For years, Pancho communicated by spelling words on a computer using a pointer mounted on a baseball cap, an arduous method that allowed him to type about five correct words per minute.
“I had to bow/tilt my head forward, crouch and poke each letter one by one to write,” he emailed.
Last year, researchers gave him another device that involved a head-controlled rat, but it was still not as fast as the brain electrodes during the study sessions.
Through the electrodes, Pancho conveys 15 to 18 words per minute. That’s the maximum rate the study allows because the computer waits between prompts. Dr Chang says it can be deciphered faster, though it’s unclear if it matches the speed of typical conversational speech: about 150 words per minute. Speed is the main reason the project focuses on speaking, tapping directly into the brain’s word production system rather than the hand movements involved in typing or writing.
“It’s the most natural way for people to communicate,” he says.
Pancho’s bubbly personality helped the researchers overcome challenges, but also sometimes made speech recognition uneven.
“I can’t control my emotions sometimes and laugh a lot and don’t do too well with testing,” he emailed.
Dr. Chang recalls times, after the algorithm successfully identified a sentence, “you could see he was clearly shaking and it looked like he was giggling.” When that happens or when, during repetitive tasks, he yawns or gets distracted, “it didn’t work very well because he wasn’t really focused on hearing those words. So we have some work to do because we obviously want it to work all the time. “
Algorithms sometimes confuse words with similar sounds, identifying “go” as “carry”, “do” as “you” and words starting with “F” – “faith”, “family” , “feel” – is the V -word, “very.”