Development of a cerebellar ataxia diagnosis model using conditional GAN-based synthetic data generation for visuomotor adaptation task

• Published in: BMC medical informatics and decision making Vol. 24; no. 1; pp. 336 - 15
• Main Authors: Kim, Jinah, Woo, Sung-Ho, Kim, Taekyung, Yoon, Won Tae, Shin, Jung Hwan, Lee, Jee-Young, Ryu, Jeh-Kwang
• Format: Journal Article
• Language: English
• Published: England BioMed Central Ltd 12.11.2024; BioMed Central; BMC
• Subjects: Accuracy; Adaptation; Adaptation, Physiological; Adult; Age; Ataxia; Atrophy; Brain; Cerebellar ataxia; Cerebellar Ataxia - diagnosis; Cerebellum; Classification; Comparative analysis; Conditional generative adversarial network; Deep Learning; Diagnosis; Digital healthcare; Female; Health aspects; Humans; Machine learning; Male; Medical diagnosis; Medical research; Middle Aged; Movement disorders; Neural networks; Neural Networks, Computer; Patients; Psychomotor Performance - physiology; Random noise; Risk factors; Sensorimotor integration; Synthetic data; Virtual reality; Visuomotor adaptation task; Young adults; South Korea; Visuomotor adaptation task; Conditional generative adversarial network; Cerebellar ataxia diagnosis; Digital healthcare; Synthetic data
• ISSN: 1472-6947; 1472-6947
• DOI: 10.1186/s12911-024-02720-y

This study proposes a synthetic data generation model to create a classification framework for cerebellar ataxia patients using trajectory data from the visuomotor adaptation task. The classification objectives include patients with cerebellar ataxia, age-matched normal individuals, and young healthy subjects. Synthetic data for the three classes is generated based on class conditions and random noise by leveraging a combination of conditional adversarial generative neural networks and reconstruction networks. This synthetic data, alongside real data, is utilized as training data for the patient classification model to enhance classification accuracy. The fidelity of the synthetic data is assessed visually to measure the validity and diversity of the generated data qualitatively while quantitatively evaluating distribution similarity to real data. Furthermore, the clinical efficacy of the patient classification model employing synthetic data is demonstrated by showcasing improved classification accuracy through a comparative analysis between results obtained using solely real data and those obtained when both real and synthetic data are utilized. This methodological approach holds promise in addressing data insufficiency in the digital healthcare domain, employing deep learning methodologies, and developing early disease diagnosis tools.