Over the years, Puncho communicated words on a computer using a pointer attached to a bench cap, a difficult way that allowed him to write about five true words per minute.

He emailed me, “I had to tilt my head forward / down, and pause the key letter one by one to write.”

Last year, researchers gave it another device associated with a head-controlled mouse, but in research sessions it is not as fast as a brain electrode.

Through the electrodes, Pancho communicated 15 to 18 words per minute. This was the maximum rate that allowed the study, as the computer asks in between. Dr. Chang says fast decoding is possible, although it’s not clear if it will reach the typical conversational speech speed: about 150 words per minute. Speed ​​is the main reason why projects focus on speaking, tapping directly into the brain’s word production system, rather than hand movements to get involved in typing or writing.

“This is the most natural way for people to communicate,” he said.

Puncho’s enthusiastic personality has helped researchers explore challenges, but also occasionally makes speech recognition uneven.

He emailed, “Sometimes I can’t control my emotions and laugh a lot and don’t do much better in the experiment.

Dr. Chang recalled a time when the algorithm successfully recognized a sentence, after which, “you could see him shaking his eyes and it felt like he was doing some kind of gesture.” It happened when, during repetitive tasks, he would yawn or get distracted, “It didn’t work very well, because he didn’t really focus on getting those words. So, some of us have to work. Things have gotten because we clearly want it to work all the time.

The algorithm is sometimes confused with words with similar phonetic sounds, “bring” as “go”, “do” as “you”, and words starting with “f” – “faith,” “family,” “feeling”. – Word as V, “too.”