MRI Mind Reading: Reconstucting Visual Images, Even Dreams, Straight from a Brain Scan

This is fascinating stuff: researchers at the University of California at Berkeley are already testing out a form of MRI scan that can reliably discern what sorts of images people are looking at when the scan is performed. They are claiming that with refinement, this technology may be able to reconstruct entire visual images just from brain scans alone, perhaps even including the non-sensory images in people’s vivid memories or dreams (We still aren’t positive that these latter images map to the visual cortex in quite the same way as direct sensory images, but there’s plenty of tantalizing circumstantial evidence that they do).

I’m a little skeptical of the idea of reconstructing new images without pre-observing them: the way the brain actually maps out different details in an image is probably not anywhere near as simple as a sort of linear pixel map. The “data” might be heavily layered or interpreted in ways that make it difficult to reconstruct, just as optical illusions produce experiences of visual effects that aren’t there on the paper itself. The scientists may currently be observing predictable inputs to the system and getting reliable hits off of that, but that’s no guarantee that the particular sorts of things they are looking at can easily translate right back into normal images again. The best that might be possible would be statistical generalizations (i.e. we can tell that he’s looking at something square, or something blue, but can’t reproduce the exact contours and fine details of what it is).

Still, very cool, and I can’t wait to see if more really is possible.

I’m not sure I really see the potential privacy worries mentioned though: the technique is essentially just a proof of concept that visual perceptions are mapped onto the brain in a mechanical fashion. Considering that a subject would have to be strapped into a huge MRI machine in order to do a scan, and that the scan could only reproduce what the patient was currently looking at, this seems more like the most expensive way imaginable to look at a Polaroid rather than getting into “thought-crime” territory. The machine doesn’t observe thoughts, per se: it merely observes what the brain is observing (or, perhaps, envisioning).

And it would be utterly wondrous if scientists could piece together even brief, cloudy snapshots of the images in people’s dreams: allowing others to finally catch a glimpse of what has always been a subjective and isolated experience.

Disclaimer: take this all with a skeptical grain of salt. It’s science journalism after all.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: