“Pay attention!” That is often very good advice, but sometimes the advice is hard to obey.   The brain’s limited attentional resources can be overwhelmed when attention has to be distributed among multiple objects.  And the challenge is even greater when the objects are moving.  For example, imagine that you’re driving on Route 128 at rush hour.  You must attend not only to your own car’s path, but also to the whims and surprising behaviors of the cars all around you.   Working in my lab, Heather Sternshein and Yigal Agam, two PhD students in the Neuroscience Program, developed a novel electroencephalographic (EEG) technique to study how selective attention is apportioned in a task that can be described as “Route 128-on Sterioids.”   We were especially interested in the neural correlates of failures of attention, the kind of failure that, on the road, might have serious consequences.

Subjects in our experiment watched as ten identical black discs moved about randomly on a computer display for ten seconds.  The hard part was to keep track the entire time of particular, pre-designated target discs –either three, four or five.  Because all ten moving discs were identical, there were no physical features to distinguish target from non-target discs.  At the end of eight seconds, all discs came to standstill, and a subject tried to identify the discs that he or she had been tracking.  The task required attentive tracking of a subset of identical multiple moving objects, something even more challenging than navigating Route 128 at rush hour.

Every once in a while during the eight-second tracking period, one of the ten discs flashed brightly for 100 msec.  Sometimes, the flashed disc was a target disc, that is, one the subject was trying to track; sometimes the flashed disc was a non-target disc, that is, one that the subject could be ignoring.   The flash evoked a response in the subject’s brain, and our EEG system picked up that response from the subject’s scalp.  Knowing that the evoked response would be larger if the flash were delivered to an object that was being attended,  we used responses to target and non-targets as an index of how attention was distributed among the multiple moving objects.  We focused our analysis on electrodes located over occipital and parietal lobes, toward the brain’s posterior.

As expected, the relative sizes of responses to the two kinds of stimuli differed: on average, flashes on target discs evoked larger responses than flashes on non-target discs.  This difference confirmed that on average subjects were paying more attention to targets than to non-targets. But as the number of discs that had to be tracked increased —from three to four to five– subjects found the task increasingly harder, and made more errors when they had to identify the discs that they had been trying to track.  The EEG revealed the neural correlate of these failures of attention.  The difference between evoked responses to flashed targets and flashed non-targets decreased as the number of targets increased.  This shrinking difference between the two sets of neural responses could explain the systematic increase in errors as the the number of targets increased.  As additional items have to be kept track of, it becomes harder for subjects to apportion attentional resources in a way that preserves a sufficient advantage for targets over non-targets.  As a result, subjects make more errors –mistaking non-targets for targets.

We plan to adapt this basic experimental strategy to study the neural basis of attention in various groups whose performance on our task is likely to abnormal:  older adults (who show impaired behavioral performance) and habitual video-game players (who show far-better-than normal performance).  Yigal Agam is now at MGH’s Martinos Center; Heather Sternshein is in the Department of Neurobiology, Harvard University.

Sternshein H, Agam Y, Sekuler R. EEG Correlates of Attentional Load during Multiple Object Tracking. PLoS One. 2011;6(7):e22660..