Deep Learning Techniques in Neuroergonomics

Abstract

There is increasing interest in using deep learning (DL) for neuroergonomics research that investigates the human brain in relation to behavioral performance in natural environments and everyday settings. But a better understanding of how to design and implement DL techniques is still needed for neuroergonomists. Written for novice neuroergonomists as well as experienced investigators, this chapter presents the history of advancements in DL, its concepts, and applications of DL in neuroergonomics research. In addition to artificial neural network (ANN) which is a basic model for DL, this chapter introduces popular DL models such as the multilayer perceptron (MLP), deep belief network (DBN), convolutional neural network (CNN), and recurrent neural networks (RNN). DL-based neuroergonomics research on four main research areas (i.e., mental workload, motor imagery, driving safety, and emotion recognition) will then be reviewed. Insights into how to model and apply DL techniques will be helpful for neuroergonomics researchers, in particular those who are not familiar with DL, but want to predict and classify brain states under various contexts.