A Unified Analytical Framework With Multiple fNIRS Features for Mental Workload Assessment in the Prefrontal Cortex

Area of Science:

• Neuroscience
• Biomedical Engineering

Background:

• Accurate mental workload assessment is crucial for effective brain-computer interface (BCI) cognitive training.
• Limited signal extraction from brain regions may not reflect true cognitive states.

Purpose of the Study:

• To develop an analytical framework using functional near-infrared spectroscopy (fNIRS) for precise mental workload assessment in the prefrontal cortex (PFC).
• To introduce and evaluate novel features for enhanced signal analysis.

Main Methods:

• Utilized a multi-channel, multi-distance fNIRS device to measure brain activity.
• Introduced 'deep contribution ratio' alongside conventional features like hemodynamic slope.
• Employed a linear support vector machine (SVM) classifier with the number of activated channels as input.

Main Results:

• The combined features (hemodynamic slope and deep contribution ratio) achieved 80.6% accuracy in classifying mental workload levels.
• This significantly outperformed conventional features alone, which yielded 59.8% accuracy.
• The analytical framework effectively suppressed false detection rates.

Conclusions:

• The proposed analytical framework with multiple features demonstrates feasibility for accurate mental workload assessment.
• This advancement holds promise for improving the efficacy of fNIRS-based BCI applications in cognitive training.