Investigation of the neural correlation with task performance and its effect on cognitive load level classification

Area of Science:

• Neuroscience
• Cognitive Psychology
• Machine Learning

Background:

• Electroencephalogram (EEG)-based cognitive load assessment is crucial in psychological research.
• Conventional assumptions link increased task difficulty directly to higher cognitive workload.
• This study challenges the universal applicability of this conventional hypothesis.

Purpose of the Study:

• To investigate whether cognitive load is solely determined by task conditions or also influenced by participant performance.
• To explore the relationship between individual performance during tasks and measured cognitive load.
• To develop more accurate, subject-independent models for cognitive load assessment.

Main Methods:

• Collected EEG data from 36 participants during resting and mental arithmetic task conditions.
• Utilized Empirical Mode Decomposition (EMD) for feature extraction from EEG signals.
• Employed Support Vector Machine (SVM) for classification of cognitive load states.

Main Results:

• Demonstrated that cognitive load can vary based on participant performance, not just task demands.
• Hypotheses regarding the impact of subject performance on cognitive load were formulated and statistically validated.
• Classification accuracy provided evidence for the performance-dependent nature of cognitive load.

Conclusions:

• The conventional hypothesis that cognitive workload solely increases with task difficulty is not always correct.
• Participant performance is a significant factor in determining cognitive load.
• Findings support the development of more appropriate, subject-independent machine learning models for cognitive load assessment.