Ed Young of Not Exactly Rocket Science has a new post, about a machine developed by researchers at Berkeley that tells what you’re seeing from your brain activity. More specifically, the device built by Kendrick Kay and coworkers uses fMRI readings of the visual cortex to identify which image amongst 120 readings the test subject is viewing. As noted, it’s accuracy is not great: 92% and 82% accuracy for Kay and fellow researcher and test subject Thomas Naselaris, respectively, when 13 fMRI readings are averaged; and 51% and 32% for Kay and Naselaris when using only a single reading. Raising the number of images to 1000, the success rate for Kay’s brain dropped to 82% from the 92% of 120 images.
While there is still a long way to go to a real-time brain reader, this is still a big step.
Tags: Brain Scan, fMRI, Mind reading, Neurology, Neuroscience, Science
