Divergent 1/f and Power Spectral Features Distinguish Cognitive from Headache Phenotypes in Traumatic Brain Injury
Joshua Abata1, Maral Sakayan1, Miranda Saathoff1, Maria Thereza Paulino1, Amanda Fang1, Stephania Tovar Vargas1, Hayley Kristinsson1, Bernadette Boden-Albala1, Mark Mapstone1, Jefferson Chen1, Michael Lopez1, Patrick Chen1, Brian Jung1
1UC Irvine
Objective:
To investigate if cognitive and headache traumatic brain injury (TBI) phenotypes exhibit distinct quantitative EEG (qEEG) features.
Background:
TBI is associated with heterogeneous post-acute symptoms. COACH-CV classification characterizes 7 TBI clinical phenotypes based on patient-reported descriptions. We hypothesized that cognitive and headache phenotypes exhibit distinct qEEG features.
Design/Methods:
Thirteen sequential adults with TBI who underwent scalp EEGs were retrospectively . These patients comprised a sample of patients in our TBI program that obtained EEGs. Cognitive (n = 9) or headache (n = 4) phenotypes were based self-reported most disabling TBI-symptom. EEGs were preprocessed with independent component analysis to remove ocular and ECG artifacts, and band-pass filtered from 0.5 to 55 Hz using a finite impulse response filter. The remaining EEG was binned into 4-second epochs, and power spectral density (PSD) was calculated for each epoch with 25% overlap using Welch’s method. Regional relative power across each canonical frequency band was calculated at the sensor level. 1/f slope was also estimated by fitting a first-degree regression to log-power versus log-frequency between 0.5 and 30 Hz. Inter-group differences were assessed using the Wilcoxon rank sum test.
Results:
The cognitive phenotype exhibited a widespread increase in 1/f slope, reaching significance at T8 and Cz (p < 0.05) and showing trends toward significance at Fp1, Fp2, C4, P3, P4, F8, T6, Fz, and Pz (p < 0.10). Relative power analysis indicated elevated delta and theta power in the cognitive phenotype, particularly in bilateral temporal regions (p < 0.02), whereas the headache phenotype showed higher alpha and beta power.
Conclusions:
EEG power spectral features differentiate TBI patients with cognitive and headache phenotypes. The skew towards lower frequencies in the cognitive phenotype resembles power spectral signatures commonly observed in mild cognitive impairment and Alzheimer's dementia. 1/f and relative power are potential biomarkers for TBI phenotypes that could help guide individualized management.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.