AI software that "visualizes" depression diagnosis has been developed—Ministry of Health, Labour and Welfare grants the first domestic pharmaceutical approval

Table of Contents

1. Background and Significance of Approval

2. Mechanism: Calculating "Depression Score" with fMRI × AI

3. Development History: Over 10 Years of Data Collection and Learning

4. Process to Regulatory Approval and Future Schedule

5. Three Impacts on Mental Health Care

6. Operational Image in Clinical Settings

7. Remaining Challenges: Accuracy, Bias, Ethics, Privacy

8. Comparison with Overseas Trends and Japan's Uniqueness

9. Future Prospects: Insurance Coverage, Disease Expansion, Drug Development Collaboration

10. Conclusion: Towards the Integration of AI and Mental Health DX

1. Background and Significance of Approval

In recent years, there has been a rapid increase in "SaMD (Software as a Medical Device)" using AI and digital biomarkers. Particularly in the field of mental disorders, the lack of objective indicators has been a long-standing issue. The software approved by the Ministry of Health, Labour and Welfare is Japan's first medical device that visualizes depression diagnosis in numerical form, symbolizing the digital transformation of psychiatric care.mainichi.jpmainichi.jp

2. Mechanism: Calculating "Depression Score" with fMRI × AI

• Additional 10-minute Imaging: Adds 10 minutes of task-based fMRI to a standard MRI exam.

• AI Analysis: Analyzes brain activity patterns using convolutional neural networks, learning differences from normal brains.

• Output: Displays a "Depression Score" from 0 to 100 and estimated probability. Clinical trials showed approximately 70% diagnostic accuracy.plus-web3.com

3. Development History: Over 10 Years of Data Collection and Learning

A team led by Mitsuo Kawato (ATR Brain Information Communication Research Laboratory Group) collected fMRI data from 944 depression patients and 2,031 healthy individuals between 2008 and 2018. They used a vast amount of brain activity maps as training data to extract multivariate features representing symptom severity. The collaboration of university hospitals and 52 psychiatric clinics nationwide, covering diverse cases, was also a key factor in high evaluation.plus-web3.com

4. Process to Regulatory Approval and Future Schedule

5. Three Impacts on Mental Health Care

1. Standardization of Diagnosis: Provides objective indicators for cases where evaluations differ among doctors.

2. Treatment Monitoring: Visualizes the effects of pharmacotherapy and CBT through score transitions.

3. Patient Empowerment: Supports awareness and treatment continuation by sharing numerical data.

6. Operational Image in Clinical Settings

• Target Patients: Differential diagnosis at first visit, treatment-resistant cases, remission confirmation, etc.

• Workflow: MRI appointment → Additional task imaging → AI analysis (about 15 minutes) → Doctor explains results in consultation room.

• Cost: Estimated at around 10,000 yen per test after insurance coverage, with an approximate 3,000 yen out-of-pocket cost for psychiatric outpatient visits.

7. Remaining Challenges: Accuracy, Bias, Ethics, Privacy

• 70% Accuracy Issue: Diagnosis solely by AI is not possible. Must be used in conjunction with interviews and DSM-5 evaluation.

• Data Bias: The focus on Japanese data may lead to performance decline in overseas deployment.

• Accountability: Development of XAI (Explainable AI) is urgent to avoid black-boxing.

• Privacy: Brain activity data is the ultimate personal information, requiring rules for secondary use under the revised Personal Information Protection Law.

8. Comparison with Overseas Trends and Japan's Uniqueness

In the U.S., an AI-based PTSD screening tool received FDA approval in 2024, and digital mental health devices are advancing in insurance reimbursement in Europe. Japan's strengths lie in its high-quality universal health insurance and the world's largest number of MRI installations. This technology, utilizing existing MRI, holds advantages in cost and implementation.facebook.com

9. Future Prospects: Insurance Coverage, Disease Expansion, Drug Development Collaboration

• Insurance Coverage: Accelerates dissemination through rapid entry into evaluation medical care.

• Disease Expansion: Ongoing research for application to bipolar disorder, schizophrenia, and anxiety disorders.

• Drug Development Support: Expected to be used as an objective biomarker in clinical trial endpoints.

10. Conclusion: Towards the Integration of AI and Mental Health DX

This approval marks the first step in transforming psychiatric care, which has relied on subjective evaluations, into a data-driven approach. The key to the future will be whether we can build an ecosystem where patients, healthcare providers, and society can benefit by enhancing accuracy and establishing ethical frameworks.

List of Reference Articles

• Mainichi Shimbun "Quantifying Depression with AI Analysis: Diagnostic Support Software Approved by Ministry of Health" (June 30, 2025)mainichi.jp

• PlusWeb3 "AI Diagnostic Support Software for Depression Approved by Ministry of Health: Numerical Analysis of fMRI Images" (July 1, 2025)plus-web3.com

• Mainichi Shimbun (Print Edition) "AI Analysis for Depression Diagnosis: Quantification Software Approved" (July 1, 2025 Morning Edition)mainichi.jp

• Japan Association of Mental Health Social Workers Official X Post "Quantifying Depression with AI Analysis" (July 1, 2025)x.com

• FDA Press Release “FDA Authorizes First AI-Based PTSD Screening Tool” (2024)