Task-Oriented Learning for Automatic EEG Denoising
- 4 March 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 1-15
Abstract
Electroencephalogram (EEG) denoising methods typically depend on manual intervention or clean reference signals. This work introduces a task-oriented learning framework for automatic EEG denoising with only task labels without clean reference signals. EEG recordings are first decomposed into components based on blind source separation (BSS) techniques. A learning-based selector then assigns a retention probability to each component, and the denoised signal is reconstructed as a probability-weighted combination. A downstream proxy-task model evaluates the reconstructed signal, with its task loss supervising the selector in a collaborative optimization scheme that relies solely on task labels, eliminating the need for clean EEG References. Experiments on three datasets, covering two paradigms and multiple noise conditions, demonstrate average gains of 2.59% in task accuracy and 0.80 dB in the signal-to-noise ratio (SNR) under standard signal-quality metrics. Further analyses demonstrate that the task-oriented learning framework is algorithm-agnostic, as it accommodates diverse decomposition techniques and network backbones for both the selector and the proxy model. These results indicate the proposed framework can denoise EEG using only task labels, supporting its potential utility in neuroscience research and EEG-based interaction systems.Keywords
All Related Versions
Funding Information
- National Key Research and Development Program of China (2023YFC2415100)
- National Natural Science Foundation of China (62373351)
- National Natural Science Foundation of China (62503474)
- National Natural Science Foundation of China (82327801)
- National Natural Science Foundation of China (62303463)
- Chinese Academy of Sciences Project for Young Scientists in Basic Research (YSBR-104)
- Beijing Natural Science Foundation (F252068)
- Beijing Natural Science Foundation (4254107)
- Beijing Nova Program (20250484813)
- China Postdoctoral Science Foundation (2024M763535)
- Postdoctoral Fellowship Program of CPSF (GZC20251170)
- Xiaomi Young Talents Program/Xiaomi Foundation
This publication has 45 references indexed in Scilit:
- EEGDiR: Electroencephalogram denoising network for temporal information storage and global modeling through Retentive NetworkComputers in Biology and Medicine, 2024
- Toward reliable signals decoding for electroencephalogram: A benchmark study to EEGNeXBiomedical Signal Processing and Control, 2023
- EEGDnet: Fusing non-local and local self-similarity for EEG signal denoising with transformerComputers in Biology and Medicine, 2022
- IC-U-Net: A U-Net-based Denoising Autoencoder Using Mixtures of Independent Components for Automatic EEG Artifact RemovalNeuroImage, 2022
- Review of challenges associated with the EEG artifact removal methodsBiomedical Signal Processing and Control, 2021
- A novel end-to-end 1D-ResCNN model to remove artifact from EEG signalsNeurocomputing, 2020
- Deep learning with convolutional neural networks for EEG decoding and visualizationHuman Brain Mapping, 2017
- An Improved Denoising Model Based on the Analysis K-SVD AlgorithmCircuits, Systems, and Signal Processing, 2017
- A comparative study of automatic techniques for ocular artifact reduction in spontaneous EEG signals based on clinical target variables: A simulation caseComputers in Biology and Medicine, 2008
- Denoising based on time-shift PCAJournal of Neuroscience Methods, 2007