The Influence of Prior Sensory Context on Meditative Neural States
Journal of Behavioral and Brain Science January 1, 2025 Shenhan Qiu 1 citation
Consumer-grade EEG recordings can distinguish meditation, alert rest, and drowsy rest using machine learning, but the choice of model materially shapes conclusions. Non-linear models (Random Forests and Support Vector Machines) achieve higher predictive performance than Linear Discriminant Analysis (LDA). However, LDA provides transparent, physiologically interpretable weightings linking increased alpha/theta power and reduced beta activity to meditative engagement. Prior context—such as recent sleep, baseline arousal, and mindfulness exposure—can inflate apparent neural markers if not controlled. A two-tiered workflow is recommended: non-linear screening for sensitivity, then linear confirmatory modeling for interpretability with explicit contextual covariates.