Our next speaker is Abdullah Bin Zahid, who is speaking on automated eye tracking for detection of blast brain injury after a natural gas explosion. Thank you so much for having me over here. I am Abdullah Bin Zahid. I am a research scientist at Hanover County Medical Center and University of Minnesota. These are my conflict of interest. As you can see, don't believe what I say. Believe the objective data. So we are having an active discussion about concussion and how should we define it, how should we measure it, and all the aspects of concussion, and how to diagnose it, and how to treat concussion. But now we have something that is another variety of concussion, which is blast injury, in which there is no direct impact to the brain. So in blast injury, the shock waves enter the brain, and because we all know that sound travels at different speed in different medium. So the sound travels at different speed in blood, versus the brain parenchyma, versus the interstitium. So whenever a shock wave enters the cranial cavity, it disrupts those connections because it is traveling at different speeds in the different tissues inside the brain and cranial cavity. So this leads to injury, but that injury is not visible on MRI or the CT scans or any other conventional modalities. So how can we measure the impact of a blast on an individual? So we are trying to use eye tracking to quantify this impact, and we are trying to use eye tracking because eyes are one of the organs that have like such a close relationship with the brain. Almost every single area of the brain has something to do with the eyes, from frontal eye fields to occipital lobes to the centers in the brainstem. Almost every single area in the brain has some sort of connection to the eyes. So with such a diffuse injury, blast injury, we can probably detect the impact of it using eye tracking. So we are trying to use 500 hertz infrared camera to diagnose this injury. Here is a typical eye tracking setup. The observer is sitting in the left lower corner of the picture, and observer is watching a music video. The music video rotates around the periphery of the monitor. The infrared camera below the monitor tracks the eyes, and the operator is on the right-hand side of the picture. The operator operates the setup using a conventional laptop while the observer is watching or the subject is watching the video. The entire process takes about 220 seconds. Here is how it looks on the operator screen. The crosshair shows that the camera is catching the eyes. And then the video goes around the monitor screen four times, 10 seconds, 10 seconds, and then 10 seconds, and then 10 seconds. So these are 40 seconds for each cycle, and video goes around the screen for five cycles. And then over there, you can see the eye tracking session in progress. The yellow and green lines show how well the eyes move along the y-axis. And the blue and the red lines show how well eyes are moving along the x-axis. The missing data is a blink that the person did. And then we had an unfortunate natural gas blast explosion in Minneapolis, and the subjects were brought into the hospital for the treatment. Many of those subjects didn't have any signs or symptoms, and they didn't even seek medical attention. We sought them out and brought them into the lab for the participation. So this was the building before the blast explosion, and this was the building after the blast explosion. As you can see, there was significant damage. And then here is a normal eye tracking. As you can see, for all the five cycles, the eye tracking trajectories are superimposed pretty tightly, which means that the eyes are moving well. The blue areas under the curve in the second and the fourth plot shows how well eyes are moving together along x-axis and along y-axis. The lesser the blue area, the better it is. So in a blast survivor, you can see almost immediately after blast, the eyes are not moving well together. In the four boxes, you can see that the five cycles are not superimposed on top of each other, which means that the eyes are certainly going in different directions. And then in the blue, in the second and the fourth graph, you can see that there is a lot of blue area, which means that the eyes are not moving together well along the x-axis and along the y-axis, especially so along the y-axis. So involuntary, like whenever we are looking at a screen and we follow it along the smooth pursuit movements, doing the smooth pursuit movements, both left and right eyes move together. And this is something involuntary. And over here, we can see that there is a deficit in that involuntary portion of that eye movement in a blast survivor. Can I go back? Yes. So over here, you can see the location of the participants along with their BISC score. The BISC stands for blast impact score that we created a model using machine learning and specifically using logistic regression by comparing the blast survivors with their counterparts from the normal population. And we came up with a BISC score, which is blast impact score. And the blast impact score of greater than 10 was considered to be the threshold where one should recommend seeking care. And in this thing, the yellow triangle shows the site of the blast, the blue two dots show the two unfortunate fatalities. And then you can see the closer you are to the blast epicenter, the higher your BISC score. And then as you move out from that, your BISC score goes down. It is very low in the people who are like outside of the building. The blast did more damage to the people who are inside the building versus those who are outside because shockwaves spread quickly outside. So with this, sorry for the interruption and thank you so much. Any questions? We have time for one question. Is there any follow-up to how long it takes for those findings to resolve rather than, you know, to be tested? This research was fairly recent. We are currently doing follow-up. I don't have any follow-up data right now. Thank you. All right, that slide was worth the wait. It's good.

Abdullah Bin Zahid

research scientist

automated eye tracking

blast brain injury

natural gas explosion