Evaluation of divided attention using different stimulation models in event-related potentials

Area of Science:

• Cognitive Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Divided attention is a significant societal challenge, often assessed via questionnaires or physiological signals.
• Electroencephalography (EEG) and evoked potentials (EPs) are key physiological measures for studying attention.
• Existing research often lacks focus on healthy, high-attention populations.

Purpose of the Study:

• To differentiate between sustained and divided attention states using physiological signals.
• To investigate the effectiveness of EEG and EP analysis for attention classification.
• To explore classification methods in a healthy, high-attention demographic.

Main Methods:

• Collected EEG and EP data from 48 university students (18-25 years old) under visual, auditory, and combined stimuli.
• Applied Fourier-based filtering (0.01-30 Hz) and extracted features like fractal dimensions, entropy, power spectral densities, Hjorth parameters, and P300 components.
• Reduced feature set size by eliminating less informative features.
• Utilized a support vector machine with a linear kernel for classification.

Main Results:

• Achieved a maximum classification accuracy of 88.89% in distinguishing between selective and divided attention.
• Identified specific EEG and EP features that effectively differentiate attention states.
• Demonstrated that divided attention can be successfully classified, though it may be more challenging than selective attention.

Conclusions:

• Accurate classification of divided attention is achievable using advanced signal processing and machine learning techniques on EEG and EP data.
• The study provides insights into the neural underpinnings of attention in a healthy, high-attention cohort.
• Findings suggest potential for objective, physiological assessment of attention states.