FaceID veteran raises $52M for brain diagnostics

Hemispheric bets big on brain data

Gidi Littwin, a co-inventor of Apple’s FaceID and Vision Pro technology, has spent the past six years building Hemispheric, a startup aiming to decode brain activity with deep learning. The company has now raised $52 million after collecting data from 100,000 people’s brains.

According to WIRED, the goal is to examine the brain without invasive procedures and eventually help diagnose cognitive disorders.

From Apple hardware to brain health

Littwin left Apple in 2020 after being contacted on LinkedIn by Hagai Lalazar, who had started developing technology to study the brain without surgery and wanted a commercially minded cofounder. Lalazar had spoken to around 75 candidates before finding him.

At Apple, Littwin helped develop FaceID and later worked on hand-tracking for Vision Pro. He told WIRED that those projects required “hundreds of thousands of subjects' worth of data” to train the underlying deep learning models.

“There were massive data collection operations behind these projects and we knew we had to build something very similar at Hemispheric,” Littwin says, “and we have.”

A quarter-million hours of brain recordings

Diagnosing conditions such as depression, Alzheimer’s, and Parkinson’s has often depended on questionnaires and behavioral observation because brain activity varies widely from person to person.

Recommended reading

Mira Murati’s lab unveils Inkling, a 975B open model

To build a broader model, Littwin and Lalazar collected what the company calls its “most prized possession”: a quarter of a million hours of brain data from 100,000 paid volunteers across Asia, as well as Tel Aviv and Boston. Participants completed activities designed like games that activated different brain regions.

Hemispheric says that dataset was used to train a frontier model that infers brain function from electrical activity inside the skull, which WIRED describes as analogous to how large language models statistically analyze text.

The company then tested the generalized model on groups including people diagnosed with PTSD, schizophrenia, and depression, and said it made accurate deductions about their brain health. The team is also working on a clinical study to test whether the model can diagnose and predict Alzheimer’s.

First FDA submission planned next year

Hemispheric’s first product is focused on PTSD. The startup plans to submit it to the FDA for approval early next year, with hopes of making it publicly available later in 2027.

The process is relatively simple: a patient wears a lightweight EEG headset for about 15 minutes while using an app on a tablet. Hemispheric says its model can then help clinicians:

• decode brain signals for diagnosis
• choose the most effective intervention by predicting treatment response
• monitor patient progress

“The future that we envision is one where this is akin to a blood test,” Lalazar says. “The device is going to be very, very cheap; it will be able to be sold and distributed throughout mental health clinics, hospitals, and even psychologists' offices.”

More money, more data, more hardware

The company’s early-stage backers include American and Israeli venture capital firms and individual investors such as early Uber backer Howard Morgan. Hemispheric says it will use the funding to build partnerships with governments, health care organizations, and pharmaceutical firms, expand hiring in the US, and push toward regulatory approval.

It also plans to gather brain data from millions of people to improve the model further.

At the same time, the founders are developing their own brain scanners, arguing that conventional EEG systems are not well suited to modern model training.

“These devices were never built for machine learning and definitely not deep learning,” Littwin says.