Examining Cognitive Shifts Through EEG: Insights from Resting State to Neurofeedback Game Engagement

Gochhait, Saikat; Leonova, Irina; Kiran, Prabha; Salau, Ayodeji Olalekan; Ali, Aitizaz; Ting, Tin Tin · 2025 · Crossref

DOI: 10.47852/bonviewjcce52025505

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the neurophysiological mechanisms underlying cognitive shifts during gamified neurofeedback training, specifically comparing resting-state brain activity with active engagement in a brain-computer interface (BCI) game. While neurofeedback is established for cognitive enhancement and stress regulation, there is limited empirical evidence regarding specific brainwave modulations in sensorimotor regions (C3, C4) among healthy individuals during interactive tasks. The research aims to characterize these dynamics to inform the development of precision-targeted neurofeedback protocols for mental health and performance optimization. The study employed a within-subjects design with twenty healthy participants aged 18–65. EEG data were collected using a single-channel dry-sensor BCI device at a sampling rate of 256 Hz. Each participant underwent a three-minute baseline session with eyes open, followed by a ten-minute neurofeedback game session. The game required participants to modulate their attention, measured via Beta and Alpha wave activity, to control an on-screen avatar, with adaptive difficulty maintaining consistent cognitive load. Data were preprocessed using artifact rejection and band-pass filtering. Analysis involved Power Spectral Density estimation via Welch’s method for five frequency bands (Delta, Theta, Alpha, Beta, Gamma) and a hybrid deep learning model combining 1D Convolutional Neural Networks (1D-CNN) for feature extraction and Bidirectional Long Short-Term Memory (BI-LSTM) networks for temporal modeling. Results indicated significant reductions in Alpha, Beta, and Sensorimotor Rhythm (SMR) power during the neurofeedback task compared to baseline, particularly at C3 and C4 electrodes. Specifically, mean Beta levels at C3 decreased by 10.82% and Alpha by 16.03%, consistent with increased cortical arousal and attentional focus. Conversely, Delta and Theta bands remained statistically unchanged. The 1D-CNN + BI-LSTM model achieved high classification accuracy (79.2–91.4%) in distinguishing baseline from task states and predicted attention scores with a low root mean square error (RMSE of 0.29). Furthermore, the analysis confirmed a significant negative correlation (r = −0.45) between baseline Alpha power and attention performance, suggesting that lower baseline Alpha is associated with enhanced attentional responsiveness during neurofeedback. The findings demonstrate that gamified neurofeedback induces distinct, measurable shifts in sensorimotor brainwave activity, primarily through the suppression of Alpha and Beta rhythms. The integration of deep learning models proved effective in capturing non-linear temporal dependencies and predicting individual responsiveness, outperforming traditional statistical methods. These results provide empirical support for the neurophysiological basis of neurofeedback in healthy populations, highlighting its potential for cognitive augmentation, stress resilience training, and personalized BCI applications. The study underscores the utility of advanced machine learning techniques in decoding real-time neural dynamics for optimized cognitive interventions.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• neuro workload indices