Mapping EEG Metrics to Human Affective and Cognitive Models: An Interdisciplinary Scoping Review from a Cognitive...

Background: Electroencephalography (EEG) offers millisecond-precision measurement of neural oscillations underlying human cognition and emotion. Despite extensive research, systematic frameworks mapping EEG metrics to psychological constructs remain fragmented. Objective: This interdisciplinary scoping review synthesizes current knowledge linking EEG signatures to affective and cognitive models from a neuroscience perspective. Methods: We examined empirical studies employing diverse EEG methodologies, from traditional spectral analysis to deep learning approaches, across laboratory and naturalistic settings. Results: Affective states manifest through distinct frequency-specific patterns: frontal alpha asymmetry (8-13 Hz) reliably indexes emotional valence with 75-85% classification accuracy, while arousal correlates with widespread beta/gamma power changes. Cognitive processes show characteristic signatures: frontal-midline theta (4-8 Hz) increases linearly with working memory load, alpha suppression marks attentional engagement, and theta/beta ratios provide robust cognitive load indices. Machine learning approaches achieve 85-98% accuracy for subject identification and 70-95% for state classification. However, significant challenges persist: spatial resolution remains limited (2-3 cm), inter-individual variability is substantial (alpha peak frequency: 7-14 Hz range), and overlapping signatures compromise diagnostic specificity across neuropsychiatric conditions. Evidence strongly supports integrated rather than segregated processing, with cross-frequency coupling mechanisms coordinating affective-cognitive interactions. Conclusions: While EEG-based assessment of mental states shows considerable promise for clinical diagnosis, brain-computer interfaces, and adaptive technologies, realizing this potential requires addressing technical limitations, standardizing methodologies, and establishing ethical frameworks for neural data privacy. Progress demands convergent approaches combining technological innovation with theoretical sophistication and ethical consideration.