2nd Place Winner BCI Award 2021 - A direct speech BCI for decoding words and sentences скачать в хорошем качестве

2nd Place Winner BCI Award 2021 - A direct speech BCI for decoding words and sentences

David A. Moses*,1,2, Sean L. Metzger*,1,2,5, Jessie R. Liu*,1,2,5, Gopala K. Anumanchipalli1,2, Joseph G. Makin1,2, Pengfei F. Sun1,2, Josh Chartier1,2, Maximilian E. Dougherty1, Patricia M. Liu3, Gary M. Abrams4, Adelyn Tu-Chan4, Karunesh Ganguly2,4, Edward F. Chang1,2,5 A direct-speech BCI for decoding words and sentences in a person with severe paralysis and anarthria Speech is the most effortless, expressive, and efficient method of communication for most individuals. For people with severe paralysis who have also lost t he ability to speak, neurotechnology designed to restore communication has the potential to drastically improve their autonomy and quality of life. Remarkable progress has been made with existing brain-computer interfaces (BCIs) for communication restoration, enabling users to type or write intended messages. A complementary and potentially faster approach may be to directly decode speech from brain areas that typically control the vocal tract. Previous efforts have shown that a paralyzed person could use an implanted two-channel microelectrode device and an audiovisual interface to generate some vowel sounds and phonemes, but word-level decoding was not demonstrated. In this work, which was conducted as part of a larger clinical trial, we used signals acquired from a 128-channel, high-density electrocorticography (ECoG) array implanted over the surface of the speech sensorimotor cortex of a brainstem stroke survivor with severe paralysis and anarthria (the inability to articulate speech) to determine if full words and sentences could be decoded from brain activity during attempts to speak. 1 Dept. of Neurological Surgery 2 Weill Institute for Neurosciences 3 Dept. of Rehab. Services 4 Dept. of Neurology 5 University of California, Berkeley, CA, USA. https://www.bci-award.com