Deep Leaning-Based Approach for Mental Workload Discrimination from Multi-channel fNIRS
As a non-invasive optical neuroimaging technique, functional near infrared spectroscopy (fNIRS) is currently used to assess brain dynamics during the performance of complex works and everyday tasks. However, the deep learning approaches to distinguish stress levels based on the changes of hemoglobin concentrations have not yet been extensively investigated. In this paper, we evaluated the efficiencies of advanced methods differentiating the rest and task periods during stroop task experiments. First, we explored that the apparent changes of oxy-hemoglobin (HbO) and deoxy-hemoglobin (HbR) concentrations associated with two mental stages did exist across each participant. Then, a novel discrimination framework was studied. Deep learning approaches, including convolutional neural network (CNN), deep belief networks (DBN), have enabled better classification accuracies of 84.26 ± 9.10% and 65.43 ± 1.59% as our preliminary study.
KeywordsStroop task experiments Functional near infrared spectroscopy Convolutional neural networks Deep belief networks
This work was supported by the Basic Science Research Program through the NRF funded by the Ministry of Education (NRF-2017R1D1A1B03036423) and the Brain Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2016M3C7A1905477, NRF-2014M3C7A1046050). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with animals performed by any of the authors.
- 2.Nguyen, H. T., Van Nguyen, H., Truong, K. Q. D., & Van Vo, T. (2013). Analysis of oxy-Hb signals to determine relationship between jaw imbalance and arm strength using fNIRS. American Journal of Biomedical Engineering, 3, 107–118.Google Scholar
- 4.Hai, N. T., Cuong, N. Q., Khoa, T. Q. D., & Toi, V. V. (2013). Temporal hemodynamic classification of two hands tapping using functional near infrared spectroscopy. Frontiers in Human Neuroscience, 7, Article 516, 1–12.Google Scholar
- 5.Molteni, E., Baselli, G., Bianchi, A. M., Caffini, M., Contini, D., Spinelli, L. et al. (2009). Frontal brain activation during a working memory task: A time-domain fNIRS study. Photonic Therapeutics and Diagnostics V, 71613 N. https://doi.org/10.1117/12.808972.
- 6.Sassaroli, A., Zheng, F., Coutts, M., Hirshfield, L. H., Girouard, A., Solovey, E. T., et al. (2009). Application of near-infrared spectroscopy for discrimination of mental workloads. In Proceedings of SPIE 7174, Optical Tomography and Spectroscopy of Tissue VIII (pp. 71741H).Google Scholar
- 8.Suthaharan, S. (2016). Support vector machine. In Machine learning models and algorithms for big data classification (pp. 207–235). Springer.Google Scholar
- 9.Chu, F., & Zaniolo, C. (2004). Fast and light boosting for adaptive mining of data streams. In Pacific-Asia Conference on Knowledge Discovery and Data Mining (pp. 282–292). Springer.Google Scholar
- 11.LeCun, Y., Kavukcuoglu, K., & Farabet, C. (2010). Convolutional networks and applications in vision. In Proceedings of the IEEE International Symposium on Circuits and Systems (pp 253–226).Google Scholar
- 12.Krizhevsky, A., Sutskever, I., & Hinton, G. (2012). ImageNet classification with deep convolutional neural networks. In Proceedings of NIPS (pp. 1097–1105).Google Scholar
- 13.Tsinalis, O., Matthews, P. M., Guo, Y., & Zafeiriou, S. (2016). Automatic sleep stage scoring with single-channel EEG using convolutional neural networks. arXiv:1610.01683.
- 15.Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. CoRR. arXiv:1409.1556.
- 16.Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., & LeCun, Y. (2014). Overfeat: Integrated recognition, localization and detection using convolutional networks. In ICLR.Google Scholar
- 17.Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 580–587).Google Scholar