Cognitive-state decoding

Estimating an internal cognitive variable — engagement, workload, fatigue, confidence, encoding success, or attentional focus — from neural activity, as opposed to decoding an intended movement or word. It underlies both diagnostic monitoring and the trigger logic of state-dependent stimulation. Signals range from scalp-EEG features (frontal-midline theta, parietal alpha, the P300) and fNIRS to intracranial population activity, and the target variable is often defined by behavior (e.g. subsequently-remembered vs forgotten) rather than by a physical act.

The hard part is that cognitive states are latent, slow, subject-specific, and confounded: an alpha increase can mean drowsiness or relaxed attention; a workload marker can be driven by task difficulty or by stress. Robust cognitive-state decoding therefore needs careful construct validity, within-subject calibration, and honest reporting, and is especially prone to overclaiming in consumer neurotechnology.