Artificial intelligence modelling human mental fatigue: A comprehensive survey

Mental fatigue refers to the decline in cognitive abilities that can occur as a result of prolonged mental exertion. Neuroscientists have been studying mental fatigue for a while. They clearly understand some underlying mechanisms of mental fatigue, such as brain chemistry and neural activity changes. However, defining mental fatigue is still an open research question. Despite this, neuroscience and cognitive psychology has made significant progress in understanding the causes and consequences of mental fatigue. In contrast, computer scientists presumably have a limited understanding of mental fatigue. This lack of understanding leads to inadequate models of mental fatigue in computer science. However, the ever evolving field of artificial intelligence (AI) shows a great potential to answer the open challenges in mental fatigue modelling. For instance, AI with fuzzy rules, machine learning or deep learning algorithms, as well as the methods of model explanation can be a valuable tool for creating accurate models of mental fatigue. Artificial intelligence models can learn from large amounts of data and accurately predict mental fatigue. However, modelling efforts are often more focused on acquiring parameters than studying and validating them within the model. Models developed in this way suffer problems with reliability making it challenging to understand the underlying causes of mental fatigue. Therefore, it is essential balance between these parameters’ correct acquisition, validation and interaction to create more accurate models of mental fatigue. In our survey, we have observed that an unspecified modelling impact the model at four scale : experimental design, data acquisition and processing, choice of indicators or parameters and reasoning. We provide some useful guidelines through criticisms for modelling, which would be closer to reality.