Neural Oscillations for Encoding and Decoding Declarative Memory using EEG Signals

Declarative memory has been studied for its relationship with remembering daily life experiences. Previous studies reported changes in power spectra during encoding phase related to behavioral performance, however decoding phase still needs to be explored. This study investigates neural oscillations changes related to memory process. Participants were asked to perform a memory task for encoding and decoding phase while EEG signals were recorded. Results showed that for encoding phase, there was a significant decrease of power in low beta, high beta bands over fronto-central area and a decrease in low beta, high beta and gamma bands over left temporal area related to successful subsequent memory effects. For decoding phase, only significant decreases of alpha power were observed over fronto-central area. This finding showed relevance of beta and alpha band for encoding and decoding phase of a memory task respectively.

💡 Research Summary

The present study investigates how neural oscillations measured by electroencephalography (EEG) differ between successful and unsuccessful trials in a pure declarative memory task, focusing separately on the encoding (study) and decoding (recognition) phases. Seven right‑handed healthy adults (age 21‑31) participated. During encoding, participants viewed five lists of 50 common English nouns (total 250 words). Each word appeared for 2 seconds after a 1‑second fixation, and participants classified the word as abstract or concrete, a judgment unrelated to later memory. After a 20‑minute arithmetic distraction task, the decoding phase presented all previously studied words intermixed with 150 novel words. Participants indicated “old” or “new” on a 4‑point confidence scale.

EEG was recorded with a 62‑channel BRAINAMP system at 1 kHz (ground Fpz, reference FCz) and later down‑sampled to 250 Hz. Signals were band‑pass filtered (0.5‑40 Hz) and baseline‑corrected across the entire epoch. Epochs spanned –1 s to +2 s relative to stimulus onset and were divided into 100 ms sliding windows. Power spectral density was computed via fast Fourier transform for five canonical bands: theta (3‑7 Hz), alpha (7‑13 Hz), low beta (13‑17 Hz), high beta (17‑30 Hz), and gamma (30‑40 Hz).

Statistical comparisons between remembered and forgotten trials employed non‑parametric Wilcoxon rank‑sum tests for behavioral measures and Kruskal‑Wallis tests for spectral power across five scalp regions (pre‑frontal, fronto‑central, parietal, left temporal, right temporal). Post‑hoc tests were Bonferroni‑corrected (α = 0.05).

Behaviorally, participants remembered on average 84.8 % of studied words with a false‑positive rate of 18.6 %. Reaction times during decoding were significantly longer for forgotten items (≈1030 ms) than for remembered items (≈510 ms; p = 0.03). No significant relationship was found between encoding reaction time and later memory performance.

Neurophysiologically, the encoding phase showed a robust decrease in low‑beta (13‑17 Hz) and high‑beta (17‑30 Hz) power over fronto‑central electrodes both before stimulus onset and during stimulus presentation. Over left temporal sites, low‑beta, high‑beta, and gamma (30‑40 Hz) power were also reduced. These findings suggest that a pre‑stimulus state characterized by reduced beta activity—potentially reflecting expectancy or attentional readiness—facilitates successful encoding, while gamma reductions may relate to early consolidation processes in temporal cortex.

During the decoding phase, the only significant spectral difference was a decrease in alpha (7‑13 Hz) power over fronto‑central sites during stimulus presentation for remembered versus forgotten items. Alpha suppression is commonly linked to increased cortical excitability and attentional allocation, indicating that successful retrieval engages attentional mechanisms that manifest as reduced alpha power. No consistent beta or gamma effects were observed in decoding.

The authors discuss how low‑beta reductions align with prior work linking beta desynchronization to attentional shifts and memory formation, while gamma reductions may reflect reduced local synchrony associated with successful encoding. The alpha decrease during retrieval is interpreted as a marker of semantic processing and attentional engagement, consistent with earlier studies.

Limitations include the small sample size, an imbalance between the number of remembered and forgotten trials (participants performed above typical levels, possibly due to the high familiarity of the word pool), and the absence of a long‑delay retest to assess true long‑term retention. Moreover, the baseline correction across the entire epoch may have attenuated some effects.

In conclusion, the study demonstrates that distinct frequency bands and scalp regions are differentially involved in the encoding and decoding of declarative memories: beta and gamma dynamics dominate successful encoding, whereas alpha modulation characterizes successful retrieval. These results provide a neurophysiological basis for future brain‑computer interface applications or non‑invasive brain stimulation protocols aimed at enhancing memory performance. Future work should expand the participant pool, incorporate delayed recall sessions, and explore real‑time modulation of the identified oscillatory signatures.