Precision Psychiatry Arrives

Why Is Biomarker-Guided Diagnosis a Breakthrough for Depression Care?

Major depressive disorder (MDD), bipolar disorder and schizophrenia can present with overlapping symptoms in early stages, yet the appropriate pharmacological treatments differ sharply. Antidepressant monotherapy in undiagnosed bipolar disorder, for example, can precipitate manic switching or rapid cycling. The World Health Organization estimates that more than 280 million people live with depression globally, and a substantial proportion of those initially diagnosed with unipolar depression are later reclassified, with bipolar disorder often taking years to identify correctly.

Near-infrared spectroscopy (NIRS) measures changes in blood oxygenation in the prefrontal cortex during cognitive tasks. Researchers in Japan have studied prefrontal hemodynamic patterns as a candidate biomarker to help distinguish these conditions, and the country's Pharmaceuticals and Medical Devices Agency (PMDA) had previously authorised NIRS-based decision support for differential diagnosis under specific clinical conditions. The newly approved AI-augmented device builds on that foundation, applying machine-learning analysis to the same underlying signal to support clinicians at the point of decision-making.

How Could Brain Imaging Change Antidepressant Selection?

The landmark STAR*D trial, sponsored by the U.S. National Institute of Mental Health, found that only about a third of patients with major depression achieved remission after a first-line selective serotonin reuptake inhibitor, and remission rates fell with each subsequent treatment step. Treatment-resistant depression — typically defined as inadequate response to at least two adequate antidepressant trials — affects an estimated 20 to 30 percent of patients with MDD, contributing disproportionately to disability, suicide risk and healthcare costs.

Imaging-based and electrophysiological biomarkers have been investigated for years as potential predictors of antidepressant response. Functional MRI studies have linked specific patterns of subgenual cingulate and amygdala activity to differential response to cognitive behavioural therapy versus pharmacotherapy. Combining such signals with AI analysis is the central premise of precision psychiatry: matching the patient to the treatment most likely to work, rather than cycling through medications empirically. Regulatory acceptance of biomarker-based devices in major markets is a prerequisite for that vision.

What Does Japan's Approval Mean for Other Regulators and Patients?

The U.S. Food and Drug Administration has authorised several AI-based software-as-a-medical-device tools across radiology, cardiology and ophthalmology, and has issued guidance on machine-learning-enabled devices. Psychiatric indications have lagged behind, in part because objective ground-truth labels are harder to define and external validation across ethnic and demographic groups is essential. Devices that perform well in a single national population may not generalise without further study.

For patients, the practical implication is that psychiatric diagnosis is moving — slowly — toward the kind of objective testing common in other branches of medicine. Clinicians caution that no current biomarker replaces a thorough clinical assessment; rather, these tools are intended as decision support. The field is also watching parallel developments in pharmacogenomic testing, blood-based inflammatory markers and digital phenotyping, all of which aim to bring depression care closer to the precision-medicine standards already established in oncology.