RESUMO
Early-stage Alzheimer's disease (AD) and frontotemporal dementia (FTD) share similar symptoms, complicating their diagnosis and the development of specific treatment strategies. Our study evaluated multiple feature extraction techniques for identifying AD and FTD biomarkers from electroencephalographic (EEG) signals. We developed an optimised machine learning architecture that integrates sliding windowing, feature extraction, and supervised learning to distinguish between AD and FTD patients, as well as from healthy controls (HCs). Our model, with a 90% overlap for sliding windowing, SVD entropy for feature extraction, and K-Nearest Neighbors (KNN) for supervised learning, achieved a mean F1-score and accuracy of 93% and 91%, 92.5% and 93%, and 91.5% and 91% for discriminating AD and HC, FTD and HC, and AD and FTD, respectively. The feature importance array, an explainable AI feature, highlighted the brain lobes that contributed to identifying and distinguishing AD and FTD biomarkers. This research introduces a novel framework for detecting and discriminating AD and FTD using EEG signals, addressing the need for accurate early-stage diagnostics. Furthermore, a comparative evaluation of sliding windowing, multiple feature extraction, and machine learning methods on AD/FTD detection and discrimination is documented.
RESUMO
Accurate sleep stage detection is crucial for diagnosing sleep disorders and tailoring treatment plans. Polysomnography (PSG) is considered the gold standard for sleep assessment since it captures a diverse set of physiological signals. While various studies have employed complex neural networks for sleep staging using PSG, our research emphasises the efficacy of a simpler and more efficient architecture. We aimed to integrate a diverse set of feature extraction measures with straightforward machine learning, potentially offering a more efficient avenue for sleep staging. We also aimed to conduct a comprehensive comparative analysis of feature extraction measures, including the power spectral density, Higuchi fractal dimension, singular value decomposition entropy, permutation entropy, and detrended fluctuation analysis, coupled with several machine-learning models, including XGBoost, Extra Trees, Random Forest, and LightGBM. Furthermore, data augmentation methods like the Synthetic Minority Oversampling Technique were also employed to rectify the inherent class imbalance in sleep data. The subsequent results highlighted that the XGBoost classifier, when used with a combination of all feature extraction measures as an ensemble, achieved the highest performance, with accuracies of 87%, 90%, 93%, 96%, and 97% and average F1-scores of 84.6%, 89%, 90.33%, 93.5%, and 93.5% for distinguishing between five-stage, four-stage, three-stage, and two distinct two-stage sleep configurations, respectively. This combined feature extraction technique represents a novel addition to the body of research since it achieves higher performance than many recently developed deep neural networks by utilising simpler machine-learning models.