Objective:

Background:

PTE affects nearly one-third of severe TBI survivors, impacting long-term outcomes. Identifying early EEG biomarkers that portend PTE may improve risk stratification, an essential step for designing future therapeutic interventions.

Design/Methods:

This case-control study included 45 severe TBI patients (17 with PTE and 28 without PTE), matched for age and Glasgow Coma Scale (GCS) score at admission. EEG data from the first 20 minutes of recordings were preprocessed and analyzed for PSWEs. Descriptive statistics were used for exploratory analysis. Logistic regression models and the leave-one-out cross-validation (LOOCV) technique were used to assess predictive performance. 

Results:

Patients with PTE had significantly longer time in PSWEs (p = 0.040) and a lower median power frequency (MPF) of PSWEs (p = 0.020) on initial EEG recordings, along with increased time in PSWEs between initial and follow-up EEGs (p = 0.035). Lower MPF of PSWEs was associated with increased PTE risk (OR 5.88, 95% CI, 1.08-32.01; p = 0.041). Multivariate logistic regression identified decompressive hemicraniectomy (DHC), time in PSWEs, and MPF of PSWEs on the initial EEG as significant predictors of PTE, achieving an AUC of 0.866 (95% CI: 0.794-0.982, p < 0.0001). This model maintained strong predictive performance under LOOCV (AUC of 0.834, 95% CI: 0.682-0.964, p < 0.0001, accuracy 80%).

Conclusions:

Combining PSWEs biomarkers with DHC status improves the prediction of PTE risk following severe TBI. The model’s robust performance suggests good generalizability and resistance to overfitting. Compared to prior analyses using other EEG features, these results underscore the added value of quantitative PSWE analysis, which captures the temporal dynamics of paroxysmal slowing.