Discriminative Analysis of Brain Functional Connectivity Patterns for Mental Fatigue Classification
- 670 Downloads
Mental fatigue is a commonly experienced state that can be induced by placing heavy demands on cognitive systems. This often leads to lowered productivity and increased safety risks. In this study, we developed a functional-connectivity based mental fatigue monitoring method. Twenty-six subjects underwent a 20-min mentally demanding test of sustained attention with high-resolution EEG monitoring. Functional connectivity patterns were obtained on the cortical surface via source localization of cortical activities in the first and last 5-min quartiles of the experiment. Multivariate pattern analysis was then adopted to extract the highly discriminative functional connectivity information. The algorithm used in the present study demonstrated an overall accuracy of 81.5% (p < 0.0001) for fatigue classification through leave-one-out cross validation. Moreover, we found that the most discriminative connectivity features were located in or across middle frontal gyrus and several motor areas, in agreement with the important role that these cortical regions play in the maintenance of sustained attention. This work therefore demonstrates the feasibility of a functional-connectivity-based mental fatigue assessment method, opening up a new avenue for modeling natural brain dynamics under different mental states. Our method has potential applications in several domains, including traffic and industrial safety.
KeywordsCross-validation Electroencephalography (EEG) Partial directed coherence (PDC) Permutation Psychomotor vigilance test (PVT) Multivariate pattern analysis (MVPA)
The authors thank the National University of Singapore for supporting the Cognitive Engineering Group at the Singapore Institute for Neurotechnology (SINAPSE) under Grant Number R-719-001-102-232. This study was also supported by the NEUROEN Grant R3940000059232. We acknowledge the assistance of Sheralyn Tan and Ong How Hwee in EEG data collection and Jie Fu in data processing.
- 2.Astolfi, L., F. de Vico Fallani, F. Cincotti, D. Mattia, M. G. Marciani, S. Bufalari, S. Salinari, A. Colosimo, L. Ding, J. C. Edgar, W. Heller, G. A. Miller, B. He, and F. Babiloni. Imaging functional brain connectivity patterns from high-resolution EEG and fMRI via graph theory. Psychophysiology 44:880–893, 2007.PubMedCrossRefGoogle Scholar
- 3.Babiloni, F., F. Cincotti, C. Babiloni, F. Carducci, D. Mattia, L. Astolfi, A. Basilisco, P. M. Rossini, L. Ding, Y. Ni, J. Cheng, K. Christine, J. Sweeney, and B. He. Estimation of the cortical functional connectivity with the multimodal integration of high-resolution EEG and fMRI data by directed transfer function. Neuroimage 24:118–131, 2005.PubMedCrossRefGoogle Scholar
- 5.Bishop, C. M. Pattern Recognition and Machine Learning. New York: Springer, 2006, 326 pp.Google Scholar
- 6.Borghini, G., L. Astolfi, G. Vecchiato, D. Mattia, and F. Babiloni. Measuring neurophysiological signals in aircraft pilots and car drivers for the assessment of mental workload, fatigue and drowsiness. Neurosci. Biobehav. Rev. 2012. doi: 10.1016/j.neubiorev.2012.10.003.
- 8.Chang, C. C., and C. J. Lin. LIBSVM: a library for support vector machines. ACM Trans. Intell. Syst. Tech. (TIST) 2:27, 2011.Google Scholar
- 16.Dockree, P. M., S. P. Kelly, J. J. Foxe, R. B. Reilly, and I. H. Robertson. Optimal sustained attention is linked to the spectral content of background EEG activity: greater ongoing tonic alpha (approximately 10 Hz) power supports successful phasic goal activation. Eur. J. Neurosci. 25:900–907, 2007.PubMedCrossRefGoogle Scholar
- 24.Jung, T. P., C. Humphries, T. W. Lee, S. Makeig, M. J. McKeown, V. Iragui, and T. J. Sejnowski. Extended ICA removes artifacts from electroencephalographic recordings. Adv. Neural Inf. Process. Syst. 10:894–900, 1998.Google Scholar
- 27.Kendall, M. G. Rank Correlation Methods. New York: Oxford University Press, 1948.Google Scholar
- 38.Lin, C. T., Y. C. Chen, T. Y. Huang, T. T. Chiu, L. W. Ko, S. F. Liang, H. Y. Hsieh, S. H. Hsu, and J. R. Duann. Development of wireless brain computer interface with embedded multitask scheduling and its application on real-time driver’s Drowsiness detection and warning. IEEE Trans. Biomed. Eng. 55:1582–1591, 2008.PubMedCrossRefGoogle Scholar
- 40.Liu, F., W. Guo, J. P. Fouche, Y. Wang, W. Wang, J. Ding, L. Zeng, C. Qiu, Q. Gong, W. Zhang, and H. Chen. Multivariate classification of social anxiety disorder using whole brain functional connectivity. Brain Struct. Funct. 2013. doi: 10.1007/s00429-013-0641-4.
- 43.Maroco, J., D. Silva, A. Rodrigues, M. Guerreiro, I. Santana, and A. de Mendonca. Data mining methods in the prediction of dementia: a real-data comparison of the accuracy, sensitivity and specificity of linear discriminant analysis, logistic regression, neural networks, support vector machines, classification trees and random forests. BMC Res. Notes 4:299, 2011.PubMedCentralPubMedCrossRefGoogle Scholar
- 45.Ojala, M., and G. C. Garriga. Permutation tests for studying classifier performance. J. Mach. Learn. Res. 99:1833–1863, 2010.Google Scholar
- 50.Sturm, W., A. de Simone, B. J. Krause, K. Specht, V. Hesselmann, I. Radermacher, H. Herzog, L. Tellmann, H. W. Muller-Gartner, and K. Willmes. Functional anatomy of intrinsic alertness: evidence for a Fronto-Parietal-Thalamic-brainstem network in the right hemisphere. Neuropsychologia 37:797–805, 1999.PubMedCrossRefGoogle Scholar