A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification
Functional near-infrared spectroscopy (fNIRS), a non-invasive optical technique, is widely used to monitor brain activities for disease diagnosis and brain-computer interfaces (BCIs). Deep learning-based fNIRS classification faces three major barriers: limited datasets, confusing evaluation criteria...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9828508/ |
| _version_ | 1827927061554003968 |
|---|---|
| author | Zenghui Wang Jun Zhang Yi Xia Peng Chen Bing Wang |
| author_facet | Zenghui Wang Jun Zhang Yi Xia Peng Chen Bing Wang |
| author_sort | Zenghui Wang |
| collection | DOAJ |
| description | Functional near-infrared spectroscopy (fNIRS), a non-invasive optical technique, is widely used to monitor brain activities for disease diagnosis and brain-computer interfaces (BCIs). Deep learning-based fNIRS classification faces three major barriers: limited datasets, confusing evaluation criteria, and domain barriers. We apply more appropriate evaluation methods to three open-access datasets to solve the first two barriers. For domain barriers, we propose a general and scalable vision fNIRS framework that converts multi-channel fNIRS signals into multi-channel virtual images using the Gramian angular difference field (GADF). We use the framework to train state-of-the-art visual models from computer vision (CV) within a few minutes, and the classification performance is competitive with the latest fNIRS models. In cross-validation experiments, visual models achieve the highest average classification results of 78.68% and 73.92% on mental arithmetic and word generation tasks, respectively. Although visual models are slightly lower than the fNIRS models on unilateral finger- and foot-tapping tasks, the F1-score and kappa coefficient indicate that these differences are insignificant in subject-independent experiments. Furthermore, we study fNIRS signal representations and the classification performance of sequence-to-image methods. We hope to introduce rich achievements from the CV domain to improve fNIRS classification research. |
| first_indexed | 2024-03-13T05:47:41Z |
| format | Article |
| id | doaj.art-4df29871c13f49a7b2a7b38af2cbf9be |
| institution | Directory Open Access Journal |
| issn | 1558-0210 |
| language | English |
| last_indexed | 2024-03-13T05:47:41Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| spelling | doaj.art-4df29871c13f49a7b2a7b38af2cbf9be2023-06-13T20:07:50ZengIEEEIEEE Transactions on Neural Systems and Rehabilitation Engineering1558-02102022-01-01301982199110.1109/TNSRE.2022.31904319828508A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy ClassificationZenghui Wang0https://orcid.org/0000-0002-0803-6206Jun Zhang1https://orcid.org/0000-0002-5985-8023Yi Xia2https://orcid.org/0000-0003-4433-6502Peng Chen3https://orcid.org/0000-0002-5810-8159Bing Wang4School of Electrical Engineering and Automation, Anhui University, Hefei, ChinaSchool of Artificial Intelligence, Anhui University, Hefei, ChinaSchool of Artificial Intelligence, Anhui University, Hefei, ChinaNational Engineering Research Center for Agro-Ecological Big Data Analysis and Application, School of Internet, and Institutes of Physical Science and Information Technology, Anhui University, Hefei, ChinaSchool of Electrical and Information Engineering, Anhui University of Technology, Maanshan, ChinaFunctional near-infrared spectroscopy (fNIRS), a non-invasive optical technique, is widely used to monitor brain activities for disease diagnosis and brain-computer interfaces (BCIs). Deep learning-based fNIRS classification faces three major barriers: limited datasets, confusing evaluation criteria, and domain barriers. We apply more appropriate evaluation methods to three open-access datasets to solve the first two barriers. For domain barriers, we propose a general and scalable vision fNIRS framework that converts multi-channel fNIRS signals into multi-channel virtual images using the Gramian angular difference field (GADF). We use the framework to train state-of-the-art visual models from computer vision (CV) within a few minutes, and the classification performance is competitive with the latest fNIRS models. In cross-validation experiments, visual models achieve the highest average classification results of 78.68% and 73.92% on mental arithmetic and word generation tasks, respectively. Although visual models are slightly lower than the fNIRS models on unilateral finger- and foot-tapping tasks, the F1-score and kappa coefficient indicate that these differences are insignificant in subject-independent experiments. Furthermore, we study fNIRS signal representations and the classification performance of sequence-to-image methods. We hope to introduce rich achievements from the CV domain to improve fNIRS classification research.https://ieeexplore.ieee.org/document/9828508/Functional near-infrared spectroscopy (fNIRS)brain–computer interfaces (BCIs)classificationdeep learningGramian angular difference field |
| spellingShingle | Zenghui Wang Jun Zhang Yi Xia Peng Chen Bing Wang A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification IEEE Transactions on Neural Systems and Rehabilitation Engineering Functional near-infrared spectroscopy (fNIRS) brain–computer interfaces (BCIs) classification deep learning Gramian angular difference field |
| title | A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification |
| title_full | A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification |
| title_fullStr | A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification |
| title_full_unstemmed | A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification |
| title_short | A General and Scalable Vision Framework for Functional Near-Infrared Spectroscopy Classification |
| title_sort | general and scalable vision framework for functional near infrared spectroscopy classification |
| topic | Functional near-infrared spectroscopy (fNIRS) brain–computer interfaces (BCIs) classification deep learning Gramian angular difference field |
| url | https://ieeexplore.ieee.org/document/9828508/ |
| work_keys_str_mv | AT zenghuiwang ageneralandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT junzhang ageneralandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT yixia ageneralandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT pengchen ageneralandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT bingwang ageneralandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT zenghuiwang generalandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT junzhang generalandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT yixia generalandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT pengchen generalandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification AT bingwang generalandscalablevisionframeworkforfunctionalnearinfraredspectroscopyclassification |