Magnetic resonance eye tracking

Scientists routinely monitor the way our brains work during various cognitive tasks with functional magnetic resonance imaging (fMRI), which reveals the brain areas that are working the hardest.

In many of these tests, researchers need to know where a subject is looking, which normally requires the use of eye-tracking equipment that bounces an infrared beam off the eye to determine where it is pointing.

This is expensive, difficult to set up in the narrow confines of an fMRI machine, and also time consuming. So Xiaoping Hu and colleagues at Emory University in Atlanta, US, have developed software that can detect eye orientation using the fMRI scanner alone.

With the subject staring at a calibration target, the software learns to identify the eye orientation from the fMRI data and is then able to work out future eye movements with reference to this, by identifying matching signals in the fMRI data.

Read the full magnetic resonance eye tracking patent application.

Justin Mullins, New Scientist consultant