EEG-Based Characterization of Auditory Attention and Meditation: An ERP and Machine Learning Approach

Omar AlaniEyad Attar^*

• King Abdulaziz University, Jeddah, Saudi Arabia

This scientific investigation explored how meditation influences neural sound stimulus responses by employing EEG techniques during both meditative states and auditory oddball tasks. The study evaluated event-related potentials alongside theta, alpha and beta spectral power while employing machine learning techniques to distinguish meditative states from cognitive tasks. The study utilized data from thirteen participants aged 24 to 58, which researchers obtained through an openly accessible OpenNeuro dataset. Examination of ERPs demonstrated that P300 amplitude showed significant growth when responding to oddball stimuli, which indicates increased attention allocation (p < 0.05). Spectral power analysis demonstrated an increase in frontal alpha and beta power during meditation while central theta power decreased, which suggests reduced cognitive load and enhanced internal focus. Meditation experience showed a statistical relationship with frontal alpha power, where r = 0.45 and p < 0.03. A Random Forest classifier reached 86. The system achieved a 7% accuracy rate in differentiating cognitive from meditative states while identifying P300 amplitude and frontal alpha power, together with beta power as significant predictors. The EEG-based neurofeedback systems demonstrate potential alongside real-time cognitive state detection for healthcare brain-computer interfaces and mental health applications. The study of meditation's effects on brain activity reveals its benefits for emotional regulation and concentration improvement. The research findings deliver strong evidence that meditation induces distinct neural modifications detectable through ERP and spectral analysis. The potential for meditation to enhance cortical efficiency alongside emotion self-regulation indicates its viability as a mental health support tool. The integration of EEG biomarkers with machine learning methods emerges as a potential pathway for real-time cognitive and emotional state monitoring which enables tailored interventions through neurofeedback systems and brain-computer interfaces to boost cognitive function and emotional health across clinical settings and everyday life.