UCSF Researchers Synthesize Speech The Brain Waves

Researchers led by language neuroscientist Edward Chang at the University of California San Francisco (UCSF) on Tuesday reported their achievement at decoding speech efforts in real time by studying the action from the speech centers of test subjects’ brains.

Three persons of speech, who were being treated for epilepsy participated in the analysis.

They allowed the researchers to use tiny recording electrodes that were put to map the origins of the seizures.

The technique, known as”electrocorticography,” or ECoG, provides richer and more detailed information about brain activity compared to technology such as EEG or fMRI.

Since ECoG electrodes don’t penetrate the brain tissue, they are a better choice for long-term brain-computer interfaces (BCIs) compared to electrodes physically placed into the brain.

Chang is a part of the Weill Institute for Neurosciences at UCSF.

The Inner Workings of the Experiment

The team of chang gave a choice of 24 possible answers to them and asked the participants nine easy questions.

The group developed machine learning algorithms to decode speech sounds from the test participants’ brain activity.

The algorithms could discover when participants heard a question or starting to identify which participants gave, and to react, together with as much as 61 percent accuracy.

The algorithms’ speed and accuracy improved when they used the test subjects’ brain activity to identify to be able to offer context which they’d heard. That method given up to 75 percent accuracy.

The technology can help people”who’ve lost the ability to talk but still have cognitive capacity that would enable understanding and the capacity to form speech-related thoughts,” noted Rob Enderle, principal analyst at the Enderle Group.

However, training the calculations”could be quite resource-intensive,” he told TechNewsWorld. “Generally that instruction would be part of finding the system related to the person’s brain.”

Funding From Facebook

The study program is overseen by UCSF and works with volunteer subjects, although facebook scientists have provided input and technology support to Chang’s laboratory. The Facebook scientists have limited access to data, which remain on site under the control of the university and at UCSF.

That research aims to ascertain if neural interface implants may be used to restore communication and motion skills to patients paralyzed because of nerve or brain difficulties.

FRL is financing part of the BRAVO study”which is focused solely on demonstrating the capacity to enable the participant to create text on a monitor with their brain activity,” the spokesperson told TechNewsWorld. “Our understanding is that UCSF’s BRAVO analysis is wider in scope.”

Facebook would like to utilize the technology for reality glasses.

Privacy and Ethical Issues

Facebook isn’t known for its adherence. The FTC fined the firm US$5 billion for breaching users’ privacy, a punishment criticized as insufficient.

Facebook is a part of the Information Transfer Project with Google Apple and Twitter, an attempt to create systems to transfer information between providers.

Facebook monitors people ubiquitously on the internet though its”Like” button, which can be used by over 8 million sites, and its FB software development kits, which can be embedded in over 60 percent each of the best iOS and Android programs.

That raises the question of what issues arise with research and product development, and whether the information gleaned from research is safe.

About the author

Related

JOIN THE DISCUSSION