The new brain interface explains the inner monologues with amazing accuracy

Scientists can now decode the brain activity related to silent internal monologue in people’s heads with a resolution of up to 74 %, according to a new study.

In a new research published today in cell, Scientists from Stanford University deciphered the imagined words from four participants with severe paralysis due to Als or stroke in the brain. Aside from being very wild, the results can help people who cannot speak to communicate more easily using computer interfaces in the brain (BCIS), the researchers say.

“This is the first time that we have been able to understand the form of brain activity when you only think about speaking,” said lead author Irene Konz, a student of high studies in Electrical Engineering at Stanford University. “For people with speech disabilities and intense mobility, BCIS can help them to decipher the interior speech in communicating more easily and more natural.”

Previously, scientists managed to decipher the speech using BCIS. When people physically try to speak loudly by engaging the muscles related to speech, these techniques can explain the resulting brain activity and write what they are trying to say. But despite its effectiveness, the current methods of communication with the help of BCI are still stressful for people with limited muscles. The new study is the first to take an internal speech.

To do this, the researchers recorded the activity in the kinetic cortex – the area responsible for controlling volunteer movements, including words – using micro -electrodes planted in the kinetic cortex of the four participants.

The researchers found that the attempt to speak that tried and imagined is similar patterns, although not identical, for brain activity. They trained the AI model to explain these imagined speech signals, so the camel was one of the vocabulary of 125,000 words with a resolution of up to 74 %. In some cases, the system chose unproductive internal ideas, such as the participants in the numbers that were silently calculated during the mission.

For people who want to use new technology but do not always want their internal ideas in the full explosion, the team added a mechanism to control the password that prevented BCI from deciphering internal speech unless the participants in the password (“Chitty Chitty Bang Bang” in this case) added. The system is known for the password with more than 98 %.

While the resolution of 74 % is high, the current technology still makes a large amount of errors. But researchers hope to do so, the most sensitive recording devices and the best algorithms can enhance their performance more.

“The Future of the Bcis is bright.” “This work gives a real hope that BCIS talk can restore communication one day, natural, and comfortable like conversation speech.”