Electroencephalography (EEG) is a non-invasive and cost-effective technique for monitoring brain activity, increasingly used in computer-aided diagnostic systems. While machine learning has shown promise in identifying disease-specific EEG patterns, potentially supporting clinicians in the early diagnosis of neurological conditions, most studies rely on single datasets with homogeneous acquisition protocols and focus exclusively on binary classification (e.g., healthy vs. one disorder). This thesis addresses these limitations by investigating the automatic classification of three neurological disorders — Alzheimer's disease (AD), epilepsy (EP), and amyotrophic lateral sclerosis (ALS) — from EEG signals. To this end, 6 publicly available EEG datasets were collected and integrated into a unified analysis framework, comprising a total of 525 subjects, including 312 healthy controls, 157 patients with Alzheimer's disease, 50 patients with epilepsy, and 6 patients with amyotrophic lateral sclerosis.