EEG decoders track memory dynamics.

Area of Science:

• Neuroscience
• Cognitive Science
• Psychology

Background:

• Mnemonic success depends on both encoding and retrieval neural activity.
• Predicting memory performance on a trial-by-trial basis remains a challenge.

Purpose of the Study:

• To determine if electroencephalographic (EEG) activity can reliably predict individual trial encoding and retrieval success.
• To investigate the relationship between neural variability, memory prediction, and item-to-context binding.

Main Methods:

• A delayed recall task involving 98 participants and 576 word lists across 24 sessions.
• Logistic regression classifiers trained on spectral features from EEG data.
• Analysis of neural activity during both word encoding and preceding spoken recall.

Main Results:

• Classifiers accurately predicted whether words would be recalled based on spectral features during encoding and retrieval.
• Neural activity during encoding predicted subsequent recall and the temporal/semantic organization of recalled items.
• Neural variability predictive of memory success was linked to item-to-context binding.

Conclusions:

• EEG spectral features can reliably predict mnemonic success at the individual trial level.
• Neural activity during encoding plays a crucial role in predicting recall and memory organization.
• Findings support the link between neural correlates of memory prediction and the cognitive process of item-to-context binding in episodic memory.