Identifying ADHD boys by very-low frequency prefrontal fNIRS fluctuations during a rhythmic mental arithmetic task
Sergio Ortuño-Miró1, Sergio Molina-Rodríguez2, Carlos Belmonte2
1Department of physiology, Miguel Hernandez University, San Joan d´Alacant, Alicante, Spain.
Journal of Neural Engineering
|May 23, 2023
Summary
Functional near-infrared spectroscopy (fNIRS) effectively identifies attention-deficit/hyperactivity disorder (ADHD) in boys. Machine learning models accurately distinguished ADHD from controls using specific brain activity patterns.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Clinical Psychology
Background:
- Computer-aided diagnosis aids ADHD assessment, but neuroimaging translation faces barriers.
- Few studies use functional near-infrared spectroscopy (fNIRS) for individual ADHD discrimination.
- Developing feasible and explainable fNIRS methods is crucial for clinical application.
Purpose of the Study:
- To develop an fNIRS-based method for identifying ADHD in boys.
- To utilize technically feasible and explainable machine learning approaches.
- To identify objective neuroimaging biomarkers for ADHD.
Main Methods:
- Collected fNIRS data from ADHD boys and controls during a mental arithmetic task.
- Computed time-frequency synchronization measures to identify group-specific oscillatory patterns.
- Employed machine learning models with feature selection for binary classification.
Main Results:
- Logistic regression and linear discriminant analysis achieved near 100% accuracy, sensitivity, and specificity.
- Classification was highly accurate using only three key features from very-low frequency oscillations.
- Statistical significance was confirmed via resampling procedures (p<.001).
Conclusions:
- Very-low frequency fNIRS fluctuations can accurately differentiate ADHD boys from controls.
- The proposed fNIRS approach shows promise for reliable and interpretable functional biomarkers.
- This method could potentially inform ADHD clinical practice with objective indicators.
