Prediction of dispositional dialectical thinking from resting-state electroencephalography

Area of Science:

• Neuroscience
• Cognitive Psychology
• Computational Psychiatry

Background:

• Dialectical thinking, a complex cognitive process, has been challenging to quantify objectively.
• Resting-state electroencephalography (EEG) offers a non-invasive window into brain activity.
• Previous research has explored neural correlates of cognitive functions, but direct prediction of dispositional dialectical thinking using EEG is less explored.

Purpose of the Study:

• To investigate the feasibility of predicting dispositional dialectical thinking levels from resting-state EEG signals.
• To identify specific EEG frequency bands and brain regions associated with dialectical thinking.
• To develop and validate machine learning models for this prediction task.

Main Methods:

• Collected resting-state EEG data from 34 participants who also completed a dialectical thinking self-report measure.
• Processed EEG signals, including wave filtration and artifact removal, and analyzed time-frequency representations across delta, theta, alpha, and beta bands.
• Employed functional principal component analysis for feature reduction and applied ensemble machine learning models (SVR, LASSO, KNN, RF, GBDT) for prediction.

Main Results:

• The alpha wave (7-13 Hz) in the early resting period (12-15 s) was the most significant predictor of dialectical thinking.
• A data-driven electrode selection approach (FC1, FCz, Fz, FC3, Cz, AFz) yielded a prediction model with an average R-squared of 0.45.
• A positive correlation was observed between alpha activity in the right dorsal anterior cingulate cortex and self-reported dialectical thinking scores.

Conclusions:

• Resting-state prefrontal and midline alpha oscillations are effective predictors of dispositional dialectical thinking.
• These findings suggest the involvement of specific brain structures, particularly the anterior cingulate cortex, in supporting dialectical thinking.
• The study demonstrates the potential of EEG-based machine learning for assessing higher-order cognitive traits.