Utilizing Phase Locking Value to Determine Neurofeedback Treatment Responsiveness in Attention Deficit Hyperactivity...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Neurofeedback is a non-invasive brain training technique for hyperactivity disorder, but treatment effectiveness varies significantly among individuals.
• Many patients are irresponsive to neurofeedback, highlighting the need for predictive methods to assess treatment efficacy before initiation.
• Previous studies on predicting neurofeedback success, particularly using slow cortical potentials, have shown limited accuracy.

Purpose of the Study:

• To explore functional brain connections across EEG frequency bands to predict neurofeedback treatment effectiveness.
• To develop and validate an algorithm for predicting the success of neurofeedback therapy for hyperactivity disorder.
• To identify optimal brain electrodes for neurofeedback treatment to enhance efficiency.

Main Methods:

• Utilized EEG data from 60 hyperactive students (aged 7-14) undergoing neurofeedback.
• Applied a five-step algorithm involving signal preprocessing, extraction of alpha and beta bands, and phase lock value analysis to assess functional brain lobe connectivity.
• Employed machine learning classifiers, including support vector machine and boosting methods, after feature selection using t-tests and genetic algorithms to identify optimal electrodes (C3, FZ, F4, CZ, C4, F3).

Main Results:

• The proposed algorithm achieved a 90.6% accuracy in predicting the treatability of individuals with hyperactivity.
• Effectively identified key electrodes for neurofeedback treatment, reducing the number from 32 to 6.
• Demonstrated the ability to predict treatment outcomes rapidly using limited data.

Conclusions:

• Introduced a highly accurate algorithm (90.6%) for predicting neurofeedback treatment outcomes in hyperactivity disorder.
• Significantly enhances treatment efficiency by identifying optimal electrodes and reducing the number required.
• Enables prediction of patient responsiveness to neurofeedback therapy, conserving time and financial resources by avoiding unnecessary sessions.