CT-Net: an interpretable CNN-Transformer fusion network for fNIRS classification.

IF 2.6 4区 医学 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS
Lingxiang Liao, Jingqing Lu, Lutao Wang, Yongqing Zhang, Dongrui Gao, Manqing Wang
{"title":"CT-Net: an interpretable CNN-Transformer fusion network for fNIRS classification.","authors":"Lingxiang Liao, Jingqing Lu, Lutao Wang, Yongqing Zhang, Dongrui Gao, Manqing Wang","doi":"10.1007/s11517-024-03138-4","DOIUrl":null,"url":null,"abstract":"<p><p>Functional near-infrared spectroscopy (fNIRS), an optical neuroimaging technique, has been widely used in the field of brain activity recognition and brain-computer interface. Existing works have proposed deep learning-based algorithms for the fNIRS classification problem. In this paper, a novel approach based on convolutional neural network and Transformer, named CT-Net, is established to guide the deep modeling for the classification of mental arithmetic (MA) tasks. We explore the effect of data representations, and design a temporal-level combination of two raw chromophore signals to improve the data utilization and enrich the feature learning of the model. We evaluate our model on two open-access datasets and achieve the classification accuracy of 98.05% and 77.61%, respectively. Moreover, we explain our model by the gradient-weighted class activation mapping, which presents a high consistent between the contributing value of features learned by the model and the mapping of brain activity in the MA task. The results suggest the feasibility and interpretability of CT-Net for decoding MA tasks.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"3233-3247"},"PeriodicalIF":2.6000,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Medical & Biological Engineering & Computing","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1007/s11517-024-03138-4","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/5/30 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS","Score":null,"Total":0}
引用次数: 0

Abstract

Functional near-infrared spectroscopy (fNIRS), an optical neuroimaging technique, has been widely used in the field of brain activity recognition and brain-computer interface. Existing works have proposed deep learning-based algorithms for the fNIRS classification problem. In this paper, a novel approach based on convolutional neural network and Transformer, named CT-Net, is established to guide the deep modeling for the classification of mental arithmetic (MA) tasks. We explore the effect of data representations, and design a temporal-level combination of two raw chromophore signals to improve the data utilization and enrich the feature learning of the model. We evaluate our model on two open-access datasets and achieve the classification accuracy of 98.05% and 77.61%, respectively. Moreover, we explain our model by the gradient-weighted class activation mapping, which presents a high consistent between the contributing value of features learned by the model and the mapping of brain activity in the MA task. The results suggest the feasibility and interpretability of CT-Net for decoding MA tasks.

Abstract Image

CT-Net:用于 fNIRS 分类的可解释 CNN 变换器融合网络。
功能性近红外光谱(fNIRS)是一种光学神经成像技术,已被广泛应用于大脑活动识别和脑机接口领域。现有研究针对 fNIRS 分类问题提出了基于深度学习的算法。本文建立了一种基于卷积神经网络和变换器(Transformer)的新方法,命名为 CT-Net,用于指导心算(MA)任务分类的深度建模。我们探索了数据表示的效果,并设计了两种原始发色团信号的时间级组合,以提高数据利用率并丰富模型的特征学习。我们在两个开放获取的数据集上评估了我们的模型,分类准确率分别达到了 98.05% 和 77.61%。此外,我们还通过梯度加权类激活映射来解释我们的模型,结果表明模型学习到的特征贡献值与 MA 任务中的大脑活动映射高度一致。这些结果表明 CT-Net 对 MA 任务解码的可行性和可解释性。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Medical & Biological Engineering & Computing
Medical & Biological Engineering & Computing 医学-工程:生物医学
CiteScore
6.00
自引率
3.10%
发文量
249
审稿时长
3.5 months
期刊介绍: Founded in 1963, Medical & Biological Engineering & Computing (MBEC) continues to serve the biomedical engineering community, covering the entire spectrum of biomedical and clinical engineering. The journal presents exciting and vital experimental and theoretical developments in biomedical science and technology, and reports on advances in computer-based methodologies in these multidisciplinary subjects. The journal also incorporates new and evolving technologies including cellular engineering and molecular imaging. MBEC publishes original research articles as well as reviews and technical notes. Its Rapid Communications category focuses on material of immediate value to the readership, while the Controversies section provides a forum to exchange views on selected issues, stimulating a vigorous and informed debate in this exciting and high profile field. MBEC is an official journal of the International Federation of Medical and Biological Engineering (IFMBE).
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信