Japan Transforms Brain Signals into Images

Japanese specialists have successfully developed, for the first time in the world, a technique that allows the use of generative artificial intelligence to create images based on processing signals from the human brain. According to Kyodo News, "during the experiment, a group of researchers from the National Institute of Advanced Industrial Science and Technology recorded brain signals while some participants viewed images of objects and landscapes. Approximately 1,200 different images were presented to the participants in the experiment. During this process, brain activity was analyzed using functional magnetic resonance imaging (fMRI). The images themselves were uploaded to an artificial intelligence database, which was tasked with learning to associate them with brain signals." Previous studies had already shown that it is possible to reconstruct images viewed by humans based on brain signals processed using fMRI. However, it was found that only a limited range of images can be reproduced this way. Now, based on previous achievements, Japanese scientists have developed a program that converts brain signals into digital values, which the already trained artificial intelligence uses to reconstruct images. The closest to reality was an image of a tiger with clearly drawn mouth and ears, along with its spotted coat color. Scientists believe this technique could be used in developing communication devices, as well as in understanding the mechanisms of human hallucinations and dreams.