A “brain-computer interface” clinic, and what that really means
In December, the Shanghai Mental Health Center opened what it calls China’s first psychiatric brain-computer interface clinic. The name sounds like science fiction, but there is no implant. Patients get brain scans (MRI, EEG and a light-based scan called fNIRS), engineers analyze the scans to find personal stimulation targets, and then patients receive an intensive course of TMS aimed at those targets: 30 to 50 sessions over 6 to 8 days.
The clinic treats a range of conditions. For someone with schizophrenia, the doctors describe aiming stimulation at the brain circuits linked to auditory hallucinations, and adjusting targets if depression or obsessive-compulsive symptoms are also present. Demand was immediate: every initial appointment was booked within 10 minutes of the announcement.
Here’s what families should know: this is a description of a new service, not a study. The doctors who run it write that they have no data yet on how many patients were treated, whether it helped, or side effects, and that they “cannot draw definitive conclusions about its safety or effectiveness.” Scan-guided TMS is a promising idea, and it’s worth watching. For schizophrenia, it is still unproven.
The wearable nobody talks about for serious mental illness
When we talk about wearables and serious mental illness, we usually mean sleep and mood trackers. A small UK study points somewhere else: the glucose sensor. Researchers led by the University of Sheffield offered a two-week continuous glucose monitor to adults living with serious mental illness and type 2 diabetes. Sixteen of 21 said yes, and they wore it an average of 11 of the 14 days.
Participants found the sensor far easier than finger-prick testing, and no adverse events were reported. The biggest problem was practical: sensors sometimes fell off after a knock. It’s an early study, and the sensor didn’t show people their readings, because the question was simply whether they’d wear one.
Why it matters for our families: people with serious mental illness are two to three times more likely to develop type 2 diabetes, and they lose an estimated 15 to 20 years of life, mostly to preventable physical illness. If your loved one has diabetes, it’s worth asking their doctor whether a continuous glucose monitor is an option.
AI that listens for early signs of psychosis
Researchers at NTU Singapore and Singapore’s Institute of Mental Health built an AI tool that looks for early signs of psychosis in everyday speech. Young people aged 14 to 29 spoke for 45 seconds about something neutral, like why they deserve a “most interesting person” award. They never had to describe a symptom.
The tool reads emotional cues in the transcript, such as fear, sadness, joy and trust. It identified up to 93% of the young people at high risk, with 67% overall accuracy, across 429 participants. Standard AI chatbots asked to do the same job performed no better than chance.
This is research, not something your family can use yet. The team says it needs more testing in the real world, and the published figures don’t say how many young people were wrongly flagged. But it’s a hopeful counterpoint to what we’re learning about chatbots and first-episode psychosis: the same young people, the same vulnerable window, and a very different kind of AI.
Read the visit notes, especially now
More doctors are using AI “scribes” that listen to the appointment and write the visit note. A new analysis of more than 5,700 real consultations found that, without safety checks, about 3% of note content wasn’t supported by anything said in the room. About 2.9% was content that could change a clinical decision. Safety checks cut those numbers roughly in half. The authors work for the company that makes the scribe.
One error type matters especially in psychiatry: the scribe can take a diagnosis the doctor was only thinking out loud about and record it as fact. A spoken “we should rule out schizophrenia” can become a diagnosis in the chart, and it follows a person to every future appointment.
If your loved one shares their patient portal with you, read the visit notes after each appointment, using the same high-risk checks doctors are being told to use: medications, allergies, diagnoses and the plan. If something doesn’t match what was said, ask the care team about it. Patients have the right to request a correction to their medical record.
ResourcesforSMI. Come join the fun
