Artificial Intelligence-Based Diagnostic Support System for Functional Dyspepsia Based on Brain Activity and Food...

Area of Science:

• Neuroscience
• Gastroenterology
• Artificial Intelligence

Background:

• Disorders of Gut-Brain Interaction (DGBI), including functional dyspepsia (FD) and irritable bowel syndrome (IBS), lack detectable organic abnormalities.
• Brain activity patterns in individuals with DGBI differ significantly from those in healthy controls.
• Artificial intelligence (AI) shows promise in identifying biomarkers for DGBI diagnosis.

Purpose of the Study:

• To develop an AI-based diagnostic support system for DGBI.
• To integrate food preferences and prefrontal cortex brain activity data for diagnosis.
• To assess the diagnostic accuracy of an AI model in differentiating DGBI patients from healthy controls.

Main Methods:

• Patients diagnosed with FD and IBS using ROME IV criteria.
• Food preferences assessed via visual analog scale.
• Brain activity measured using functional near-infrared spectroscopy (fNIRS).
• An artificial neural network model developed using fNIRS and food preference data.

Main Results:

• The study included 41 participants (25 with DGBI).
• The AI model achieved 72.3% accuracy in distinguishing DGBI patients from healthy controls.
• The AI model achieved 77.1% accuracy in differentiating FD patients from healthy controls.

Conclusions:

• AI models integrating brain activity and food preferences demonstrate significant accuracy in identifying DGBI.
• Functional near-infrared spectroscopy (fNIRS) offers objective evidence for DGBI diagnosis.
• This approach supports the development of novel diagnostic tools for functional gastrointestinal disorders.