Thursday, November 21, 2013

Augmented Cognition and some pretty cool new toys.



Dylan Schmorrow gave an intriguing talk at TEDxDrexelU.  I just saw the video up on YouTube.   

For those of you who are not familiar with Dylan, he is a phenom in the area of Augmented Cognition.  He just retired from his job as Deputy Director of Human Performance, Training, and BioSystems at the Office of the Secretary of Defense and joined SoarTech as Chief Scientist.  He is one of the most scientifically capable human factors pros I have ever known. 

The basic idea of his talk is that we don’t need to know exactly how the brain works in order to exploit what we do know.  Sometimes, having a simple (and oversimplified) connection from some kind of brain activity to some mental construct can be all we need to improve system performance using augmented cognition.  Here are a few of the examples he used.   
 
1.  We know that working memory is often the major bottleneck in human performance.  There is only so much we can keep in mind at once.  We also know that there are different brain areas responsible for different kinds of working memory, such as spatial or verbal information.  We don’t know exactly how these work.  Perhaps these areas are in charge of temporary storage.  Or maybe they are pointers to long term memory.  Or maybe they are like the orchestra conductor of working memory.  His point is that it doesn’t really matter if all we want to do is predict when someone’s working memory in one of these modalities is getting close to full.  He says that we have the ability now to attach a sensor to a person’s forehead and assess if one of these areas is getting close to full (or not).  If it is close to full, we can reframe additional information into a different modality or we can reduce the overall task load.

2. We also know that there is a particular signal called the P200 spike that indicates when a piece of information was really processed. We don’t know exactly what this is signifying, but we know it happens at about the same time. So we can attach a sensor to monitor the P200 and if important information is presented we can check for a spike.  If there is no spike, the person needs a reminder to pay attention.  Dylan jokes that this can be important when your spouse reminds you to pick up milk on the way home from work and you automatically agree without really hearing what she said.  But this could be really valuable when something unexpected happens on the road while you are driving.  A sensor in the car can see if you noticed it, and if not, get your attention. 

He had a few other examples, but I think you get the point.  There are things we can do right now with our limited models of brain processing that would work pretty reliably.  This is what he has been doing at SecDef and I suspect now at SoarTech. 

Pretty cool stuff, don’t you think?