Apple FaceID Co-Inventor's AI Startup Aims to Decode Brain Health With $52M in Funding

The co-creator behind Apple's FaceID and Vision Pro hand-tracking technology has spent the last six years developing a frontier artificial intelligence system designed to interpret electrical activity in the human brain, with the goal of diagnosing cognitive disorders without invasive procedures.

Gidi Littwin's startup, Hemispheric, has now secured $52 million in funding after assembling a dataset covering 100,000 individuals' brains to train its deep-learning models.

From FaceID to Brain Decoding

Littwin departed Apple in 2020 seeking a new challenge. He found it through a LinkedIn message from Hagai Lalazar, who would become his Hemispheric cofounder. Lalazar had already begun developing AI to study the brain without surgery and was searching for a commercially driven partner to propel the company forward — a search that had led him through approximately 75 candidates before connecting with Littwin.

At Apple, Littwin had contributed to FaceID and was later working on hand-tracking for the Vision Pro augmented reality headset. That work required collecting what he described as "hundreds of thousands of subjects' worth of data" to train the deep learning models powering those consumer technologies.

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

Training AI on 100,000 Brains

Because every person's brain activity manifests differently, medical professionals have long depended on subjective questionnaires and behavioral observations to identify conditions such as depression, Alzheimer's, and Parkinson's disease. To overcome this limitation, Littwin and Lalazar gathered what they consider their most valuable asset: 250,000 hours of brain data collected from 100,000 paid volunteers located across Asia, Tel Aviv, and Boston. Participants completed a series of tasks resembling games that stimulated different regions of the brain.

This dataset was used to train a frontier AI model that infers brain function from electrical activity within the skull, operating on principles similar to how large language models extract meaning through statistical analysis of text. The team subsequently tested the generalized model on specific subgroups, including individuals diagnosed with PTSD, schizophrenia, and depression, and reported that it produced accurate assessments of those individuals' brain health. A clinical study is currently underway to evaluate whether the model can diagnose and potentially predict Alzheimer's disease.

Toward FDA Approval and Public Rollout

Hemispheric's first product, intended for studying PTSD, is slated for submission to the U.S. Food and Drug Administration early next year. The company aims to make the product available to the public sometime in 2027.

The diagnostic process involves a patient wearing a lightweight EEG headset that records electrical brain activity for approximately 15 minutes while engaging with an application on a tablet. Hemispheric states that its AI model then assists clinicians in interpreting the signals to formulate diagnoses, predict the most effective treatment options, and track patient progress over time.

Lalazar envisions a future where the technology functions much like a routine blood test. "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," he said.

Building Better Brain Scanners

AI-assisted diagnostic tools are not entirely new to medicine — similar systems for detecting conditions such as lung cancer are already in clinical use across Europe, helping accelerate access to treatment. At the same time, major AI companies including OpenAI and Anthropic are pushing deeper into the healthcare sector, increasing competitive pressure on the growing field of health-focused startups.

Hemispheric's $52 million in early-stage funding came from American and Israeli venture capital firms along with individual investors, among them Howard Morgan, an early backer of Uber. The company plans to allocate the capital toward forging partnerships with governments, healthcare organizations, and pharmaceutical companies; expanding its U.S. workforce; pursuing regulatory approval; and collecting brain data from millions more individuals to further refine its model.

The team is also developing proprietary brain scanners designed to capture information that the company believes will prove more useful for its models than conventional EEG devices. As Littwin noted, existing EEG equipment was never engineered with machine learning or deep learning applications in mind.

As Hemispheric pushes toward clinical validation and regulatory clearance, its approach could represent a significant shift in how cognitive and mental health conditions are identified and treated. If the vision holds, a quick, non-invasive brain scan could one day be as routine as checking your blood pressure. Found this article interesting? Share it with your network and join the conversation about the future of AI-driven healthcare.

Source: Wired