Depression is a prevalent mental health disorder affecting over 300 million people globally. Currently, the diagnosis of depression worldwide, including in Taiwan, relies on symptom-based criteria listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM). This process involves patient cooperation through interviews to articulate their current mental state, heavily relying on the physician's assessment to understand the patient's psychopathology. However, due to the historical stigma surrounding mental illness, patients often feel uncomfortable and reluctant, sometimes even concealing their symptoms, leading to diagnostic challenges.

To address these issues, Professor Cheng-Ta Li’s team at Taipei Veterans General Hospital has developed the "Depression-AI PreDIC^®" This system utilizes three different types of brain imaging to examine patients' brain functions, combined with artificial intelligence (AI) to objectively assess and classify the severity of depression. It also predicts the effectiveness of repetitive transcranial magnetic stimulation (rTMS) therapy, marking a revolutionary breakthrough in depression treatment.

Objective assessment of depression severity and classification

The "Depression-AI PreDIC^®," developed by Professor Cheng-Ta Li's team, enhances traditional scale assessments by incorporating three different types of brain imaging to examine brain function. These include:

Electroencephalography (EEG): To detect abnormalities in brain wave spectra.

Functional Magnetic Resonance Imaging (fMRI): To evaluate abnormalities in neural connectivity.

Positron Emission Tomography (PET): To assess metabolic conditions of neural cells.

By integrating these imaging modalities with AI, the system provides an objective understanding of depression severity. AI analysis of brain imaging data enables the classification and grading of depression, allowing physicians to offer more precise treatments tailored to each subtype of depression.

Predicting rTMS treatment efficacy and accelerating recovery

Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive treatment for depression, but its effectiveness varies among individuals. Historically, clinicians have observed abnormalities in depression patients through various brain imaging tools, yet subsequent treatment plans often lacked coherence. Even when abnormalities were detected, personalized treatment was not always arranged.

The research team has collaborated with the Precision Psychiatry Center at Taipei Veterans General Hospital to continuously collect data, validate, and optimize the process. Using linear and non-linear EEG analyses combined with specialized AI analytical models, they developed a pioneering technique that can predict rTMS treatment outcomes through EEG before treatment begins. This technique optimizes treatment plans, significantly reduces operational complexity, and decreases interpretation time to just two minutes. In the future, physicians will be able to quickly determine the most effective treatment method for patients, expediting their recovery from depression.

Professor Cheng-Ta Li expressed the hope that the development of this system will help more depression patients receive more effective treatment. The team will continue to collect data, validate, and optimize processes, aiming to extend this innovative approach to the diagnosis and treatment of other mental disorders in the future.

Resource (Mandarin):

《新創動態》「A/腦影像」評估憂鬱症 客觀分級預測治療效果