Hemispheric, a startup founded by Gidi Littwin, a former Apple engineer known for developing FaceID, has announced a significant $52 million funding round. The company uses artificial intelligence to analyze electrical brain activity and diagnose disorders such as depression, PTSD, and Parkinson's disease. The goal is to make neural diagnosis as simple and affordable as a blood test.
Gidi Littwin's background and the birth of Hemispheric
Littwin left Apple in 2020 after contributing to key projects like FaceID and hand tracking for the Vision Pro. During his tenure, he collected hundreds of thousands of data points to train deep learning models. The turning point came when Hagai Lalazar, co-founder of Hemispheric, contacted him via LinkedIn. Lalazar had already developed an idea for studying the brain without surgery but was looking for a commercially minded co-founder. After speaking with about 75 candidates, he found the right person in Littwin.
Sponsored Protocol
How the diagnosis works: EEG headset and artificial intelligence
The Hemispheric system requires the patient to wear a lightweight EEG headset for about 15 minutes while interacting with an app on a tablet. The headset measures electrical brain activity, which is then analyzed by an advanced AI model. This model, described as a "frontier" model, interprets signals similarly to how large language models deduce meaning from text. This allows clinicians to diagnose disorders, select therapeutic interventions, and monitor progress with greater accuracy than traditional subjective questionnaires.
The data collected and the frontier model for the brain
To train the model, Hemispheric gathered a quarter-million hours of brain data from 100,000 paid volunteers across Asia, Tel Aviv, and Boston. Volunteers performed game-like activities designed to activate different brain regions. This dataset, described by Littwin as the "most prized possession," enabled the creation of a generalizable model. Subsequent tests on patients with PTSD, schizophrenia, and depression showed that the model can make accurate deductions about individuals' brain health. The team is currently working on a clinical study to test the model's ability to diagnose and even predict Alzheimer's disease.
Sponsored Protocol
Future prospects and FDA approval
Hemispheric plans to submit its first product, focused on PTSD diagnosis, to the FDA early next year. If approved, it could be made available to the public by 2027. The raised funds, from American and Israeli venture capital firms including early Uber backer Howard Morgan, will be used to expand partnerships with governments, healthcare organizations, and pharmaceutical companies, as well as to hire staff in the United States. The startup is also developing proprietary brain scanners, believed to be more suitable for machine learning than traditional EEGs. For a comparison with other innovations, see the article on PsiQuantum, another startup aiming to revolutionize its field. More information on EEG technology can be found on Wikipedia.
Sponsored Protocol
