August 18, 2020

NeuroChat with Chris Berka featuring Thomas Petros #002

Thomas Petros, a professor at the University of North Dakota, joins us on this episode of NeuroChat with Chris Berka.


[00:06– 05:25]


Chris: Thank you so much for taking time to have a conversation with us and we hope you'll share this conversation with other researchers or other people who might be interested in doing the type of work that you're doing. So, if you could start out, just tell a little bit about yourself, your background, your current position and go on into your research interests.

Thomas: All right, I am a professor of Psychology at the University of North Dakota. I’m starting my 41st year this fall.

Chris: Wow, congratulations.It's fantastic.

Thomas: I’m trained as aCognitive Psychologist, but I did EEG work in the 80s actually. But really got disillusioned with it because at that time it took so much development time and you know technicians that were willing to help you out come and go. So, I just chucked it because I wasn't getting anything done. So, I opened up other avenues of research but in the early 2000s, a young researcher from Honeywell came and talked to us. Santosh Madden and Murray Cooper the business manager there,, in the course of that conversation, he told us about ABM and some of the systems you have and just trying to do things with them. You have really developed things that reach out to researchers that they don't have to spend three quarters of their time doing coding and practicing how to put electrodes on because the impedances are a little bit easier than they were 30 years ago.


Chris: That's great and then thank you for that. I also started doing research with EEG in the late 70s, when I was at Ohio state and you know, we had a room full of computers. We had a room full of analogue devices and everything took lots and lots of time and was very cumbersome. Still exciting but cumbersome.

Thomas: Cumbersome, like writing your stuff to tapes and well and then. So, we also have a very well developed school of aviation here and I had been doing work with them for 20 years on different human factors things. So, that's like hypoxia and dual task. Did some work with alcohol hangover and then some other things regarding UAV sensors and then sensor work and piloting so and I had made contacts there.

So, we have been starting with your system trying to make some things happen and you know with grant opportunities small things here and there. Our first project was actually published in ergonomics with air traffic control. The first author was Kyle Bernhard. I think he said he had talked to you. We were looking at you know trying to validate the ability of the system to discriminate between various levels of workload. We were able to tap into our air traffic control school and we were able to test more experience and less experienced controllers controlling different levels of congested airspace and we hope to parlay that into something with them, you know with the new things that are coming out with ATC next gen. They always call it, but they called it next gen 15 years ago. So, we're looking at some of those, I mean those things. So, getting into this kind of work I’ll talk about more has really revitalized my passion for my research.

Chris: That's great. That's the most important thing, right. Especially after 40 years, you've got to stay passionate and excited.

Thomas: Well, exactly. I had done a bunch of work. I’ve been in individual different things with aging, cognitive aging. I want to go back to some earlier things with age ergonomics with olderadults and safety and stuff, but we started with the air traffic control thing and now I’ll just tell you a couple of the current things. We're working on one is we're collecting EEG from people in flight, in actual flight but what's andI know people have done that but the one thing that we're also collecting is data from the aircraft system.

[05:26 – 10:09]

That's feeding in, like pitch bank and I had some colleagues in aviation that were just collecting data from the aircraft to see if they could profile critical patterns that might predict upcoming accidents. For example, the final turn to approach is a real problem for pilots losing control in accidents and private pilot stuff. So, I’ve been working with a kid on a dissertation like that but we basically to not get too far into weeds but want to combine characteristics of the plane with some physiological measures to predict some safety criterion. That's ongoing. We have been collecting data from the middle of last summer but come march the pandemics shut it down.

Chris: Oh, yeah.

Thomas: So, the other current thing-

Chris: What type of plane? Is it a jet? A Lear jet or?

Thomas: No, it is- Oh god I’m going to embarrass myself. It's got wings and two motors.

Chris: A single pilot though.

Thomas: Yes, it's a single pilot.

Chris: Okay, so single pilot and then how many passengers are possible or none?

Thomas: One passenger in the back.There's that and the other thing we're working on is, I wanted to start doing some things and using your measures as indices of learning. I see you had been involved in some work a few years back to look at training and reduced workload.So, that's, as you know the specific usable outcome I thought would be really important, would be informing instructors in terms of who needs a few more trials of repetition and what I’m thinking about is the two pilots who are flying their final, their check right or whatever they both produce the same behavioral measures but one is let's say workload maxed out for lack of a better term, you know.

The other thing we're doing is, we've collected the data on a vigilance study. I was very intrigued in 2010-11 and those and I think there's been some of these in the education literature that published some work on the alpha power diminished alpha power or I’m sorry higher alpha amplitude and diminished attention. There was some of that published and so I thought that would be, if we could do something like that in the lab it might be some only-

Chris: Indicator of fatigue.

Thomas: Yeah, indicator of fatigue and pilots and perhaps inefficient scanning and sensor operators and air traffic controllers.

Chris: Right.

Thomas: So, we did a vigilant study and I’m still analyzing the data where basically the stimulus appeared in the centre of the screen and they picked out the target along around a circle on the screen and then we would measure. We collected EEG data. So, I’m recovering the mind-set but your folks who sent us lab-x I mean; I cannot believe how easy that has made.

Chris: That's good.

Thomas: I mean, I was monkeying around with our subroutines and mat lab and that is quite a fine product.

Chris: Yeah, I think people who are entering EEG research for the first time, have no idea how much data they're going to end up with. Even if it's just, you know 20 minutes of recording. You still have a lot to do in terms of sorting and analyzing and looking at it from different perspectives. So, we tried to build a lab x tool that allowed you to do batch processing of data. So, you can load in a lot of files and then look at it from different perspectives.

Thomas: Well, that tool is amazing that is just amazing, and I greatly appreciate that it just makes life easier, you know. I’m too old to spend time developing things.

[10:10 – 15:12]

[Chris laughing]

So, it's nice to know products that are out there. So, then the final thing we're doing is with UAV, we have a predator or aviation school but they have a program where they're teaching the beginning pilots for UAV predators and we have a predator simulator and so I’ve got the instructors to construct a couple of patterns and they fly their patterns. Well, they do a couple of things. They run their checklists and early in training and later in training and they fly patterns. They have them do early in training and later in training. So, we're trying to see if we can use your software basically to document learning.

Chris: Interesting.

Thomas: Yeah, I’m hoping that data comes in and I can get the government interested in it.

Chris: Yeah. So, I have had a number of conversations with the folks at AFRL about that whole transition from being a pilot of an aircraft to piloting a drone or operating a drone and a big transition and it's not necessarily the same skill set that you need to be a pilot is when you're sitting and operating a drone and from what I’ve heard. If you're a really good video gamer, you're likely to be a better drone operator than if you're a really good pilot.

Thomas: Oh, for sure, some of those pilots that I’ve met out here that really hate getting in the seat to fly a drone. They're too impulsive, too motor involved and all that sort of thing.

Chris: Right. So, I didn't realize. So, you have an air traffic controllers training program as part of the university.

Thomas: Yeah, it's part of our aviation school. So, they train air traffic control they. We train drones and we train fixed-wing aircraft. Now, the harder thing about it is, they're the hard thing from my perspective as a researcher is, they're largely directed toward training and not as much research. They're pushing that over the last few years and, but you know, you have to have money to pay for simulators. You have to have money to in that sort of thing. They're pretty much self-funded but there's a lot of opportunities here regarding air traffic control. We had work with a company in sending a proposal in out in D.C. Now I can't remember their name to utilize our air traffic control work with to see if we can utilize some kind of physiological monitoring to help direct when different next-gen things are going to be required, you know.

The next-gen thing, they talk a lot about it as automating air traffic control but they it's clear they've never read any stuff on passive fatigue, you know. When people looking at screens for long periods of time

Chris: Yeah and I think with air traffic controllers, you've got a dual problem. One is when they're overloaded and fatigued but also then there's periods, where they're under loaded but they still have to maintain a high level of vigilance and you know we had talked with one group who was looking at automating the workload. So, if you have four air traffic controllers, they all have various flight paths and plans that they're following. But sometimes traffic is light and sometimes it's heavy. So, could you use the brain state of drowsiness or cognitive load to determine who was ready for more work or who was overloaded and maybe needed a break and that's kind of the vision of the future. We would love to build that system with somebody.

Thomas: Well you know, maybe I should send this white paper out to you and that we've sent to this company. I had sent it to Murray Cooper. Do you know Murray from-?

Chris: Yeah, sure. Yeah,

Thomas: Murray knows everybody, and he's got some swamp land for you too.

Chris: Exactly.

Thomas: But that’s exactly what we have been holding-

Chris: He’s designed a little prototype display where you would you see the four heads and every time, they're in a red zone for fatigue or they're in a red zone for overload or if they're under loaded, you know. It lights up different colors. So, presumably, a supervisor could reassign.

Thomas: Oh, you have a product like that already?

[15:13 – 20:48]

Chris: Well we had a prop- semi working prototype.

Thomas: No kidding.

Chris: Yeah and I'd be happy to share it with you. It didn't really go anywhere. We had floated a couple people that had asked for it and it's a very interesting research problem but also a very pragmatic one because it's challenging work that's, I mean it's interesting that you guys have a training program. It's not for everybody number one. It's only going to get more challenging once flights pick up again.

Thomas: One of the limitations in doing that kind of work is our faculty are all former controllers and they always tell me; well the union won't ever let them wear this stuff and this and that. They kind of pooh-pooh, when you talk about the eventual end state, but I tell them, one of the things we try to analyze is the minimum number of sites that you can use and still get good workload measures. I mean I think a lot of your original algorithms were done with your x10.

Chris: That's right.

Thomas: You get some pretty good predictive power.

Chris: That's right. We really only need five sensor sites for workload and drowsiness and we've never been able to get it smaller than that. I mean we've done some other things with just for drowsiness and sleep on set you can get by with three sensors to get the whole spectrum of alertness, drowsiness, sleep onset but workload's a little more complicated and again it's different definitions of workload but this notion of mental workload and mental overload, we really require about five sensor sites in order to capture that accurately.

Thomas: Do you do mostly midline for those?

Chris: For workload, we need lateral ones as well. We need the right and left parietal and then the C3, C4as well. So, if the workload requires a few more sensors for that.

Thomas: Well wow, it might be interesting if there's some money available. If you were interested, we could somehow set up a protocol out here where we have trained students who could run some. I wonder if that might be something of interest to, you know if they come out with some of these. You could validate your protocol here with our-

Chris: Definitely, you in your previous work did you compare expert traffic controllers with novices, or did you do any of those initial comparisons? Because that's something we haven't done yet.

Thomas: Well that steering lies the problem. We had second year controllers and senior level controllers. So, they haven’t, and we did not find much of an experience effect, but they had not had that much time controlling either because when they leave here. But what part of the protocol we had sent in have sent to different places was to recruit expert controllers from the region here and the FAA wasn't able to shake lose any money or they didn't like our project or both there's sort of two things. We had ideas on with that, but one is we have quite an alumni group, who are out active controllers now and our faculty felt pretty confident. They could recruit them to come on in and so and I will if you don't mind, I’ll send you these proposals.

Chris: That would be perfect, yeah.

Thomas: We've said maybe we can mesh something together and the general idea there was our original paper did not have a very strong manipulation of experience. So, we hope to follow that up with actually practicing controllers who have quite a bit of experience but piggybacking on that was the idea of bringing older controllers in, you know the man that you may know the mandatory retirement age is 55 and some of them might still want to work and so we thought and the other part of that if you look at the man machine teaming with the physiological sensors activating different machine interfaces would that not greatly expand the air traffic control supply chain, you know. Especially as you go into next-gen stuff that has more passive monitoring. So, those are some of the ideas that were embedded in these papers that we have been shopping around hoping to go good. Well I’d love to send them to you.

Chris: Yeah, we'd love to collaborate with you on that.

Thomas: Yeah, you know if you find something you want to set up a meeting after that and go a little forward, we had been like I said working on the old controllers thing but then got busy with other things pandemic hit and some things just die on the vine.

[20:49 -25:27]

Chris: Yeah and I’m very interested in your research on live flight. We've done a little bit of that work and we're hoping to do some more with some of the new helicopter flights.I mean generally, we've found that our system works pretty well on a learjet or in in some of the helicopters that we've looked at but you always have to test it first. I’m assuming that you didn't have any interference from the electronics on board the flight or any other equipment interactions or you wouldn't have gotten cleared to fly with our b alert system.

Thomas: Oh yeah. No, there was no interference. I am not quite done looking at the data yet and cleaning it to see how much the artifacts we have gotten but I’m hoping your ears open but the thing about using your system or the general theme would be using your system in combination with data off the airplane. The angle of attack thing the turn to final is. Well, I’m kind of jumping around one of the things that the military does to get around the problem of turn final turn is they have an angle of attack device on the aircraft. Basically, to make sure the approach is at an appropriate angle and I’m kind of butchering this because I’m could barely drive let alone fly but apparently that's a major problem for private pilots. So, the idea being, if we could you know, along with characteristics of the aircraft. We're hoping to put that into some machine learning algorithms when we're done with this and hopefully something magical spits out but that doesn't always work.

Chris: Yes, absolutely. I mean that's what we do with data too.

Thomas: Those are the directional hope to go.

Chris: Yes, exciting and didn't you do some work with environmental with hazmat suits or some sort of unusual gear, where you put our headset on with other equipment?

Thomas: No, that wasn't us.

Chris: Okay, that was part of your-

Thomas: No that's so I’m into a lot of things and air traffic controllers.

Chris: Let me ask you kind of a big vision question if suddenly somebody, some agency dropped several million dollars in research funds, what is kind of your dream research project that you'd like to do?

Thomas: Oh man! Well, there's- This the reminds me-

Chris: Of my EEG.

Thomas: It reminds me of my comprehensive exams.

Chris: There's no right answer.

Thomas: Yeah, well one would be something to do with safety in older adults as they may stay in the work forces I mentioned the older adults air traffic controller. We hope you know that could be a model for other sorts of things safety during driving now. I’m not again so that might be one general broad thing of events but the other one broadly speaking would be a lot of different approaches toward using the B-Alert system or smaller electrode subsets to facilitate education and training.I not only the model I’m using now is with predator pilots with our students. We also have general atomics here and I’ve been talking with them about they have a training site here in for the border the customs and border patrol their people are taught by some friends of mine and the idea being to facilitate the efficiency of their training and some of that would be not everybody needs to stay an extra day or two, but if some people do and then their skills don't fade as fast. I think that would be valuable not just with aviation but kind of like as people are living all longer, people staying in the workforce longer.

[25:28 – 30:33]

Chris: Yes, absolutely and as we age, we have to operate all this equipment now too.

Thomas: Oh well, they're like zoom and everything else.

Chris: Yeah, I mean we've all had a crash course in zoom and video cameras, and you know all sorts of things.I think that that's a really important area because my generation, the baby-boomers, we're living longer and we're working longer, yeah choice but because we have to and by definition, we have to interact with technology. So, the ergonomics when you're 21 is very different from when you're 61.

Thomas: Exactly and you know I’ve never felt more enthused about stuff and I’m, you know now that I have a Lab-X and I got more ideas and I have time you know teaching class gets in the way but you know that's sometimes fun too. You met one of my students or maybe over the phone Kyle Bernhard.

Chris: Yes, he's a great guy. Oh my god, conversations he's got a lot of ideas too.

Thomas: He's at Boeing now working for them.

Chris: Yes, I heard.

Thomas: Oh okay. So, anyway. So, now that he's gone myself and my other colleague Dmitry Petoskey. We have to do a lot of this technical stuff on our own and let me just mention really quick. He's done some stuff up independent of me. He collaborates on this aviation stuff, but he has an interest in the use of your system and he works with Optometrists around the country using some of your measures. I believe it and I don't know this very well evoke potentials and some of your measures to discriminate people who have had concussion history.

Chris: Oh, that's interesting.

Thomas: So, I don't know much more about that. Maybe if we get a chance to talk again. I’ll make sure he's with us and can tell you some of his ideas and he's been publishing quite a few things in there has an NIH grant to do some of that.

Chris: That's fantastic and so do you do you have other do you have graduate students in your laboratory as well?

Thomas: At this point for theEEG, no. I do some other work and we're basically a clinical psych program and so a lot of my research has been working with clinical students doing things like that. I have had a couple undergraduates who left that were just dynamite.I’m hoping to get one of them come back because she helped with some of the EEG work, we've done but for now we work mostly with undergrads, train them up and try to get them excited about things.

Chris: That's great.

Thomas: Yeah, it's a lot of exciting fun stuff.

Chris: Good. Is there anything else that we can do to help you facilitate additional research? It's been areal pleasure to work with you over the years.

Thomas: Back at you. No, you guys have been so giving when Alfonso said would you do a promotional video or what kind of talk with Chris, I said of course, I will but that's, I feel indebted to you folks and yeah, you can read the white papers I sent when I said them and then-

Chris: I have tried to keep up with at least your publications and I’m happy to read the white papers and let's discuss if we can collaborate with something for air traffic control or for pilots’ live flight. Those are areas of interest for us. Definitely!

Thomas: Well, why don't I send you. We have some small funding to support the pilots in flight. We had to write a small proposal to get that internally. I get into the office tomorrow the next day. I’ll collect everything and send you an email and then shoot mean email. Have your guy arrange a time if you want to chit chat about it. We'll go from there.

Chris: That’ll be great and if you have any still photos or any video images because that, you know the live flight lends itself well to having a nice photograph of the pilot with our headset on or anything else that you might have any images associated with your work that we could publicize. We're trying to you know we there are customers like yourself that we've worked with for I think a decade, almost a decade probably and have done some incredible work and so we want to sort of, circulate that, publicize it more. Generate some research collaboration and hopefully get you some additional funding.

[30:33 – 33:10]

Thomas: Okay, there you go then.That would be well, I would look forward to that. I’ve read all your stuff and a big fan of all your stuff and I would really look forward to that.

Chris: All right, fantastic. Anything else? So as far as coved are you completely shut down from doing any human subject work, right.

Thomas: Yeah, the research office is trying to come up with a protocol for doing human subjects. I think it's going to be slow in the fall. I know that they've started flight operations, but both the co-pilot and the student have to wear masks and we don't want to continue collecting data there and introduce hypoxia compound. So, whatever, that's been a bit shut down for now. So, we'll see what happens. I’m hopeful, you know.We'll start up after Christmas if the virus ever, if we ever get serious and start working together to beat the virus.

Chris: I know. So, the university will students be coming back. Are you doing a hybrid virtual?

Thomas: Kind of a hybrid the students will be coming back, and a lot of the faculty are just doing online.I’m doing online and in person but at least for the fall. I’m a statistician also and I teach all our graduate stats and I’ve done that online. So, I’m going to teach a couple of my in-person classes what they call flipped where they're going to watch some of my lectures recorded which I’ve already done and then have them come in on zoom and we'll do some hands-on, I guess they like to call it and we'll put your cap on there and see if the engagement measure stays high.

Chris: Oh, you should do that. Definitely,I know I’ve thought about that a lot when I get on these zoom calls with multiple people and you're trying your attention is split into multiple directions. I’d love to cap on and just kind of see who's paying attention and who-

Thomas: Yeah and call them out, number one you're zoning out now. But anyway. It's really been a pleasure talking to you.

Chris: Thanks so much for your time. Let's work more together sounds good.

Thomas: Thank you Chris.

Chris: All right, take care.

Thomas: Bye.


More News