This company uses AI to spot concussions


Brightlamp, which is a computer vision and machine learning startup out of Purdue University, is working on a medical software that could diagnose possible concussions. This is important in many sports, most notably football where up 40 percent of pro players have shown signs of dementia, Alzheimer’s and depression due to repeated concussive blows to the head.

The company has developed an app that uses deep learning to spot concussions in those who’ve been hit hard in the head. It works by flashing a bright light from a smartphone into the person’s eye and then measuring pupil dilation, which reportedly helps determine if there is a traumatic brain injury after a blow to the head. From there, the app sends the image to a cloud-based neural network to figure out if the person has the characteristic markers of someone with a concussion, and returns the result to the user in about five seconds.

According to Brightlamp’s co-founder Kurtis Sluss, the app has a near perfect diagnostic accuracy of 98 percent when compared to the same data from MRIs. He added that once the software is out, it will be available in two different versions — one for coaches and professionals and the other for athletes and others who want to test themselves.

Brightlamp’s app has yet to be FDA approved to be able to provide medical diagnosis, and in the meantime the company is working to patent-protect its technology.

Sluss doesn’t want to stop at concussions — he aims to use the tech to determine other diseases with just your smartphone in hand, though he hasn’t revealed what’s next for the company.

Sluss is currently a chemical engineering student at Purdue. Other members of the Brightlamp team include Jonathan Holt, Craig Wilhite, Michael Heims, James Waggoner and Dylan Sinks, who also are students in various engineering departments at Purdue.