NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery
Optical brain monitoring, such as near-infrared spectroscopy (NIRS), has facilitated numerous brain studies, including those based on machine learning techniques. A large and diverse dataset is necessary for training machine learning algorithms to avoid overfitting a limited amount of data. However,...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2023-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/10089413/ |
_version_ | 1797839191336288256 |
---|---|
author | Zephaniah Phillips V Seung-Ho Paik Seung-Hyun Lee Eun-Jeong Choi Beop-Min Kim |
author_facet | Zephaniah Phillips V Seung-Ho Paik Seung-Hyun Lee Eun-Jeong Choi Beop-Min Kim |
author_sort | Zephaniah Phillips V |
collection | DOAJ |
description | Optical brain monitoring, such as near-infrared spectroscopy (NIRS), has facilitated numerous brain studies, including those based on machine learning techniques. A large and diverse dataset is necessary for training machine learning algorithms to avoid overfitting a limited amount of data. However, recruiting sufficient subjects is challenging owing to time and budget constraints. Therefore, we propose an NIRS data generation algorithm that scales NIRS signal components, such as hemodynamic response function, physiological systemic noise, and instrumental spike noise, based on the source-detector distance to augment the training data. Experimental self-paced left- and right-hand motor imagery data were augmented with generated NIRS data to train a convolutional neural network and classify the motor imagery data. Augmenting the training dataset with 1000 generated data increased the classification accuracy to 86.3 ± 4.1%, a 26% increase compared with training on experimental data only. In addition, we applied Guided Gradient-weighted Class Activation Mapping (Grad-CAM) to visualize the class discriminative features of the input data. The peaks of Guided Grad-CAM heatmaps aligned with the oxy-hemoglobin peaks during self-paced motor imagery. We concluded that the increased cerebral oxygenation, especially in the contralateral hemisphere, was the class-discriminative feature for classifying left- and right-hand motor imagery. |
first_indexed | 2024-04-09T15:54:06Z |
format | Article |
id | doaj.art-7a2c436ac37747af840e18dc00101c67 |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-04-09T15:54:06Z |
publishDate | 2023-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-7a2c436ac37747af840e18dc00101c672023-04-25T23:00:42ZengIEEEIEEE Access2169-35362023-01-0111373133732310.1109/ACCESS.2023.326348910089413NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor ImageryZephaniah Phillips V0https://orcid.org/0000-0001-5021-2368Seung-Ho Paik1https://orcid.org/0000-0001-6993-3229Seung-Hyun Lee2https://orcid.org/0000-0002-1109-6787Eun-Jeong Choi3https://orcid.org/0009-0008-7534-4680Beop-Min Kim4https://orcid.org/0000-0002-1056-5078Global Health Technology Research Center, College of Health Science, Korea University, Seoul, Republic of KoreaGlobal Health Technology Research Center, College of Health Science, Korea University, Seoul, Republic of KoreaInterdisciplinary Program in Precision Public Health, Korea University, Seoul, Republic of KoreaDepartment of Biomedical Engineering, Korea University, Seoul, Republic of KoreaDepartment of Biomedical Engineering, Korea University, Seoul, Republic of KoreaOptical brain monitoring, such as near-infrared spectroscopy (NIRS), has facilitated numerous brain studies, including those based on machine learning techniques. A large and diverse dataset is necessary for training machine learning algorithms to avoid overfitting a limited amount of data. However, recruiting sufficient subjects is challenging owing to time and budget constraints. Therefore, we propose an NIRS data generation algorithm that scales NIRS signal components, such as hemodynamic response function, physiological systemic noise, and instrumental spike noise, based on the source-detector distance to augment the training data. Experimental self-paced left- and right-hand motor imagery data were augmented with generated NIRS data to train a convolutional neural network and classify the motor imagery data. Augmenting the training dataset with 1000 generated data increased the classification accuracy to 86.3 ± 4.1%, a 26% increase compared with training on experimental data only. In addition, we applied Guided Gradient-weighted Class Activation Mapping (Grad-CAM) to visualize the class discriminative features of the input data. The peaks of Guided Grad-CAM heatmaps aligned with the oxy-hemoglobin peaks during self-paced motor imagery. We concluded that the increased cerebral oxygenation, especially in the contralateral hemisphere, was the class-discriminative feature for classifying left- and right-hand motor imagery.https://ieeexplore.ieee.org/document/10089413/Cerebral oxygenationclass activation mappingconvolutional neural networkfunctional near-infrared spectroscopymachine learningoptical monitoring |
spellingShingle | Zephaniah Phillips V Seung-Ho Paik Seung-Hyun Lee Eun-Jeong Choi Beop-Min Kim NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery IEEE Access Cerebral oxygenation class activation mapping convolutional neural network functional near-infrared spectroscopy machine learning optical monitoring |
title | NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery |
title_full | NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery |
title_fullStr | NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery |
title_full_unstemmed | NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery |
title_short | NIRS Data Augmentation Technique to Detect Hemodynamic Peaks During Self-Paced Motor Imagery |
title_sort | nirs data augmentation technique to detect hemodynamic peaks during self paced motor imagery |
topic | Cerebral oxygenation class activation mapping convolutional neural network functional near-infrared spectroscopy machine learning optical monitoring |
url | https://ieeexplore.ieee.org/document/10089413/ |
work_keys_str_mv | AT zephaniahphillipsv nirsdataaugmentationtechniquetodetecthemodynamicpeaksduringselfpacedmotorimagery AT seunghopaik nirsdataaugmentationtechniquetodetecthemodynamicpeaksduringselfpacedmotorimagery AT seunghyunlee nirsdataaugmentationtechniquetodetecthemodynamicpeaksduringselfpacedmotorimagery AT eunjeongchoi nirsdataaugmentationtechniquetodetecthemodynamicpeaksduringselfpacedmotorimagery AT beopminkim nirsdataaugmentationtechniquetodetecthemodynamicpeaksduringselfpacedmotorimagery |