Improving Outcome Prediction for Traumatic Brain Injury From Imbalanced Datasets Using RUSBoosted Trees on Electroencephalography Spectral Power

Abstract

Reliable prediction of traumatic brain injury (TBI) outcomes based on machine learning (ML) that is derived from quantitative electroencephalography (EEG) features has renewed interest in recent years. Nevertheless, the approach has suffered from imbalanced datasets. Hence, to get a reliable predictive model for predicting outcomes, specifically in a high proportion of moderate TBI with good outcomes, could be challenging. This work proposes an improved outcome predictive model that combines the absolute power spectral density (PSD) as input features for training random under-sampling boosting decision trees (RUSBoosted Trees) as a classifier. Resting-state, eyes-closed EEG data were obtained from 27 moderate TBI patients with follow-up visits. Patient outcome at 4-10 weeks to 12-month was dichotomized based on the Glasgow Outcome Scale as poor (GOS score ≤ 4) and good outcomes (GOS score = 5). The predictive values of absolute PSD from five frequency bands: δ (0.5-4Hz), θ (4-7Hz), α (7-13Hz), β (13-30Hz) and γ (30-100Hz) were evaluated to identify the most informative predictors for reliable prediction outcomes. RUSBoosted Trees performed best at discriminating patients into two outcomes categories (G-Mean = 92.95%, TP rate =100%, TN rate = 86.4%) of absolute PSD in δ and γ bands, which was excellent compared to the other state-of-the-art methods. The highest area under the curve (AUC) of absolute PSD in δ (AUC δ = 0.97) and γ (AUC γ = 0.95) revealed their predictive values as robust prognostic markers for prediction outcomes. The RUSBoosted Trees presents a promising result in prognosis prediction of highly imbalanced data, making it an accessible prediction tool for clinical decision-making, unlike the black-box approaches.

Prediction of Recovery from Traumatic Brain Injury with EEG Power Spectrum in Combination of Independent Component Analysis and RUSBoost Model

“…The ensemble DT technique is more flexible and less prone to overfitting (i.e., has a high bias but low variance), demonstrating the generalization power of RUSBoost in predicting outcomes. The present results support the previously reported development of a predictive model using the ensemble DT and resampling is better suited in predicting TBI outcomes than using an individual algorithm (i.e., DT, SVM and k-NN) [40].…”

“…The learning strategy of the RUSBoost algorithms offered advantages in improving the prediction of the poor outcomes with a slight decrease in the good outcomes class. The undersampling strategy, which balances the class distribution in the dataset, is highly beneficial in learning from skewed training data [40,55]. In sleep spindles detection [59], the RUSBoost algorithms enable an automatic sleep spindles detection with an F-measure of 0.70 and sensitivity of 76.9% without requiring threshold calibration.…”

“…On the contrary, Alam et al [49] have found that the selection of artifact removal had distinct effects on the PSD calculation. However, the recent findings in Noor et al [40] showed that the artifacts rejection by the automatic continuous rejection and experts confirmation provided promising results in predicting TBI outcomes of specific frequency bands. Although an automated artifacts rejection can help isolate between the neural or non-neural source components from ICA decomposition, subjective method (i.e., visual classification by experts) is still typically advisable [33,70].…”

“…The great majority of research included in the review [23,28] found that identifying the most informative qEEG that characterizes the recovery outcome level is crucial to ensure early targeted prediction after TBI post-injury. More recently, Noor et al [40] have demonstrated that qEEG features of PSD in δ, θ, α, and γ showed promising results in the prediction of recovery outcome of moderate TBI patients. A key strength is that the frequency distribution of neuronal activity provides information on the patient level of arousal, restful alertness, and general capacity for focus mental activity [8].…”

“…Recent studies revealed that the ML based on qEEG characteristics yields superior performance in classifying the outcomes of TBI patients. The results highlighted that the ML algorithm (i.e., random under-sampling boosting decision trees (RUSBoosted Trees)) that uses qEEG features (i.e., PSD in specific frequency band (e.g., δ, θ, α and γ)) demonstrated promising results to predict outcomes of highly imbalanced moderate TBI dataset [40]. The present study extends our prior work [40] by including a modification of adding an automatic artifacts rejection method (i.e., multiple artifact rejection algorithm (MARA)), an independent component analysis (ICA)-based algorithm in EEG preprocessing steps and exploring their effects on the predictive performance of the RUSBoost prediction model.…”