2026-02-19 00:00 China Papers Foundations & Methods Translated from EN

Improved Spontaneous EEG Signal Decoding Efficiency by Function Predefined Convolutional Neural Network

Summary Researchers propose a function predefined convolutional neural network (FPCNN) for decoding spontaneous EEG in brain-computer interfaces. Its learnable function predefined convolution (FPC) layer searches for the key spatial-frequency parameters of spontaneous EEG so the parameters carry clear physical meaning, and it builds trainable orthogonal detectors on the FPC to capture complex phase-varying signals. On three spontaneous EEG datasets, FPCNN outperformed state-of-the-art methods by 2.09%, 3.08% and 3.41%, with single-round training and testing taking just 67.96 and 19.36 seconds on non-GPU hardware, which the authors say makes it suited to EEG processing in diverse environments.

Why it matters The combination — modest gains over state-of-the-art plus CPU-only training and testing measured in seconds — points to a decoder cheap enough to run in resource-limited or edge settings, a practical constraint that accuracy-focused EEG papers rarely address.

This study, published in IEEE Transactions on Neural Networks and Learning Systems, proposes a function predefined convolutional neural network (FPCNN) for spontaneous EEG BCI decoding, using a learnable function-predefined convolutional layer to search key spatio-frequency parameters and a trainable quadrature detector to capture complex phase-change signals.

On three spontaneous EEG datasets, FPCNN improved performance by 2.09%, 3.08% and 3.41% respectively compared with state-of-the-art methods, with per-epoch training and testing times of only 67.96 and 19.36 seconds in a non-GPU environment, beneficial for EEG processing in diverse settings. PMID 41706793, DOI 10.1109/tnnls.2026.3652882

Compiled by BCIwiki from public sources

Sources · 1

pubmed.ncbi.nlm.nih.gov 2026-02-19

Beijing Research Institute of Mechanical and Electrical Technology Xidian University Decoding Algorithms China

From these organizations

2026-08 Xidian University EEG-fNIRS fusion decodes imagined handwriting 2026-04 Xidian University Xidian Team Drops Transformer Encoder, MLP Decoder Holds 0.94–0.98 AUC

Beijing Research Institute of Mechanical and Electrical Technology → Xidian University →

Read original ↗

Suggest a correction Revisions · none 中 / EN