Classifying visual fatigue severity based on neurophysiological signals and psychophysiological ratings

Abstract

This paper presents a neuroergonomics study of visual fatigue associated with visual display tasks. Symptoms of visual fatigue may include tiredness, headaches, eye soreness, eye aches, discomfort when the eyes are open, difficulties in focusing, and blurred vision. These symptoms can be caused by demands on the visual functions, such as focusing and converging of the eyes. A feed forward artificial neural network (FF-ANN) is used to classify visual fatigue based on the subjects' neurophysiological signals and psychophysiological ratings on a simulated sickness questionnaire (SSQ) scale. Inclusive of all data, a classification accuracy of 83.33 % was obtained (with 70% for training, 15% for validating, and 15% for testing from nine subjects), 90.4% accuracy was obtained using eye response data, 86.78% accuracy with eye and EEG response data, and, 84.93% with eye and hemodynamics response data. The FF-ANN model classifies about 16.7 % and 33.33% of severe and moderate ratings based on medium to difficult simulated air traffic control tasks on the SSQ scale.