Scientists say a new brain-computer interface allows users to transmit 62 words per minute

Scientists say a new brain-computer interface allows users to transmit 62 words per minute

A team of Stanford scientists claim to have tested a new brain-computer interface (BCI) that can decode speech at up to 62 words per minute, beating the previous record by 3.4 times.

That would be a big step towards real-time speech conversion at the pace of natural human conversation.

Max Hodak, who co-founded the BCI company Neuralink with Elon Musk but was not involved with the study, called the research “a significant step in the usefulness of implanted BCIs” in an email to Futurism.

As detailed in a yet to be peer-reviewed article, the team of Stanford scientists found that they only needed to analyze brain activity in a relatively small region of the cortex to convert it into coherent speech using a machine learning algorithm.

The aim was to give back the voice of those who can no longer speak or caress due to ALS. While keyboard-based solutions have allowed paralyzed people to communicate again to some degree, a brain-based voice interface could greatly speed up decoding.

“Here we demonstrated a speech BCI capable of decoding unrestricted sentences from a large vocabulary at speeds of 62 words per minute, the first time a BCI has matched the communication rates that alternative technologies for people with paralysis, such as B. eyes, has far surpassed tracking,” the researchers write.

In one experiment, the team recorded the neural activity of an ALS patient, who can move his mouth but has trouble forming words, from two small areas in his brain.

Using a recurrent neural network decoder capable of predicting text, the researchers then converted those signals into words – and at a surprisingly fast rate.

They found that analysis of these orofacial movements and the associated neural activity “was probably strong enough to support speech BCI, despite paralysis and tight coverage of the cortical surface,” according to the study.

But the system wasn’t perfect. The error rate of the researchers’ recurrent neural network (RNN) decoder was still around 20 percent.

“Our demonstration is a proof of concept that decoding attempted speech movements from intracortical recordings is a promising approach, but it is not yet a complete, clinically feasible system,” the researchers acknowledged in their work.

To improve their system’s error rate, the scientists propose examining more areas of the brain while optimizing the algorithm.

More on BCIs: Neuralink co-founder says new company’s eye implant could be “ultimate” VR technology

Leave a Reply

Your email address will not be published. Required fields are marked *