Q-MIND: enhancing ADHD diagnosis using quantum machine learning for advanced neuroimaging analysis.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Computational Psychiatry

Background:

• Attention Deficit Hyperactivity Disorder (ADHD) is a prevalent neurobehavioural disorder in children and adolescents, characterized by inattention, hyperactivity, and impulsivity.
• Accurate and timely ADHD diagnosis is challenging due to the disorder's complexity and lack of standardized testing, leading to a high number of undiagnosed cases.
• Integrating neuroimaging and phenotypic data offers potential for improved diagnostic accuracy.

Purpose of the Study:

• To develop a data-driven framework for enhancing clinical decision-making in ADHD diagnosis.
• To leverage quantum machine learning and evolutionary algorithms for improved diagnostic precision.
• To create a robust and scalable diagnostic tool for ADHD.

Main Methods:

• Utilized the ADHD-200 dataset, comprising neuroimaging and phenotypic data.
• Employed a Quantum Convolutional Neural Network (QCNN) for feature extraction from MRI and behavioral data.
• Introduced a Differential Evolution-Swarm Optimization (DE-Swarm) algorithm for feature selection and an AutoML system for model optimization.

Main Results:

• The Gradient Boosting Classifier, optimized through the proposed framework, achieved a test accuracy of 98.53%.
• The approach demonstrated superior precision, recall, and specificity compared to existing methods.
• The integrated framework effectively processed high-dimensional data, identifying nuanced patterns for diagnosis.

Conclusions:

• The study presents a robust and scalable framework for ADHD diagnosis by integrating quantum computing principles and evolutionary strategies.
• Combining phenotypic and neuroimaging data with advanced machine learning significantly enhances diagnostic precision.
• This approach supports more personalized clinical assessments for individuals with ADHD.