An article was published stating that it was possible to accurately read human thoughts by scanning the brain using a technique called fMRI and interpreting it with artificial intelligence. Technology is expected to significantly advance technologies that support communication for people who are unable to speak or move due to illness or injury.
Until now there have been technologies that scan the brain and read thoughts, but most of them use invasive devices that require electrodes or implants to be implanted in the head via surgery. The research team of the University of Texas at Austin Neuroscience announced in an article published on September 29 (bioRxiv) that they were able to read thoughts non-invasively using fMRI.
Originally, fMRI is a technology that reads brain activity by matching electromagnetic waves and measuring brain blood flow. For this reason, there is no need to embed electrodes in the head, but compared to a device that reads neuronal signals directly, it takes time to interpret and lacks real-time performance. However, the research team solved this problem by reading the meaning of high-level thinking in the brain with an AI model using machine learning instead of deciphering every word the subject was thinking about.
In developing this technology, the research team first had three people, two men and a woman in their 20s and 30s, listening to an audio drama or podcast while using the fMRI device. A decoder, an algorithm that interprets incidents from fMRI data, was created by learning from the fMRI scan data measured for 16 hours in several steps and from the original sentence. If the decoder thus created re-read the subject’s brain data, it was possible to reproduce the story heard by the subject quite accurately. Sure, there are errors in the details, but it is possible to accurately express a word or phrase and get the point.
Researchers believe the technology could help support communication in people who cannot speak or move. This is because fMRI technology is non-invasive and does not require surgery to implant electrodes in the brain. Of course, fMRI devices cannot be purchased or used by individuals on a large scale, but according to the research team, the possibility exists that this technology could be applied to MEG (magnetoencephalography) using more portable devices than fMRI.
However, a decoder made by learning a person’s thought data can only be used to decode that person’s thought and cannot be used at all for other people. This means that you won’t be able to read your thoughts unless you participate in an extensive training session at least 16 hours in advance. The researchers also studied reading inaccuracies by asking the scanned subjects to think about something else or to count. As a result, imagining animals is said to be the most effective way to prevent detoxification. Relative content this placecan be registered