Scientists at Meta have developed Brain2Qwerty v2, a system that can convert brain activity into text without requiring brain implants or surgery. Instead, it relies on a non-invasive helmet-like device that detects the brain’s magnetic signals as a person thinks about typing.
The technology uses magnetoencephalography (MEG), a brain imaging technique that records magnetic fields generated by neural activity. The system interprets imagined typing or mental letter selection and converts it into words on a screen, enabling thought-based text input.
Nature Neuroscience Highlights Brain-to-Text Breakthrough
The research was published in the journal Nature Neuroscience, where scientists described how the original Brain2Qwerty system successfully decoded sentences directly from brain signals without physically interacting with the brain.
The latest version, Brain2Qwerty v2, significantly improves performance by using artificial intelligence to process raw brain data automatically and produce text in near real time. This replaces the slower, manually guided decoding methods used in earlier versions.
To train the system, nine healthy volunteers spent approximately ten hours typing sentences while wearing the MEG helmet. During these sessions, the AI learned to associate specific patterns of brain activity with individual letters. It was also combined with a language model, similar to predictive text or smartphone autocorrect, to improve accuracy and reduce typing errors.
Testing showed a major improvement in performance. The system accurately decoded 61% of words on average, with accuracy reaching 78% in the best-performing cases. By comparison, earlier non-invasive brain-to-text methods achieved only about 8% accuracy, highlighting the significant progress made by the new technology.
Currently, helping people who have lost the ability to communicate often requires surgically implanting chips into the brain. The dataset, developed with the Basque Center on Cognition, Brain and Language, is now available for global research and development.
Helmet-Based System Could Restore Communication for People with Paralysis and ALS
A non-invasive solution that only requires wearing a helmet could offer a safer and more accessible alternative. It has the potential to help people living with conditions such as amyotrophic lateral sclerosis (ALS), severe stroke, or other forms of paralysis regain the ability to communicate with loved ones and express their needs, even if they have lost the ability to speak or move.
For now, the biggest limitation is the hardware. The brain-reading equipment is still bulky, expensive, and comparable in size to hospital imaging machines. However, researchers believe advances in sensor technology will make these devices much smaller over time, with the long-term goal of integrating the system into an affordable, wearable cap or similar head-mounted device.
To help speed up research in the field, Meta has made the Brain2Qwerty training code publicly available. Developed with the Basque Center on Cognition, Brain and Language, the dataset is now available for global research and further development.
Developers, scientists, and anyone interested can access technical documentation, demonstrations, and downloadable resources through the project’s official website. The project hub and related files are available at address
Read the original article on: pagina3
Read more:Mushrooms causing tiny human hallucinations contain no psychedelics