MEG-Derived Symptom-Sensitive Biomarkers with Long-Term Test-Retest Reliability

Area of Science:

• Neuroscience
• Biomedical Engineering
• Clinical Neurology

Background:

• Chronic sequelae of traumatic brain injury (TBI) present diagnostic and treatment challenges.
• Magnetoencephalography (MEG) offers potential for objective neuroelectric measures.
• Previous MEG studies lacked sufficient test-retest reliability and sensitivity.

Purpose of the Study:

• To evaluate the long-term test-retest reliability of novel MEG-derived regional brain measures.
• To assess the sensitivity of these measures to chronic symptoms in TBI patients.
• To establish the clinical utility of these neuroelectric measures for TBI management.

Main Methods:

• Resting-state MEG recordings from normative (CamCAN) and TBI (TEAM-TBI) cohorts.
• Automated correction of MEG data using a random forest classifier.
• Validation of regional neuroelectric measures against clinical symptom assessments (Brief Symptom Inventory, Insomnia Severity Index) and syndrome adjudication.

Main Results:

• High mean 16-month test-retest reliability (correlation = 0.67) for MEG measures in the CamCAN cohort.
• Significant sensitivity (p < 0.0003) of MEG measures to distinguish TBI patients with and without specific clinical syndromes (pain, psychological, oculomotor, vestibular, cognitive, sleep).
• Demonstrated automatic processing of MEG data without human intervention.

Conclusions:

• Novel MEG-derived regional neuroelectric measures exhibit robust test-retest reliability.
• These measures are sensitive to individual symptoms and clinical syndromes in chronic TBI.
• The findings support the clinical utility of these objective MEG measures for TBI diagnosis and monitoring.