Neuro: Machine Learning Optimized to Detect Neurodegenerative Diseases Pilot Study

Alzheimer’s Disease (AD) is a progressive neurodegenerative disorder that primarily affects cognitive function. Early detection is a crucial factor in slowing the disease’s progression, improving quality of life, and retaining cognitive function for years beyond diagnosis. Traditional detection methods have included a mix of mental tests, Magnetic Resonance Imaging (MRI) scans, and Positron Emission Tomography (PET) scans. At the same time, machine scanning methods boast an accuracy rate of up to 85%. These processes are relatively expensive, and language barriers often hinder cognitive function tests. Neuro is a study program that utilizes artificial intelligence and machine learning to analyze vocal cognitive tests for biomarkers of Alzheimer’s Disease. Using a blend of traditional machine learning (Random Forest (RF), Support Vector Machines (SVM), Gradient Boosting (GB)) mixed with neural networks, such as Recurring Neural Networks (RNNs), Convolution Neural Networks (CNNs), and Feedforward Neural Networks (FNNs), allows Neuro to retain its accuracy within its study. Neuro boasts an accuracy rate of 95%, an 88.3% F1 Score, 95% recall, 82.6% precision, and an AUC (Area Under the Curve) of 0.931. Working in tandem with diagnosis is the explanation of data; for this, we utilized SHAP (SHapley Additive exPlanations) for individual predictions. Machine learning programming is both cost-effective and accessible, allowing it to be utilized in various clinical settings.