Construction of machine learning models for recognizing comorbid anxiety in epilepsy patients based on their clinical and quantitative EEG features.

BACKGROUND: This study aimed to construct prediction models for the recognizing of anxiety disorders (AD) in patients with epilepsy (PWEs) by combining clinical features with quantitative electroencephalogram (qEEG) features and using machine learning (ML). METHODS: Nineteen clinical features and 20-min resting-state EEG were collected from 71 PWEs comorbid with AD and another 60 PWEs without AD who met the inclusion-exclusion criteria of this study. The EEG were preprocessed and 684 Phase Locking Value (PLV) and 76 Lempel-Ziv Complexity (LZC) features on four bands were extracted. The Fisher score method was used to rank all the derived features. We constructed four models for recognizing AD in PWEs, whether PWEs based on different combinations of features using eXtreme gradient boosting (XGboost) and evaluated these models using the five-fold cross-validation method. RESULTS: The prediction model constructed by combining the clinical, PLV, and LZC features showed the best performance, with an accuracy of 96.18%, precision of 94.29%, sensitivity of 98.33%, F1-score of 96.06%, and Area Under the Curve (AUC) of 0.96. The Fisher score ranking results displayed that the top ten features were depression, educational attainment, α_P3LZC, α_T6-PzPLV, α_F7LZC, ß_Fp2-O1PLV, θ_T4-CzPLV, θ_F7-PzPLV, α_Fp2LZC, and θ_T4-PzPLV. CONCLUSIONS: The model, constructed by combining the clinical and qEEG features PLV and LZC, efficiently identified the presence of AD comorbidity in PWEs and might have the potential to complement the clinical diagnosis. Our findings suggest that LZC features in the α band and PLV features in Fp2-O1 may be potential biomarkers for diagnosing AD in PWEs.

A model for the diagnosis of anxiety in patients with epilepsy based on phase locking value and Lempel-Ziv complexity features of the electroencephalogram.

OBJECTIVE: Anxiety disorders (AD) are critical factors that significantly (about one-fifth) impact the quality of life (QoL) in patients with epilepsy (PWE). Objective diagnostic methods have contributed to the identification of PWE susceptible to AD. This study aimed to identify AD in PWE by constructing a diagnostic model based on the phase locking value (PLV) and Lempel-Ziv Complexity (LZC) features of the electroencephalogram (EEG). METHODS: EEG data from 131 patients with epilepsy (PWE) were enrolled in this study. Patients were divided into two groups, anxiety disorder (AD, n = 61) and non-anxiety disorder (NAD, n = 70), according to the Hamilton Rating Scale for Anxiety (HAM-A). Support vector machine (SVM) and K-Nearest-Neighbor(KNN) algorithms were used to construct three models - the PLVEEG, LZCEEG, and PLVEEG + LZCEEG feature models. Finally, the area under the receiver operating characteristic curve (AUC) and statistical analyses were performed to evaluate the model performance. RESULTS: The efficiency of the KNN-based PLCEEG + LZCEEG feature model was the best, and the accuracy, precision, recall, F1-score, and AUC of the model after five-fold cross-validations scores were 87.89 %, 82.27 %, 98.33 %, 88.95 %, and 0.89, respectively. When the model efficiency was optimal, 29 EEG features were suggested. Further analysis of these features indicated 22 EEG features that were significantly different between the two groups, including 50 % features of the alpha (α)-band. CONCLUSIONS: The PLVEEG + LZCEEG model features can identify AD in PWE. The PLVEEG and LZCEEG characteristics of the α-band may further be explored as potential biomarkers for AD in PWE.

The potential of the Lempel-Ziv complexity of the EEG in diagnosing cognitive impairment in patients with temporal lobe epilepsy.

AIM: To analyze whether the Lempel-Ziv Complexity (LZC) in quantitative electroencephalogram differs between the temporal lobe epilepsy (TLE) patients with or without cognitive impairment (CI) and explore the diagnostic value of LZC for identifying CI in TLE patients. METHODS: Twenty-two clinical features and 20-min EEG recordings were collected from 48 TLE patients with CI and 27 cognitively normal (CON) TLE patients. Seventy-six LZC features were calculated for 19 leads in four frequency bands (alpha, beta, delta, and theta). The clinical and LZC features were compared between the two groups. A support vector machine (SVM) was subsequently constructed using the leave-one-out method of cross-validation for LZC features with statistical differences. RESULTS: Regarding the clinical features, the level of education (p < .001), hippocampal atrophy and sclerosis (p = .029), and depression (p = .037) were statistically different between the two groups. For the LZC features, there were statistically significant differences in the alpha (Fp1, Fz, Cz, Pz, C3, C4, T3, T4, T5, T6, F3, F4, F7, F8, O1, and O2), beta (Fp2), and theta (F7) oscillations. The mean LZC in the alpha band was higher in the TLE-CI group than that in the CON group, and there were no differences in the remaining bands. The SVM model showed 74.51% accuracy, 79.63% sensitivity, 84.30% F1 score, 68.75% specificity, and .85 area under the curve scores. CONCLUSIONS: The LZC in the alpha band might have the potential to be used as a biomarker for the diagnosis of TLE combined with CI. The TLE-CI group, on the other hand, exhibited a higher degree of complexity in alpha oscillations, which were widespread and occurred in all brain regions.

An objective model for diagnosing comorbid cognitive impairment in patients with epilepsy based on the clinical-EEG functional connectivity features.

Objective: Cognitive impairment (CI) is a common disorder in patients with epilepsy (PWEs). Objective assessment method for diagnosing CI in PWEs would be beneficial in reality. This study proposed to construct a diagnostic model for CI in PWEs using the clinical and the phase locking value (PLV) functional connectivity features of the electroencephalogram (EEG). Methods: PWEs who met the inclusion and exclusion criteria were divided into a cognitively normal (CON) group (n = 55) and a CI group (n = 76). The 23 clinical features and 684 PLV EEG features at the time of patient visit were screened and ranked using the Fisher score. Adaptive Boosting (AdaBoost) and Gradient Boosting Decision Tree (GBDT) were used as algorithms to construct diagnostic models of CI in PWEs either with pure clinical features, pure PLV EEG features, or combined clinical and PLV EEG features. The performance of these models was assessed using a five-fold cross-validation method. Results: GBDT-built model with combined clinical and PLV EEG features performed the best with accuracy, precision, recall, F1-score, and an area under the curve (AUC) of 90.11, 93.40, 89.50, 91.39, and 0.95%. The top 5 features found to influence the model performance based on the Fisher scores were the magnetic resonance imaging (MRI) findings of the head for abnormalities, educational attainment, PLV EEG in the beta (ß)-band C3-F4, seizure frequency, and PLV EEG in theta (θ)-band Fp1-Fz. A total of 12 of the top 5% of features exhibited statistically different PLV EEG features, while eight of which were PLV EEG features in the θ band. Conclusion: The model constructed from the combined clinical and PLV EEG features could effectively identify CI in PWEs and possess the potential as a useful objective evaluation method. The PLV EEG in the θ band could be a potential biomarker for the complementary diagnosis of CI comorbid with epilepsy.