Meta introduces Brain2Qwerty v2 AI, which can convert brain waves into text in real-time without surgery. This technology can help people who are unable to speak. Its accuracy is approaching that of surgical methods.
New Delhi [भारत]June 30 (ANI): Meta has unveiled Brain2Qwerty v2. It is a new artificial intelligence (AI) system that can convert brain activity into text without any surgical implants. This is being considered an important step in brain-computer interface research.
According to the company, Brain2Qwerty v2 is the best end-to-end system to date that can decode sentences in real-time from non-invasive brain recordings.
Meta said the technology is reaching levels of accuracy that were previously only possible with techniques like brain surgery. “Brain2Qwerty v2 is the most advanced end-to-end system for decoding sentences in real-time from non-invasive brain recordings, approaching levels of accuracy previously only possible with techniques requiring brain surgery,” the company said.
There can be hope for millions of people
The company said this research could help millions of people suffering from brain lesions and other conditions that leave them unable to communicate.
This is in contrast to invasive (surgical) methods such as stereotactic electroencephalography and electrocorticography, which require surgical implants to capture brain signals. In contrast, Brain2Qwerty uses non-invasive recording to decode text.
How does this technology work?
Meta said it trained Brain2Qwerty v2 using approximately 22,000 sentences collected from nine volunteer participants. Each participant wore a magnetoencephalography (MEG) device for approximately 10 hours while actively typing.
The company explained that the system uses end-to-end deep learning to decode language directly from raw brain signals, rather than relying on manually designed pipelines to identify neural events.
According to Meta, large language models were fine-tuned on neural data, enabling the system to use semantic context to bridge the gap between noisy brain signals and coherent language.
The company also deployed AI agents to explore optimizations in the decoding process, with the final training configuration selected by engineers.
as accurate as surgical methods
Meta said the new system achieved a word accuracy rate of 61 percent, a significant improvement compared to the word accuracy of 8 percent achieved by other non-invasive methods.
For the best-performing participant in the study, the system achieved a word accuracy rate of 78 percent, with more than half of all decoded sentences having one word error or less.
The company also found that decoding accuracy improved as more training data became available. This suggests that the performance gap between non-invasive and surgical methods can be further reduced through larger datasets.
Research will get a boost, code will be released
To support further research, Meta has announced that it is releasing the full training code for Brain2Qwerty v1 and v2. Its research partner, the Basque Center on Cognition, Brain and Language (BCBL), will also release the Brain2Qwerty v1 dataset.
Meta said the work is part of its broader efforts to develop open foundational models of the brain and advance neuroscience research aimed at improving the diagnosis, treatment and understanding of neurological disorders.
Meta also said that this research is part of its broader effort to create an open foundational model of the brain.
The company highlighted its TribeV2 model for encoding perception, NeuralSet for processing large-scale brain data, and NeuralBench for systematically evaluating brain models.
Meta said these initiatives are part of its ‘Digital Brain Project’, under which it recently announced a US$5 million fund to help create open neuroscience datasets. (ANI)
(Except for the headline, this story has not been edited by Asianetnews Editorial staff and is published from a syndicated feed.)