An exploratory analysis of longitudinal artificial intelligence for cognitive fatigue detection using...

Area of Science:

• Neuroscience
• Artificial Intelligence
• Wearable Technology

Background:

• Cognitive fatigue, stemming from stress, impairs attention and performance.
• Real-world monitoring of cognitive fatigue is vital for managing breaks and performance.
• Existing methods often rely on complex biosignals (EEG, EOG) and AI for pattern recognition.

Purpose of the Study:

• To introduce the Exploratory Analysis of Longitudinal Artificial Intelligence for Cognitive Fatigue Detection Using Neurophysiological Based Biosignal Data (EALAI-CFDNBD) approach.
• To detect cognitive fatigue using neurophysiological biosignal data.
• To develop more effective and less intrusive wearable devices for cognitive fatigue tracking.

Main Methods:

• Utilized linear scaling normalization (LSN) for input feature scaling.
• Employed binary olympiad optimization algorithm (BOOA) for feature selection.
• Used a graph convolutional autoencoder (GCA) classifier for fatigue detection.
• Applied multi-objective hippopotamus optimization (MOHO) for hyperparameter tuning.

Main Results:

• The EALAI-CFDNBD model demonstrated high accuracy in detecting cognitive fatigue.
• Experimental validation on the MEFAR dataset yielded a superior accuracy of 97.59%.
• The approach effectively reduced data dimensionality and optimized model hyperparameters.

Conclusions:

• The EALAI-CFDNBD approach shows significant promise for accurate cognitive fatigue detection.
• This AI-driven method can enhance the development of wearable devices for fatigue monitoring.
• The findings have implications for high-demand environments where cognitive performance is critical.