Resting-state fMRI seizure onset localization meta-analysis: comparing rs-fMRI to other modalities including surgical outcomes

Varina L Boerwinkle ¹, Mary A Nowlen ², Jesus E Vazquez ³

¹Division of Child Neurology, University of North Carolina, School of Medicine, Chapel Hill, NC, United States.
²Department of Obstetrics and Gynecology, Banner University Medical Center, Phoenix, AZ, United States.
³Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States.
⁴University of North Carolina, School of Medicine, Chapel Hill, NC, United States.
⁵Surgical Neurology Branch, National Institute of Neurological Disorders and Stroke, Bethesda, MD, United States.
⁶Tulane University, School of Medicine, New Orleans, LA, United States.
⁷Department of Psychiatry, University of Arizona College of Medicine, Phoenix, AZ, United States.
⁸Department of Neurosurgery, University of California, San Diego, San Diego, CA, United States.
⁹University of Arizona College of Medicine, Phoenix, AZ, United States.
¹⁰Creighton University School of Medicine, Phoenix, AZ, United States.
¹¹Department of Psychiatry, Brigham & Women's Hospital, Boston, MA, United States.

⁰Division of Child Neurology, University of North Carolina, School of Medicine, Chapel Hill, NC, United States.

Frontiers in Neuroimaging +

|

January 1, 2025

View abstract on PubMed

Summary

Resting-state fMRI shows high sensitivity but low specificity for localizing the seizure onset zone in epilepsy surgery. Agreement varies by modality, with surgical outcomes showing better correlation than EEG or iEEG.

Area Of Science

Neuroimaging
Epilepsy Research
Surgical Planning

Background

Resting-state functional MRI (rs-fMRI) is a potential tool for localizing the seizure onset zone (SOZ) in epilepsy surgery.
Prior meta-analyses suggest agreement with established modalities, but heterogeneity exists.

Purpose Of The Study

To characterize the agreement between rs-fMRI and other modalities for SOZ localization.
To investigate the impact of varied rs-fMRI analysis subtypes and comparative modalities on agreement.
To test the hypothesis that SOZ-signal strength can overcome heterogeneity.

Main Methods

Systematic literature search (PubMed, Embase, Scopus, Web of Science, Google Scholar, April 2010-April 2020) following PRISMA guidelines.
Meta-analysis of odds ratios to measure agreement between rs-fMRI and comparative modalities (EEG, surgical outcome, intracranial EEG, etc.).
Fixed- and random-effects models to assess heterogeneity and differences in agreement across modalities and rs-fMRI subtypes (e.g., ICA, seed-based analysis).

Main Results

25 studies met inclusion criteria, comparing rs-fMRI with EEG, surgical outcome, intracranial EEG, anatomical MRI, EEG-fMRI, and magnetoencephalography.
Significant heterogeneity in agreement odds ratios was observed (p < 0.001), attributed to rs-fMRI analysis variability.
Agreement varied by modality; surgical outcomes showed higher agreement than EEG (p=0.002) and iEEG (p=0.007). Population mean sensitivity was 0.91, specificity 0.09.

Conclusions

rs-fMRI demonstrates high sensitivity but low specificity for SOZ localization, with significant heterogeneity in agreement.
Variability in rs-fMRI analysis methods contributes to inconsistent results.
Agreement is influenced by the comparative modality, with surgical outcomes offering more reliable correlation than electrophysiological methods.

Keywords:

epilepsy surgery meta-analysis neuroimaging resting-state functional MRI seizure network