temporal lobe epilepsy depends on hippocampal dominance: a systematic review of the literatureJournal of Neurology, 2021
- Deep Learning‐based Classification of Resting‐state fMRI Independent‐component AnalysisNeuroinformatics, 2021
- A resting-state functional MRI study on the effect of vagal nerve stimulation on spontaneous regional brain activity in drug-resistant epilepsy patientsBehavioural Brain Research, 2020
- Network‐targeted approach and postoperative resting‐state functional magnetic resonance imaging are associated with seizure outcomeAnnals of Neurology, 2019
- Resective Epilepsy Surgery for Drug-Resistant Focal EpilepsyJAMA, 2015
- An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: Fractional ALFFJournal of Neuroscience Methods, 2008
- Classification of fMRI independent components using IC-fingerprints and support vector machine classifiersNeuroImage, 2007