A neural network to classify fatigue from human–computer interaction

Fatigue, especially in its mental form, is one of the most worrying health problems nowadays. It affects not only health but also motivation, emotions and feelings and has an impact both at the individual and organizational level. Fatigue monitoring and management assumes thus, in this century, an increased importance, that should be promoted by private organizations and governments alike. While traditional approaches are mostly based on questionnaires, in this paper we present an alternative one that relies on the observation of the individual's interaction with the computer. We show that this interaction changes with the onset of fatigue and that these changes are significant enough to support the training of a neural network that can classify mental fatigue in real time. The main outcome of this work is the development of non-invasive systems for the continuous classification of mental fatigue that can support effective and efficient fatigue management initiatives, especially in the context of desk jobs.