Neuroscientists at the University of Toronto have developed a new technique for the first time to recreate images using EEG of what people perceive based on the brain activity. The technique is developed by Dan Nemrodov who is a postdoctoral fellow in Assistant Professor Adrian Nestor's lab at U of T Scarborough. The study involved showing test subjects images of faces while they were hooked to EEG equipment. The brain activity was then used to recreate images using machine learning algorithms.
EEG captures activity at millisecond scale so that we are able to see very fine detail of how test subjects develops images. According to researchers, it takes our brain around 170 milliseconds to form a good representation of the face we observe. A lot of researchers doubted the possibility of such image reconstruction due to its limitation until the study affirmed that EEG does have the potential. The results have huge theoretical and practical prospects from a neurotechnological viewpoint. After this study, testing future prospects is already underway in the Nestor's lab. They are testing on how image reconstruction build on EEG data can be applied to other objects as well. It can most certainly have wide-ranging clinical applications too. The study could also help in providing people who are unable to verbally communicate a possible means for expressing themselves.
By: Neha Maheshwari