Multi-Source Fusion Domain Adaptation Using Resting-State Knowledge for Motor Imagery Classification Tasks

Zhu, Lei; Yang, Junting; Ding, Wangpan; Zhu, Jieping; Xu, Ping; Ying, Nanjiao; Zhang, Jianhai

doi:10.1109/jsen.2021.3101684

Cited by 19 publications

(22 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, there are five subjects (a1, a2, ..., a5) in MI1 dataset, so we have 5 × 4 = 20 S→S tasks, e.g., a1→a2 (Subject a1 as the source domain, and subject a2 as the target domain), and five M→S tasks, e.g., a1, a2, a3, a4→a5. In this paper, the balanced classification accuracy (BCA) is used as the performance measure metric, which has been widely used in EEG classification [11], [52], [53]. The evaluation metric is defined as follow:…”

Section: B Experimental Settingsmentioning

confidence: 99%

Multi-Source Transfer Learning for EEG Classification Based on Domain Adversarial Neural Network

Liu

Zhang

et al. 2023

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

Electroencephalogram (EEG) classification has attracted great attention in recent years, and many models have been presented for this task. Nevertheless, EEG data vary from subject to subject, which may lead to the performance of a classifier degrades due to individual differences. To collect enough labeled data to model would address the issue, but it is often time-consuming and laborintensive. In this paper, we propose a new multi-source transfer learning method based on domain adversarial neural network for EEG classification. Specifically, we design a domain adversarial neural network, which includes a feature extractor, a classifier, and a domain discriminator, and therefore reduce the domain shift to achieve the purpose. In addition, a unified multi-source optimization framework is constructed to further improve the performance, and the result for EEG classification is induced by the weighted combination of the predictions from multiple source domains. Experiments on two publicly available EEG datasets validate the advantages of the proposed method.

show abstract

Section: B Experimental Settingsmentioning

confidence: 99%

Multi-Source Transfer Learning for EEG Classification Based on Domain Adversarial Neural Network

Liu

Zhang

et al. 2023

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

show abstract

“…Two common public MI datasets were used in the experiments, BCI competition IV 1 dataset 1 and dataset 2a, as [6], [16]. Recording paradigms of these two datasets are similar, displayed in Fig.…”

Section: A Datasets and Preprocessingmentioning

confidence: 99%

“…Nine subjects participated in the experiment, and each of them performed 288 trials in each session. For consistency as [6], [16] and labeled samples, we only used training session trials to verify our MSTLs. Each subject had 144 trials, 72 trials for left hand or right hand.…”

Section: A Datasets and Preprocessingmentioning

confidence: 99%

“…Thus, some researchers have adopted multi-source transfer learning (MSTL) to reduce individual differences in MI-BCI field. Multi-source fusion adaptation regularization (MFAR) is based on Euclidean alignment via rest-state knowledge (EARK) [16]. Cai et al integrated two Riemannian manifolds into a TL framework to align marginal distribution and joint distribution simultaneously, named manifold embedded TL (METL) [17].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Multi-Source Transfer Joint Matching Method for Inter-Subject Motor Imagery Decoding

Wei

Jia

et al. 2023

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

Individual differences among different subjects pose a great challenge to motor imagery (MI) decoding. Multi-source transfer learning (MSTL) is one of the most promising ways to reduce individual differences, which can utilize rich information and align the data distribution among different subjects. However, most MSTL methods in MI-BCI combine all data in the source subjects into a single mixed domain, which will ignore the effect of important samples and the large differences in multiple source subjects. To address these issues, we introduce transfer joint matching and improve it to multi-source transfer joint matching (MSTJM) and weighted MSTJM (wM-STJM). Different from previous MSTL methods in MI, our methods align the data distribution for each pair of subjects, and then integrate the results by decision fusion. Besides that, we design an inter-subject MI decoding framework to verify the effectiveness of these two MSTL algorithms. It mainly consists of three modules: covariance matrix centroid alignment in the Riemannian space, source selection in the Euclidean space after tangent space mapping to reduce negative transfer and computation overhead, and further distribution alignment by MSTJM or wMSTJM. The superiority of this framework is verified on two common public MI datasets from BCI competition IV. The average classification accuracy of the MSTJM and wMSTJ methods outperformed other state-of-the-art methods by at least 4.24% and 2.62% respectively. It's promising to advance the practical applications of MI-BCI.

show abstract

“…Chai et al [ 23 ] proposed a novel subspace alignment auto-encoder to reduce the difference in data distribution among subjects or sessions, which combined auto-encoder and subspace alignment in a unified framework by using nonlinear transformations and maximum mean discrepancy (MMD). The divergences of marginal and conditional probability distributions between the different domains can be minimized via MMD [ 24 , 25 ]. Jiang et al [ 26 ] proposed a kernel-based Riemannian manifold domain adaptation framework to align the covariance matrices in the Riemannian manifold, and minimize the conditional distribution distance between the source and target domains based on MMD.…”

Section: Introductionmentioning

confidence: 99%

A dual alignment-based multi-source domain adaptation framework for motor imagery EEG classification

2022

Appl Intell

View full text Add to dashboard Cite

Domain adaptation, as an important branch of transfer learning, can be applied to cope with data insufficiency and high subject variabilities in motor imagery electroencephalogram (MI-EEG) based brain-computer interfaces. The existing methods generally focus on aligning data and feature distribution; however, aligning each source domain with the informative samples of the target domain and seeking the most appropriate source domains to enhance the classification effect has not been considered. In this paper, we propose a dual alignment-based multi-source domain adaptation framework, denoted DAMSDAF. Based on continuous wavelet transform, all channels of MI-EEG signals are converted respectively and the generated time-frequency spectrum images are stitched to construct multi-source domains and target domain. Then, the informative samples close to the decision boundary are found in the target domain by using entropy, and they are employed to align and reassign each source domain with normalized mutual information. Furthermore, a multi-branch deep network (MBDN) is designed, and the maximum mean discrepancy is embedded in each branch to realign the specific feature distribution. Each branch is separately trained by an aligned source domain, and all the single branch transfer accuracies are arranged in descending order and utilized for weighted prediction of MBDN. Therefore, the most suitable number of source domains with top weights can be automatically determined. Extensive experiments are conducted based on 3 public MI-EEG datasets. DAMSDAF achieves the classification accuracies of 92.56%, 69.45% and 89.57%, and the statistical analysis is performed by the kappa value and t- test. Experimental results show that DAMSDAF significantly improves the transfer effects compared to the present methods, indicating that dual alignment can sufficiently use the different weighted samples and even source domains at different levels as well as realizing optimal selection of multi-source domains.

show abstract

Multi-Source Fusion Domain Adaptation Using Resting-State Knowledge for Motor Imagery Classification Tasks

Cited by 19 publications

References 22 publications

Multi-Source Transfer Learning for EEG Classification Based on Domain Adversarial Neural Network

Multi-Source Transfer Learning for EEG Classification Based on Domain Adversarial Neural Network

A Multi-Source Transfer Joint Matching Method for Inter-Subject Motor Imagery Decoding

A dual alignment-based multi-source domain adaptation framework for motor imagery EEG classification

Contact Info

Product

Resources

About