Dog, human, and robot bonding: Past, present, and future

Jeffrey Katz Author

04/26/2022 Added

33 Plays

Description

This is a recording of Dr. Jeffrey Katz's presentation for the Nebraska Symposium on Motivation given on April 21, 2022.

Searchable Transcript

Search:

[00:00:12.450]Jeffrey Stevens: Alright, welcome back everyone thanks for for joining us for the afternoon session here in this nebraska symposium on motivation.
[00:00:19.560]Jeffrey Stevens: or next speaker is Dr Jeffrey Katz. Dr katz received his bachelor's degree in psychology from Ithaca college before completing his masters and PhD PhD in experimental psychology at tufts university work Robert cook.
[00:00:35.670]Jeffrey Stevens: You completed a postdoctoral fellowship at the University of Texas Medical School working with Tony right he joined auburn university's faculty in 2000.
[00:00:46.680]Jeffrey Stevens: and has been there ever since so bravo.
[00:00:50.100]Jeffrey Stevens: He is currently the Michael and leanne more Sani row endowed professor in the college of liberal arts Dr katz is very active in the comparative cognition society, which is a.
[00:01:00.660]Jeffrey Stevens: society that I frequent as well, where he won their recognition service award well so congratulations on on your service to that society, this is a.
[00:01:11.040]Jeffrey Stevens: involves an annual conference in Florida, and this is where I met Jeff was in this annual conference it's in the same hotel in Florida we're on the beach.
[00:01:18.960]Jeffrey Stevens: Beautiful it's a really great place to have a conference, but I gotta say it's a little strange not seeing you in a flowery beat shirt and actually you.
[00:01:27.870]Jeffrey Stevens: yeah yeah maybe we should have had a call the casual day tomorrow, you can wear your beat shirts, but it's great to see spiffed up for the for the symposium here.
[00:01:36.900]Jeffrey Stevens: Dr katz his research focuses on the comparative mechanisms of learning and cognition including behavioral and functional neural imaging methods to investigate dog human attachment concept learning observational.
[00:01:48.210]Jeffrey Stevens: Learning and the behavioral and neural correlates of working dogs, he also.
[00:01:52.860]Jeffrey Stevens: If you've been paying close attention has over 100 publications, which means that we have a hat trick for our speakers today.
[00:01:59.670]Jeffrey Stevens: Three over 300 publications among the among the three of them so very well very well published all of our all of our speakers so with that I will stop here and hand it over to Dr cat so please join me in welcoming Dr katz.
[00:02:22.260]Jeffrey Stevens: All right, well it's an honor to be here today, and I wanted to start by thanking Jeff for inviting me.
[00:02:28.380]Jeffrey Stevens: them nebraska symposium when I was in Grad school, I remember coming across an article by al Campbell and I just loved it and I was like Oh, this is the best thing ever.
[00:02:38.790]Jeffrey Stevens: And so, when Jeff asked me to like participate, I was like yes, because to me the scholarship is amazing over the years and i'm really honored to be here great great campus great facilities really enjoying.
[00:02:50.160]Jeffrey Stevens: So today i'm going to talk about some of our research with dogs and humans and some new stuff with the robot bonding and let's get into it.
[00:03:00.960]I can just use the mouse.
[00:03:04.560]Jeffrey Stevens: So, as many of you are aware, there's been the rise of the dogs are the last two decades and there's a number of factors that have gone into this in terms of producing more publications.
[00:03:13.260]Jeffrey Stevens: A lot of these publications have in fact dealt with the human bond now i'll talk about a little bit of our research that we've done with that.
[00:03:20.910]Jeffrey Stevens: But I came into the kind of I started off as a primate researcher and pigeons and birds way back here and kind of jumped into the rise in this kind of area.
[00:03:30.870]Jeffrey Stevens: And i've been really enjoying it and I think it's great a lot of questions to ask so let's let's go to get into this more.
[00:03:38.970]Jeffrey Stevens: So we've heard a lot of great talks today and it's really an honor to be with them, and some of the stuff you've thought you've heard a little bit about today and yesterday.
[00:03:46.500]Jeffrey Stevens: But when we study the dog human bond, and we look at canine behavior.
[00:03:50.940]Jeffrey Stevens: And their cognition we often use a number of different methods to do their number different cues.
[00:03:56.130]Jeffrey Stevens: So it can be communicated gestures as common so hand signals how that dog behaves to that how they act to that we can look at play behavior between the dog and the human and how they interact and I.
[00:04:06.600]Jeffrey Stevens: I really love the new field of embodied cognition and kind of studying dogs and humans this way I think it's really fascinating.
[00:04:13.230]Jeffrey Stevens: Human speech, we know dogs can can respond to the emotional content of the of the speech and the meaning of the words that we say and one of the common ones, is how dogs use pointing to kind of indicate.
[00:04:25.020]Jeffrey Stevens: Where we can indicate to them locations to go to based on our points for them to discriminate.
[00:04:33.120]Jeffrey Stevens: So I like to think about measuring these cognitive behaviors and other people have to do, we can put them into kind of passive and active processing and what I mean in this context and the passive processing.
[00:04:44.010]Jeffrey Stevens: is essentially the dog doesn't have to actually do anything right, so if they're watching pictures we can scan their brain measure their heart rate.
[00:04:51.480]Jeffrey Stevens: Looking at videos they don't actually have to make a discrimination, but we can measure their internal processes This way we can people do that with.
[00:04:59.310]Jeffrey Stevens: With a variety of different tasks and then there's active processing, and this is where the dog has to physically do something so respond to a command.
[00:05:06.990]Jeffrey Stevens: The guests or no or paradigm is really fascinating to look at the theory of mind and dogs and it can they actually possess this.
[00:05:13.890]Jeffrey Stevens: object choice is a common procedure as as object search and i'm going to talk a little bit about strange situation test in the unsolvable task you've heard some about that already the conference i'll say a little bit more.
[00:05:26.550]Jeffrey Stevens: So the approach i've taken in my career and many people in comparative cognition take the same approach when we think about evolution.
[00:05:35.460]Jeffrey Stevens: Is when we compare across different species what produces that mechanism where did it originate is that our residual mechanism.
[00:05:42.840]Jeffrey Stevens: And the species, so in this case we're dogs, it was before they split off from other species and it's just leftover or is it due to domestication over the last 40,000 years.
[00:05:52.380]Jeffrey Stevens: We can also look at ontological factors right, so how the dog develops in their lifetime and and that's really important to look at to.
[00:06:01.590]Jeffrey Stevens: The approach I taken balls combining a bunch of different techniques, so a lot of us do this I like to look at the behavior look look at look at the underlying cognitive processes that underlie that behavior.
[00:06:11.820]Jeffrey Stevens: will look at anatomical differences and functional neural mechanisms, and today I think it's really strong and in a lot of different areas kind of take this converging operations approach and.
[00:06:21.600]Jeffrey Stevens: And the reason why we do it is, it allows us to get reliability and validity.
[00:06:25.740]Jeffrey Stevens: In our in our findings, because one task by itself is always has a weakness, but when you can see similar things across a variety of different methods.
[00:06:33.750]Jeffrey Stevens: It gives us much stronger conclusions to make about the processes that were studied and over here on the right is Charles darwin's family.
[00:06:42.450]Jeffrey Stevens: And this is his dog Bob in the middle of it and I truly believe is Charles Darwin was alive today he would take a very similar approach in looking at evolutionary cognition and animals.
[00:06:56.070]Jeffrey Stevens: So I wanted to paint the backdrop of the dogs i'll be talking about they're working dogs we work with pet dogs too, but everything i'll talk about today is working dogs.
[00:07:04.380]Jeffrey Stevens: And I collaborate with kanye performance sciences and you know the small subset of what i'm going to talk is very small relation to everything that these dogs do.
[00:07:13.410]Jeffrey Stevens: Alright, so we look at their performance behavior their health issues we look at genetics phenomenal step out with will behavior.
[00:07:19.290]Jeffrey Stevens: Where I do most of the work at CPS is with all faction and cognition and behavior and training area that's where most by some of it dabbles in the other other areas of operational didn't dynamics and targeted detection.
[00:07:32.460]Jeffrey Stevens: But all the work we do in the studies that we're looking at they come back to asking about the auburn dog and what ways can we make the auburn dog better for detection, as we move forward and in all the factors involved with it.
[00:07:48.420]Jeffrey Stevens: So I wanted to start by talking about some of our our MRI work.
[00:07:53.220]Jeffrey Stevens: And so we've been where one of the few places that does it it's really exciting it's difficult it's challenging.
[00:07:59.190]Jeffrey Stevens: It starts off with everybody asks how do you get them the right lifestyle and for those of you do training with dogs, this is not going to be too surprising, but they start with their clicker you know.
[00:08:07.800]Jeffrey Stevens: We got free ones outside, so I figured out, but basically the dog responds to the to the click and get to treat there's going to be we're going to use a stick, and I should say we've done this method in several different ways.
[00:08:20.520]Jeffrey Stevens: And, but the the goal of all of it is to get your dog to go into the scanner and remains still for over five minutes right they need to be able to do that, for us to get quality data.
[00:08:31.350]Jeffrey Stevens: So the first method we did is actually developed by Paul Wagner.
[00:08:34.920]Jeffrey Stevens: And he did clicker treat training and basically just like I demonstrate with the click you get the tree, but then he puts in the target stick, and so the dog has to go and follow the tip of a stick.
[00:08:44.340]Jeffrey Stevens: gets rewarded for doing that, and you slowly increase the time that they station at the stack right so and you get to build up the time.
[00:08:52.770]Jeffrey Stevens: And you go into the mock scanner and i'll show you some pictures a little bit and basically bring them in and they got to sit inside the bore continue the stationing play the sounds of the scanner.
[00:09:04.350]Jeffrey Stevens: And then we go over to the real real deal and see if they transfer there.
[00:09:10.860]Jeffrey Stevens: So here's some pictures.
[00:09:13.170]Jeffrey Stevens: So there's a trainer with a clicker in the stick in the hand and the dog, you know stations to that additions to it here is one of our very expensive mock coils.
[00:09:22.650]Jeffrey Stevens: You can see the dog goes in there and in lines lives down in the prone position in the same position will be in the scanner and you build up the station and behavior there here's our original mock scanner.
[00:09:34.530]Jeffrey Stevens: it's the one you guys have here is great to i'll show you our new one it's it's you'll see if you like, at the same or not, but it's.
[00:09:42.030]Jeffrey Stevens: The Doc comes in here the station inside the head coil there and get comfortable and again building up the time for over five minutes and playing all the sounds of the environment that they're going to have to experience.
[00:09:55.830]And having had protection while they're doing it.
[00:09:58.590]Jeffrey Stevens: So here's our current mock scanner so we do all these CPS dogs right they there are some that become research dogs, and these are the ones that that we work with.
[00:10:09.330]Jeffrey Stevens: And they spend their time doing research, but they also go to prison.
[00:10:13.380]Jeffrey Stevens: And some of you are probably familiar with these programs and but essentially they work with em mates and it's a treat for the enemy is going to have this.
[00:10:21.540]Jeffrey Stevens: chance to work with dogs and they like it, and this data, my show you just a second is actually from inmates that the the original station and behavior training right.
[00:10:31.890]Jeffrey Stevens: And there are several of them that are amazing artist and they actually.
[00:10:35.670]Jeffrey Stevens: mimic our our semen scanner very similar the one you guys have here we have a little auburn you know Doc print on there, of course, so maybe it's a tiger, but you know call it a dog today.
[00:10:45.690]Jeffrey Stevens: And then you know, but essentially a dog kind of comes up the ramp goes into the to the board and stations in there, and you know, he would play all the sounds get them acclimated to the environment and then we're we're ready to go.
[00:10:59.640]Jeffrey Stevens: Here is here's Paul Wagner here's a here's our MRI sweet.
[00:11:03.990]Jeffrey Stevens: The dog we actually use a human knee coil for the scans and it's really just because the dogs had sits more comfortably in there and in the in the closer you have to the side of the coin, the better your signal that headquarter for humans is too big for for most dogs.
[00:11:19.620]Jeffrey Stevens: here's another dog good picture that new coil, then you know you could basically go into the bore and and get them in there and then start collecting your data.
[00:11:30.840]Jeffrey Stevens: Most of our research started with all faction back then, we were able to segue in some other areas which we'll talk about it a little bit I wanted to show you one video before I do that.
[00:11:44.130]Jeffrey Stevens: So dogs most dogs actually like doing this they'll just get ready to go.
[00:11:51.840]And then he's just gonna wait for the queue.
[00:11:57.300]Jeffrey Stevens: And he'll go in and it's a they're just like humans, the really good ones they fall asleep.
[00:12:04.710]Jeffrey Stevens: You know the really bad ones don't you know but it's like a you know so it's like you build up to this all right, and then you got to increase the duration.
[00:12:12.810]Jeffrey Stevens: And I want to show you some some brand new data, so we took a see bark and we looked at the energy score.
[00:12:19.950]Jeffrey Stevens: And these are the ones that were trained with with the prisoners and we said, like Okay, based on this, he barks score how long did it take how many sessions to actually get good data.
[00:12:30.210]Jeffrey Stevens: right when you went to the Mr environment, and you can see the correlation with with energy right, so if you're not a very high, you know energetic dog, you know it takes less sessions, but if he has more energy.
[00:12:42.270]Jeffrey Stevens: big surprise right it takes longer for them to actually acclimate to the environment so some of the first I think it's the first data looking at that.
[00:12:50.400]Jeffrey Stevens: and seeing how that impacts and the training it, but you know, the bottom line is after they get there, and you can get good data.
[00:12:57.750]Jeffrey Stevens: they're good they're in the saddle it's very rare that a dog quits on you, they can choose.
[00:13:02.100]Jeffrey Stevens: They can choose to do that and that but we've only had one dog that I know that to said middle of, say, like i'm not doing this anymore.
[00:13:08.550]Jeffrey Stevens: Right, it was frustrating but it's like but, but overall once they're there you can bring them back several months later.
[00:13:15.060]Jeffrey Stevens: go right in get some data it's Nice, so you did it there's some tool time to get there, but once you're there it's good.
[00:13:21.390]Jeffrey Stevens: And I should mention that you know these dogs are all purpose dread right bred to be high energy working dogs, these are super energetic dog so it's a.
[00:13:31.110]Jeffrey Stevens: it's different for them to sit there for five minutes is a challenge, or they do not want to do that.
[00:13:36.930]Jeffrey Stevens: So I often wish you're like Oh, I wish we work with a golden retriever that was very placid and you know, whatever operate a much easier in these situations, but.
[00:13:45.450]Jeffrey Stevens: But, but this is what we have and we any other thing is, we can't like cherry pick and say, like, I want to work with this one, or that one it's just you know you have what you have you know so it's so you work with it.
[00:13:56.580]Jeffrey Stevens: Alright So the first thing I want to talk about is work that I did with with several people Andy Andy tompkins former student of mine took the lead on this particular project.
[00:14:06.690]Jeffrey Stevens: So this is what the again with a working detector dogs what we're going to do here is, and this was after the old faction work and we're gonna we're going to start asking questions about the relationship between the dog and the trainer.
[00:14:19.770]Jeffrey Stevens: Right and we were interested in that for, for obvious reasons, you know what is that relationship, and of course this we're all about the motivation and bonding here, so I figured i'd focus on this aspect of the research.
[00:14:31.980]Jeffrey Stevens: So what happens is the scanner and they're going to look at pictures of faces and there's going to be faces of familiar people, which are their trainers are healers.
[00:14:41.880]Jeffrey Stevens: All right, so they've had extensive time with a dog and they're going to be unfamiliar people that they have no training experience with it all.
[00:14:48.180]Jeffrey Stevens: Right, so we do that, and then everything's Nice and counterbalance and the design, then we manipulate the emotional expressions and the pictures, so we have positive one, so these are people with SMILEY faces and i'll show you pictures in a second.
[00:15:03.450]Jeffrey Stevens: We have negative ones, and this is angry faces and then we have neutral faces, you know expression.
[00:15:10.920]Jeffrey Stevens: So we do all that we then also show them pictures of dog faces.
[00:15:15.570]Jeffrey Stevens: Right and so those dogs we didn't finish manipulate the the emotional content of these pictures, but we did show them ones that they were coupled with and familiar and kind of hung out with him that exercise whether then unfamiliar ones that they they had no experience with.
[00:15:30.630]Jeffrey Stevens: We also show the dogs pictures of human videos and i'll show you what those look like and then those videos we manipulated.
[00:15:38.610]Jeffrey Stevens: Again, the emotional Valence of the videos the time frame was the same it's five seconds for each image, the videos went for five seconds and they cycle, a little bit to kind of have the time frames.
[00:15:51.630]Jeffrey Stevens: But if you're interested in the Mr parameters I haven't found their.
[00:15:56.700]Jeffrey Stevens: Okay, so here are some of the trainers on the left, so the familiar ones, you can see a picture of positive neutral negative and i'll show you more in a second and here's you know unfamiliar one so.
[00:16:07.410]Jeffrey Stevens: saying we match this all up and we we we match for all the Valence measures for negative neutral and positive.
[00:16:15.210]Jeffrey Stevens: We want to do that, so that when we look at contrast between familiar and unfamiliar people right, we want to make sure that that the balance scores warrant you know influencing it in some way, so we bounced for that aspect of this.
[00:16:29.970]Jeffrey Stevens: And these are ratings done by humans right, so the humans are watching this is separate study they're looking at the pictures they make their scores on it categorize them in terms of the emotion and we do the same thing with the videos.
[00:16:44.970]Jeffrey Stevens: These are what the videos look like the the positive faces, by the way, former Grad students that didn't have any interaction with dogs right so like.
[00:16:53.940]Jeffrey Stevens: This is Alex Daniel going good dog good dog good dog, and this is a Alex is now a professor and Stevens at Harvard now, and so this is a.
[00:17:05.130]Jeffrey Stevens: he's going, you can do this, you can do that you could do this, you can do that so and then Adam one of my Grad students you at uab now he's gone bad bad dog that dog so it's so.
[00:17:17.820]Jeffrey Stevens: So, so they were you know doing other projects not related to dog, so they they were they had no contact with so they've served served good so, and so we match all that up to.
[00:17:27.390]Jeffrey Stevens: Alright, so first findings, I want to share with you is this when we contrast and just look at.
[00:17:34.680]Jeffrey Stevens: Familiar human faces I sorry yeah.
[00:17:38.700]Jeffrey Stevens: We just look at human faces versus Doc faces all right, and so what we found and there's only two areas as a whole, brain now since there's only two areas of which.
[00:17:48.270]Jeffrey Stevens: We found activity selective to the human faces in the dog faces and so there's the, and this is now left turbo cortex.
[00:17:57.120]Jeffrey Stevens: And the dog areas and red and the human areas and green and i'll call them dog and face Eric dog and human face here is not to say that these are selectively processing only these stimuli.
[00:18:10.980]Jeffrey Stevens: But it's just out of convenience and talking about it and the paper, so what you're looking at if you're not familiar with these kind of data plots is.
[00:18:20.070]Jeffrey Stevens: The stimulus is not here in the white area, then we turn the picture on right, and you can see if it's a human face the amount of activity, so this is changing the bold signal goes up right, and then it comes to have and we turn the stimulus off.
[00:18:35.610]Jeffrey Stevens: For the dog face right little activity, but not nearly as much as the human face right and that that's the that's the green blob up there in the in the brain slices it's the same logic on.
[00:18:48.120]Jeffrey Stevens: On the panel below but you'll see more activity for the dog faces, then the human faces there and that's that's the Little Red blob up there.
[00:18:57.090]Jeffrey Stevens: And this was you know this was this was me because other people I had shown.
[00:19:01.980]Jeffrey Stevens: That the dog brain is responding will respond to dog and human faces, but this, the first time.
[00:19:07.740]Jeffrey Stevens: We were able to separate dog and human faces other other labs jonah was just the area was overlapping for human and dog faces and so why might that be for working dogs, the other studies involved pet dog so.
[00:19:20.490]Jeffrey Stevens: It could be It could be that the fact that pet dogs are living in the household with the humans, the way they're representing the faces right it's it's they just go together but.
[00:19:31.440]Jeffrey Stevens: For Kendall dogs and we're trainers are really important right for them and humans are really important for for different reasons and they're not living, you know every hour their days with humans.
[00:19:41.760]Jeffrey Stevens: That maybe there's some plasticity that goes on here, so it could be an ontological difference that we see across working dogs dogs but but that's unclear at this point.
[00:19:52.650]Jeffrey Stevens: All right, so all right analysis, the only thing we found there was again those two areas for human dog faces.
[00:19:59.730]Jeffrey Stevens: When we look at familiar and unfamiliar dog faces are tested and pick up anything there, so there was no, I mean the brains active when they're looking at it, but there's no difference between.
[00:20:08.970]Jeffrey Stevens: Familiar and unfamiliar dog faces, so the next thing I want to talk about is the emotional content.
[00:20:15.720]Jeffrey Stevens: And so, when you look at the contrast we took all the familiar images of humans and all the unfamiliar images of humans.
[00:20:22.710]Jeffrey Stevens: In the panel a for the still images we found one area that was active the cost eight right and we looked at the the same logic, but for the cottage for the videos you get the cottage and regular.
[00:20:34.650]Jeffrey Stevens: Active right so so that was interesting and other laboratories had shown independently, that the caught, it was at active for familiarity to for odors of their of their voter owners, excuse me.
[00:20:49.350]Jeffrey Stevens: So, in terms of the emotional content, what do you see in the on the left side here is the positive versus neutral contrast so that's taking all the positive human faces comparing him to all the neutral.
[00:21:00.900]Jeffrey Stevens: human faces and what we see for the stone which is still images we get the caught it again and for videos we get the hippocampus and the amygdala.
[00:21:09.930]Jeffrey Stevens: And then, for the negative images what we get is negative negative versus neutral again we get the kata and hippocampus and they may be one cottage for for the videos.
[00:21:22.260]Jeffrey Stevens: And so, this was you know our original goal was not to kind of study the attachment network, you know we're just going to interested, but if you're familiar with.
[00:21:30.300]Jeffrey Stevens: The attachment network and humans right you'll know that these structures are very calm right there they're part of it, and you know i'll show you that in a SEC.
[00:21:41.280]Jeffrey Stevens: So this is our version of the unsolvable task you've seen other versions of it today it's.
[00:21:46.320]Jeffrey Stevens: it's great because you can do it real quick because Brian was talking about earlier dogs dog starts here at the bottom.
[00:21:51.990]Jeffrey Stevens: And can can goes over here and flips over the container and gets access to the reward and they do that right away in our version, we have a human on the left and the right side.
[00:22:04.080]Jeffrey Stevens: And, and what happens here is we kind of balance for whether it's the familiar person or the unfamiliar person.
[00:22:10.530]Jeffrey Stevens: And so the question is going to be when we get the trial when it's unsolvable right we lock it down who do they look at more all right, the person that they have training experience with or the other person.
[00:22:23.730]Jeffrey Stevens: And i'll show you that in on the right here is the trainer okay.
[00:22:37.500]Jeffrey Stevens: right away.
[00:22:39.840]Jeffrey Stevens: realizes.
[00:22:43.380]Jeffrey Stevens: trainer once looks back down.
[00:22:50.400]Jeffrey Stevens: Alright, so what we do is we we count the number of times they look in for how long they look.
[00:22:58.380]Jeffrey Stevens: And you know I gave you a nice example where they're looking at the trainer.
[00:23:03.600]Jeffrey Stevens: When you look at the data is separated by familiar unfamiliar, this is the solvable said, you know side there is not a lot of behavior not surprising because they can get the reward without any problems.
[00:23:14.010]Jeffrey Stevens: But when we lock it down what you'll see is there's a little more looks at the familiar person, they also look at the unfamiliar person, but what we have a really strong significant difference is the amount of time they look at the familiar person alright.
[00:23:30.330]Jeffrey Stevens: So what we did with that data can you know, Brian prime this great work but individual differences and dogs, we took each individual dogs data point from.
[00:23:41.850]Jeffrey Stevens: From the familiar familiar analysis there and we correlate it with their brain act and we want to know, like is the relationship between the individual dogs how much they are looking at the trader and the different contrast.
[00:23:58.770]Jeffrey Stevens: And there wasn't some situation, so, if you look over on the left side when we contrast familiar unfamiliar.
[00:24:05.940]Jeffrey Stevens: That duration value was correlated with the amount of activity that we saw in the individual log for the amygdala caught it and hippocampus for only one of the the.
[00:24:17.820]Jeffrey Stevens: The emotional contrast, and we see a relationship and the hippocampus for positive stimulate.
[00:24:23.700]Jeffrey Stevens: All right, and then for the for the family or unfamiliar for the videos we again we saw the amygdala cottage hippocampus so these three structures are important for representing the relationship, they have with the dog in terms of the trainer.
[00:24:42.810]Jeffrey Stevens: So nothing we did there is next next in this analysis is it's called conductivity fingerprints, this is developed by a group.
[00:24:51.300]Jeffrey Stevens: at Oxford and it's pretty interesting they are there they're looking at functionally homologous regions potentially between non human primates and humans.
[00:25:04.170]Jeffrey Stevens: Right, they want to ask about whether the networks during resting state or similar different right so.
[00:25:11.610]Jeffrey Stevens: So they published several papers i'm doing this in there, they use resting state data, so what resting state data is you're not familiar the you basically go into the scanner you look at a little fixation point a plus sign.
[00:25:25.350]Jeffrey Stevens: And then you think about nothing that's your instructions and it's actually pretty hard to do because thinking about nothing, for you know 10 minutes is not enjoyable.
[00:25:35.280]Jeffrey Stevens: But, but people do it and you do the same thing with dogs, they did the same thing with with with non human primates right the verbal instructions are a little different, but you just put them in and scan are on the same sequence right.
[00:25:50.400]Jeffrey Stevens: Can you match up different different networks, so the way this works is for our study we use the human connect own project right and we downloaded 154 subjects from there.
[00:26:00.750]Jeffrey Stevens: And we match them for dog ears and human years and we selected 19 target regions distributed across the brain.
[00:26:10.140]Jeffrey Stevens: All right, and so what happens you take a target region seed right and you ask when you're looking at resting state data when is it active in relationship to other regions in this network that you've kind of predefined.
[00:26:25.200]Jeffrey Stevens: And so, when you look at these plots right it's got three seeds here.
[00:26:29.880]Jeffrey Stevens: You know each representing the different points here so there's there's.
[00:26:34.620]Jeffrey Stevens: Not a lot of activity now our connection there, but there is for those two locations and so basically each one of these patterns becomes a fingerprint for the see all right.
[00:26:43.680]Jeffrey Stevens: And so you you do that, and one species all right, and then you do it in the other species, so you take your your seed from one species you take your seats to the other species and you map them on so they're kind of analogous in the regions right.
[00:26:57.540]Jeffrey Stevens: And you see if there you can find a match where they settle in statistically, and so you can see, this this image on the bottom matches for the one on the right, so you have a match there now, when we did this with the dogs.
[00:27:09.930]Jeffrey Stevens: Right, we had our 19 regions we took those two regions, what I talked about the human, you know dog area and we use that when we selected our points we ran our analysis.
[00:27:22.320]Jeffrey Stevens: And what we found is what we call the human face area maps on to the human fuser form area.
[00:27:28.830]Jeffrey Stevens: And the dog face area mapped on to the super triple gyrus if you're familiar with a face processing system in humans, this is, these are two common areas.
[00:27:38.190]Jeffrey Stevens: And the models and I don't want to you know, say that I think that these are functionally analogous across dogs and humans, but I think it's an interesting finding that we can explore more.
[00:27:50.880]Jeffrey Stevens: So i'm going to summarize what summarize what I just told you, is but we've got separate areas for humans and dogs all right, this is, you know I told you, the story about maybe it's due to.
[00:28:01.680]Jeffrey Stevens: domestication in pets that we saw that's different in terms of the studies going across labs, but we do see consistencies across three different labs and the cerebral cortex that's being involved in processing faces, so you know I want to take a moment to say like.
[00:28:17.640]Jeffrey Stevens: Your imaging and dogs is is that the star right and if you're familiar with neuroimaging in humans, there was tons and tons of problems at the beginning of it and.
[00:28:27.300]Jeffrey Stevens: All those problems exist with dogs right, so my my caveat here is, I feel that everything needs to be replicated I think we need to standardize.
[00:28:37.440]Jeffrey Stevens: Our processes more but we're we're at the Star and we're going to get there we're going to move there, I strongly believe on that, but we're still getting in really interesting finding so.
[00:28:49.260]Jeffrey Stevens: The other content in the middle of those are those are common areas for familiarity for human faces all right, those are also part of the attachment network positive and.
[00:28:58.800]Jeffrey Stevens: negative balance cotton nebula hippocampus can calm instructions We saw this really interesting connection between behavior outside of the scanner.
[00:29:07.320]Jeffrey Stevens: And behavior of the brain internally inside the scanner, and these are like the same stimulate right the same person and and now, now that was a really cool finding and something pretty novel about that study like I quite like.
[00:29:21.960]Jeffrey Stevens: Again we see these these structures are common the most common and mouse emotional arousal across different species as as the hippocampus involved with familiarity and memory.
[00:29:32.250]Jeffrey Stevens: In the in the Chi, it is also involved in in mediation mediation familiarity across a number of different labs will be looking at.
[00:29:42.030]The findings.
[00:29:44.190]Jeffrey Stevens: So we're seeing evidence for representing the social relationship between the trainer and the working Doc here.
[00:29:51.120]Jeffrey Stevens: You know the goal was not to kind of find the attachment network, but it but it's like this opens up the question like do dogs have the attachment network like humans do all right, is it do they have one is it similar is it different, how does it, how does it vary.
[00:30:05.970]Jeffrey Stevens: Now, and humans, we know a ton about this already so it's, this is just a recent review paper and just kind of highlighting structures involved in the attachment network and humor with humans with fmri data.
[00:30:17.970]Jeffrey Stevens: And those structures, I talked about before amygdala caught a big campus they're all there so so it's a alright.
[00:30:24.630]Jeffrey Stevens: So we know something's going on humans and there's been studies showing that humans right.
[00:30:29.970]Jeffrey Stevens: represent their dog right, similarly to their family members, when we look at attachment not not identical, but there, there are some differences, but the network's active right there.
[00:30:39.300]Jeffrey Stevens: So we know we get that but what about the dogs are they gonna have we're gonna have functional analogous regions or we're going to see a network of activity.
[00:30:45.840]Jeffrey Stevens: Clearly there's going to be something we all know, we bond with dogs in the attached right there's gotta be something there.
[00:30:52.440]Jeffrey Stevens: Right, so the prior slide I just went through they kind of let gave us some seeds for us to look at with the future and they weren't designed to say like Okay, can we look and study the attachment network as we move forward.
[00:31:04.890]Jeffrey Stevens: And so, this next set of studies is going to start to approach that question.
[00:31:13.020]Jeffrey Stevens: Okay, so we're pretty confident we're going to find something of interest and.
[00:31:19.410]Jeffrey Stevens: How we're going to do it is by building a bio behavioral profile right so we're we're going to do a we're going to look at behavior in different tasks we're going to look at surveys.
[00:31:31.590]Jeffrey Stevens: To the people that work with the dogs and i'll talk about the experiment in a second subjects got the hang out with dogs.
[00:31:39.210]Jeffrey Stevens: And we're going to build this we're looking at imaging data, we want to take it all together kind of look and see how these different factors are related to attachment in dogs and humans.
[00:31:54.270]Jeffrey Stevens: So the first phase of this experiment i'll call pre bonding and what we're going to do and i'll show you this a little more detail we're gonna do this cure based test.
[00:32:01.950]Jeffrey Stevens: You know modeled after the ainsworth's strange situation test we're going to do pair detachment right after we we we took a shortened version for monique noodles lab.
[00:32:11.370]Jeffrey Stevens: Who did this really, really well, and so we just pretty much copy BAT and we're going to ask a lot of questionnaires about.
[00:32:18.240]Jeffrey Stevens: to humans, we want to find out whether our personality differences and their feelings about dogs and, and these are all the question here is, but those are Those are some of them.
[00:32:27.900]Jeffrey Stevens: Then we're going to go into a bonding phase and what happens here is a human gets to hang out with a dog three hours a week for a month.
[00:32:39.030]Jeffrey Stevens: All right, they basically get to play, and this is all scripted or pretty much scripted it's like all right we're gonna give you different toys, we want you to play with this toy and do that in our a.
[00:32:47.940]Jeffrey Stevens: pet dog and rocketed and it's it's basically play with the dog all right and form a relationship right.
[00:32:55.770]Jeffrey Stevens: So, remember, these are all working dogs, you know their kin old and now they're going to get to go out and play right, so this is, this is a cool thing for them to do, also in the in the humans not surprising, they really like doing this all right to hang out with a dog.
[00:33:10.470]Jeffrey Stevens: We do surveys after each session right, so we asked them like how's the how's the experience going for you, with it, with the humans.
[00:33:18.150]Jeffrey Stevens: And we also collect saliva and urine samples at three points over this bonding at the beginning middle and end and we want to see how that data right will relate to the different tasks that I might talk about.
[00:33:35.550]Jeffrey Stevens: All right, and then at the end we do post bonding to do the secure base test that attachment test same thing over again, and several different questions.
[00:33:46.620]Jeffrey Stevens: Okay, so secure base test the the way this works is, you have the first phase each phase last two minutes.
[00:33:53.130]Jeffrey Stevens: The dog and the human are in the room together at the dog enters the semi circle around the human.
[00:33:58.470]Jeffrey Stevens: The human can pet it a couple times all right, and so you measure how often they kind of spend time with the human there's toys in there, they go over to the play with the toys.
[00:34:07.830]Jeffrey Stevens: Phase two is the human leaves the room dog stays by themselves right and then.
[00:34:13.530]Jeffrey Stevens: Phase three the human comes back the reunion phase, because the dog spend a lot of time next to the human or not and and people use these measures to kind of categorize the dog as being securely attached or insecure.
[00:34:31.560]Jeffrey Stevens: Apparently attachment test is essentially you take two people and you put them in the room, together, and this happens, right after secure based test so that ends and you have a minute the dogs by themselves in the room, by themselves, and then the two humans come in.
[00:34:47.010]Jeffrey Stevens: All right now in Phase one there's a familiar and unfamiliar person there.
[00:34:51.210]Jeffrey Stevens: Now, at the very beginning at pre bonding they're both unfamiliar all right, but the, but in this case the person left is going to actually be the person that gets on it, but there's no relationship at this point in time, except for the fact that the familiar person.
[00:35:04.650]Jeffrey Stevens: was in the secure base test, so there is a little bit of relationship there, right from the just that but six minutes.
[00:35:11.130]Jeffrey Stevens: four minutes really and then and then sort of the staff so we're not really expecting to see a difference their.
[00:35:16.740]Jeffrey Stevens: Face so basically in Phase one if the dog comes into the semi circle, you can say dog a couple times and then that's it right and then.
[00:35:26.190]Jeffrey Stevens: stay there are doing it once phase two is you can continuously hang out and pet you can pet the whole time caller's name and it's you know, much more so it's called a passive phase one and.
[00:35:39.780]Jeffrey Stevens: phase.
[00:35:42.750]Jeffrey Stevens: All right.
[00:35:44.640]Jeffrey Stevens: So you do the pre bonding you do the bonding do the post bonding after that.
[00:35:50.580]Jeffrey Stevens: You get to have your brain scanned.
[00:35:54.660]Jeffrey Stevens: And so, not only do you as a human participating, but the dog goes to the scanner to now, at the same time, but.
[00:36:02.220]Jeffrey Stevens: You know they're they're doing it in what we did here is.
[00:36:07.830]Jeffrey Stevens: The same kind of tasks I described before where it's a picture viewing right now we have three types of stimulating we have the affiliate person is the person that's bonded and played three hours a week.
[00:36:18.120]Jeffrey Stevens: We have the completely unfamiliar human so no experience with that person, and we have a familiar human so this person.
[00:36:24.030]Jeffrey Stevens: got to hang out and play with a dog five minutes per week so it's kind of a control for familiarity right, so we when we contrast affiliated versus unfamiliar we want to make sure it wasn't just a familiarity different So this was able to kind of control for that.
[00:36:43.860]Jeffrey Stevens: Now, when the humans go in, they don't look at pictures of humans, they look at pictures of dogs right so it's it's the same same logic.
[00:36:51.270]Jeffrey Stevens: The dog they're bonded with the familiar one than the unfamiliar one and a lot of either they're almost all dogs are black labs in the study, so they look similar.
[00:37:00.780]Jeffrey Stevens: Right, but they have different colored callers that helps to help them to identify a minute once you hang out with a dog long enough you start to recognize recognize a dog, but the first time you look at them if you didn't know them you you couldn't be hard to tell.
[00:37:17.160]Jeffrey Stevens: The other type of scan we do is we let them watch a video in the video goes on for five minutes right, so they watch them those bonding sessions, I was talking about right we basically take five minutes of video from that.
[00:37:30.450]Jeffrey Stevens: And you sit there, and you watch it right and we do that for the three conditions, the affiliate the familiar and the unfamiliar, and this is, I will not play a whole five minute video because you know it's you playing with the dog, but we can.
[00:37:53.310]Jeffrey Stevens: I should say we've got we had a big big area where we had like six different little pins set up and so we've got we're doing six dogs at one time.
[00:38:03.060]Jeffrey Stevens: Efficient to do it that way, but but that said you're doing that you're just watching this video for five minutes, the dog watches the exact same video alright so you're seeing the exact same thing we're going to ask questions about that relationship we do our analysis.
[00:38:21.810]Jeffrey Stevens: Now, at this point you've probably been saying, where are the robots.
[00:38:27.360]Jeffrey Stevens: Maybe you've been saying that I don't know it goes here like robots and Brian you're probably saying they were the robots i'm waiting I don't know like so like anyways.
[00:38:35.820]Jeffrey Stevens: Where the robots alright, so, while this experiment was going on there was this event that happened, it was called the super bowl you guys you've probably all heard about this thing beard award in nebraska you guys watch super bowl here, or you just do college.
[00:38:51.300]Jeffrey Stevens: i've scary so it's like because i'm not you know, an auburn we have our pro football team to, and you know it's sickening thing alright so.
[00:39:00.510]Jeffrey Stevens: i'm watching the game you won't watch the game, this year, maybe in another commercial about to tell you about.
[00:39:07.440]Jeffrey Stevens: So this was a commercial got played.
[00:39:10.440]Jeffrey Stevens: And you know we nicknamed our pro our project, the robo dog project right so but sure enough during the superbowl guess what comes out.
[00:39:20.580]Jeffrey Stevens: robot dog.
[00:39:22.710]Jeffrey Stevens: And this is robo Doc alright, so there is this commercial.
[00:39:29.850]Jeffrey Stevens: it's for an electric car i'm not going to tell you what car because I don't want to give them and I don't want to play the video and give them a free commercial but here's this guy charging up his car all right, you can see robo dogs looking out the window.
[00:39:42.570]Jeffrey Stevens: At the guy charging for this car and immediately robo dogs, like, I want to be with this man, you know they he forms an immediate bond already attached to this guy just right away.
[00:39:54.990]Jeffrey Stevens: So.
[00:39:56.490]Jeffrey Stevens: The guy pulls off in his car and robo dog burst out the store starts running down the serie chasing the car, I mean super good next scene, you see his robot dog is jumping from building to building leaping you know in single bounce you know amazing and as robo dog is doing this.
[00:40:15.120]Jeffrey Stevens: he's jumping up and that is power goes out.
[00:40:20.010]Jeffrey Stevens: Now we don't see robot dog fall to the ground and smash right, but what we do see next.
[00:40:30.570]Jeffrey Stevens: Is the guy driving the car gets robo dog notice the hand on top of robo dogs already formed a bond with him Jackson into his electric car and charges them up.
[00:40:44.070]Jeffrey Stevens: All right, so they're there they're forming this great relationship right and the next scene, we see.
[00:40:51.360]Jeffrey Stevens: Is you know robo dog notice he's all charged up a little green thing and his chest right and he's hanging out the window loving it just like your your pet dog at home in the car doing all that stuff and they've developed this amazing relationship and just a matter of a few minutes.
[00:41:08.550]Jeffrey Stevens: And you can own robo dog for $299.
[00:41:15.690]Jeffrey Stevens: Maybe in 5200 years because the real robot dog guess what he costs.
[00:41:23.220]it's more.
[00:41:24.990]Jeffrey Stevens: So so like we work with this guy.
[00:41:30.240]Jeffrey Stevens: So this is sony's Ibo and he's really cute too he doesn't chase over after cars actually does he's just kind of really slow and they get away pretty fast so it's not very effective, he doesn't leap over anything.
[00:41:45.390]Jeffrey Stevens: But he has a ball, he can kick it and you know he can sing and dance, you know it's kind of cool and he only costs about $3,000 so you know so it's.
[00:41:56.490]Jeffrey Stevens: It was kind of cool I got to live with robo dog and my son, and I for a month over the summer time to kind of see like oh what's it like living with a robo dog right do you actually form a relationship with this robot.
[00:42:10.320]Jeffrey Stevens: Can you kind of do all right it definitely happens, but it's you know it's not a real dog for sure it's not that same kind of relationship, but.
[00:42:19.050]Jeffrey Stevens: i'd be sitting there working at my computer and all of a sudden there's a little robot dog His name was toshi actually so toshi comes up and like.
[00:42:25.710]Jeffrey Stevens: he's rubbing on my leg he's whining wants me to play with them, you know sit there and play me a song and do a dance I get my attention, and so you need to feed them.
[00:42:35.280]Jeffrey Stevens: right there times, he wants to be fed a little APP that comes with them and Jim tricks and high fives and you know it's you know it's kind of amusing, you know so it's but you definitely form a relationship with your your robotic dog.
[00:42:50.940]Jeffrey Stevens: So he barks he winds he dances he does these amazing tricks sing songs happy and you know it clap your hands, I mean that's one of his favorite very he does a waltz and dances to it.
[00:43:01.710]Jeffrey Stevens: Even urinates la like in the middle of your carpet he'll pick up his leg, and nothing comes out with the sound effect is there.
[00:43:10.110]Jeffrey Stevens: And you get that and they have personalities, when you when you start with robo dog right it's a it's kind of like a blank slate right the way you interact with that dog.
[00:43:21.810]Jeffrey Stevens: forms the personality that dog so so many claims right so unlike.
[00:43:27.000]Jeffrey Stevens: And there's some software they're kind of helps but it's like they definitely do have different personalities i've hung out with several of these guys.
[00:43:32.670]Jeffrey Stevens: And so, like as you as you interact with Robert dog you help to shape it.
[00:43:37.020]Jeffrey Stevens: Right is it is it needy or not doesn't need attention and you're doing that right, and you can you can do bad dog and scolded and all that stuff if you feel the need to do it so so they do, develop a personality, as you work with them.
[00:43:55.560]Jeffrey Stevens: All right, so.
[00:43:58.260]Jeffrey Stevens: After you've done this experiment with the bonding with a real dog you get to do the whole thing over with robo dog.
[00:44:05.880]Jeffrey Stevens: All right, in actuality this is counterbalance so some people will get robo dog first and then get the real Doc second So the question is, what kind of relationship does you know your subject get with robo dog.
[00:44:18.780]Jeffrey Stevens: Right and you can see, like the subjects are smiling they're having a good time it's not staged all course you know so.
[00:44:25.920]Jeffrey Stevens: And the cool thing when you get get robot dog you get to name them.
[00:44:29.520]Jeffrey Stevens: Right, I think that's tater tot on the left and melvin on the right, you get to give them a little you know if you want to call her a bandana so you can personalize it and and form this this relationship with your robotic dog, and you know training teachers do tricks hang out pet them.
[00:44:47.910]Jeffrey Stevens: And entertain yourself for an hour three times three times a week for for three hours total a week.
[00:44:56.880]Jeffrey Stevens: Now, when the human goes in the scanner they don't look at pictures of real dogs they look at pictures of.
[00:45:02.400]Jeffrey Stevens: The subjects and the experiment, and so there were there were six different ones six different high bows and you know they they're here, you can see, is the same exact logic it's going to be seeing contrast, I described before for the human dogs.
[00:45:18.180]Jeffrey Stevens: One of my you know students that work was like, why are we going to scan the Ibo you know, like, why are we doing a robo dog, and you know it was like well you know that's not gonna work out really well so.
[00:45:29.610]Jeffrey Stevens: i'm glad he didn't take it over to the to the last week, I would have been disaster, of course, but but uh but the cool thing you can do is you can download the software, if you work with the Sony really well kind of see how that's been developed.
[00:45:45.780]So.
[00:45:47.340]Maybe.
[00:45:50.850]Jeffrey Stevens: So in the study there's there's 20 pairs alright, so we have 20 humans with with 20 different dog pairs they're actually 11 dogs, so they.
[00:45:59.760]Jeffrey Stevens: They we had one dog drop out, he was that dog I was telling you about that didn't want to scan anymore, it was very annoying so we had to get another dog in there.
[00:46:07.740]Jeffrey Stevens: And he was a great dog, to start with bell and like right away would get get his brain scan and then, just like no, you know what it is.
[00:46:17.130]Jeffrey Stevens: So he had these pairs it's a like a sales counterbalanced you we've got 20 pairs for humans and and the dogs and 20 parents for the humans and the eyeballs.
[00:46:26.790]Jeffrey Stevens: And then we have these runs I was talking about you get the videos you get the human faces the humans also do a matching task which I didn't tell you about, but if you're familiar matching the sample it's basically that with with with the pictures of different categories.
[00:46:45.960]Jeffrey Stevens: Alright, so let me show you just what the secure base test looks like for the real dog and can I bow because what the real dog first.
[00:46:56.760]Jeffrey Stevens: All right, so who do the real.
[00:47:03.300]Jeffrey Stevens: Am I going to play a whole.
[00:47:06.090]Jeffrey Stevens: You can see the dogs just sitting in.
[00:47:09.240]Jeffrey Stevens: He doesn't want to move.
[00:47:24.210]Jeffrey Stevens: So measuring how much time is it.
[00:47:28.920]Our coops and then you can get a sense of of.
[00:48:00.480]I was actually urinating sound.
[00:48:22.500]Jeffrey Stevens: Alright, so that gives you a sense of what's going on now, when you analyze and I have to say this is this experiment just finished.
[00:48:31.170]Jeffrey Stevens: At the end of last semester, and so I only have a little bit of data to share with you you're gonna have to stay tuned for some of the all the brain imaging all that stuff but it's coming.
[00:48:40.470]Jeffrey Stevens: Alright, so if I wanted to share this is all about, I was really motivated to talk with you guys about what we're doing with bonding, and so I wanted to at least show you where we are the President will get to talk about the future, to a little bit.
[00:48:55.710]Jeffrey Stevens: So when you categorize the the real dogs right for for what their attachment style is it's about 68% are secure and there's the the other ones that are insecure.
[00:49:10.380]Jeffrey Stevens: This is pre bonding and then we're like does it change over the month period of bonding.
[00:49:16.950]Jeffrey Stevens: and not so much and that's not too surprising, these are these dogs I should have told you, there were like five and six years old and four years, all the you know, a month period, with one person is not going to.
[00:49:27.000]Jeffrey Stevens: change their style too much and what you know some of the dogs went into the insecure because they just kind of want to play with their toys more.
[00:49:34.590]Jeffrey Stevens: So it's but, but it is what it is now the interesting thing, so I have, I have some data, I want to share from monique slab here.
[00:49:43.080]Jeffrey Stevens: She did a big Meta analysis on pet dogs and their their styles of attachment.
[00:49:49.110]Jeffrey Stevens: And what you'll see is that the the working dogs right at Camp performance sciences they're very similar to the pet dogs in terms of their attachment style right, so you know.
[00:49:59.940]Jeffrey Stevens: 68% and 68% right so that was pretty cool and and in her lab he did a really interesting study looking at shelter dogs and foster dogs.
[00:50:09.510]Jeffrey Stevens: And you can see, the increase going from from the shelter to the Foster actually increased the attachment style, which was really great because it shows that you can you can help your dog pretty quickly here.
[00:50:21.960]Jeffrey Stevens: Alright parrot attachment test, by the way, I should I don't have the the robotic dog attachment styles yet So those are those are those are coming i'm sorry.
[00:50:33.480]Jeffrey Stevens: So apparent attachment test let's take a look at a few of those and we'll do the real dog here.
[00:50:41.400]Jeffrey Stevens: This is the first two minutes.
[00:50:45.390]Jeffrey Stevens: Just do a quick pat stop.
[00:50:52.620]Jeffrey Stevens: once more.
[00:51:02.160]Jeffrey Stevens: So we're going to look at do they spend more time with the familiar the.
[00:51:08.670]Jeffrey Stevens: Man, you know he picked up his.
[00:51:17.400]Jeffrey Stevens: passive.
[00:51:20.820]Jeffrey Stevens: To face right.
[00:51:37.110]Now.
[00:51:57.240]Jeffrey Stevens: Now let's look at robo dog in action.
[00:52:28.020]Out there, let me show you the whole two minutes.
[00:52:31.980]Jeffrey Stevens: starts moving around.
[00:52:35.070]In the eventually stops.
[00:52:39.480]Jeffrey Stevens: years the active fee.
[00:52:44.820]Jeffrey Stevens: that's tater tot.
[00:52:51.900]he's enjoying it.
[00:52:57.960]Jeffrey Stevens: And so that's what you measure now.
[00:53:02.730]Jeffrey Stevens: So new data to share here, so if we look at this is, you know, again, this is the real dog or an attachment.
[00:53:08.160]Jeffrey Stevens: This is pre bonding so remember the, as I said earlier, to you before the familiar and unfamiliar The only difference here is the familiar person was in the secure base test.
[00:53:18.060]Jeffrey Stevens: For total of you know, four minutes interaction, but yeah the dog has a slight preference to go and hang out and check out the unfamiliar person little bit.
[00:53:27.480]Jeffrey Stevens: All right, and we weren't really expecting that the the data, where the course, but at this point we were just spreading it to be equal it wasn't.
[00:53:37.980]Jeffrey Stevens: After bonding it gets larger so it's so it was like what's going on here, it seems like the dog Brian is wants to check out new things.
[00:53:50.790]So.
[00:53:52.530]Jeffrey Stevens: Maybe that's not too surprised in the big picture for working dogs, you know that's part of what they were their job can sometimes do us to explore stuff so it's a.
[00:54:00.450]Jeffrey Stevens: So I was surprised that it would happen so quickly after just a really quick amount of time, but these again, these are you know older dogs, they have a lot of experience so.
[00:54:08.520]Jeffrey Stevens: So I thought that was an interesting finding I don't know what we're going to see where the eyeballs yet, but we will find out.
[00:54:15.120]Jeffrey Stevens: Now we did some some survey work, and this is the lexington attachment pet scale.
[00:54:23.400]Jeffrey Stevens: Maybe not a big surprise here but humans get more attached to the real dog than the robotic dog right so Okay, so you know okay That was good to someone good to know and there's no order effect that matter which one you got first.
[00:54:37.440]Jeffrey Stevens: But this is my first time using the laps and I was like you know some of these questions are not real they're more like centered for pets, not for the working dogs and robots so I was like let me look at some of the individual questions.
[00:54:49.770]Jeffrey Stevens: Maybe there'll be more insightful maybe more favorable and I should say the score for the eyeballs it's not like they didn't form an attachment right they still did it's you know it's not super low right it's just not as much as the real.
[00:55:04.320]Jeffrey Stevens: Okay, so here are some individual questions we'll just you know question one my partner, and I have a very close relationship so you know with someone agree with the real dogs somewhat disagree with Ibo and then i'm not very attached to my partner.
[00:55:22.440]Jeffrey Stevens: Well, you know somewhat disagree, for I hate these negative questions because they always screw up when you got to score him, but you know it is what it is.
[00:55:29.460]Jeffrey Stevens: And and, by the way, the real say says i'm not very attached to my pet right but we changed it to partner, because of the situation and, in the last question.
[00:55:38.700]Jeffrey Stevens: I will miss my partner when the study is over, and you can for the real dog, you know most you know they're strongly agree to a lot of people's strongly agree.
[00:55:46.410]Jeffrey Stevens: But nobody was below someone agree, and you know, and I both do you know somewhat disagree there because it's just not as strong, so you know, the bottom line is the overall mean scores to sell you the same story it's like they don't attach a strong like they're real dogs so sorry diagnose.
[00:56:03.390]Jeffrey Stevens: Okay there's also the scale, which was awesome to me how you know, do you like your robot you know robot likability, how do you feel about that and.
[00:56:13.440]Jeffrey Stevens: It didn't really i'll show you a few questions, but it really didn't change between pre and post bonding so like working with an Ibo for four months and make them all of a sudden, I like robots more right anything went down a little bit so but not simply do.
[00:56:29.040]Jeffrey Stevens: So, again, I was, like some of these questions are like would you talk to your robot for advice, you know that i'm not super relevant here, so I was like let me look at some individual questions so is the robot friendly So you can see, you know.
[00:56:42.450]Jeffrey Stevens: Pre employment, you know it's like six you know five to six, you know net and sevens like strongly agree right and one is like strongly disagree, so you figure four is like right in the middle right, so you know that's not too bad you know it's friendly.
[00:56:55.950]Jeffrey Stevens: isn't likable right about the same score, but you know it's not changing over time and then.
[00:57:02.640]Jeffrey Stevens: This robot is warm this one actually significantly goes down so they're not so favorable for Ibo but and then this last one I got for you this robot is approachable so that hasn't really changed so you know so there's so.
[00:57:17.790]Jeffrey Stevens: Why are we looking at this, let me so like.
[00:57:21.240]Jeffrey Stevens: we'll talk about about robots in a second, but let me just give you overall future future directions right.
[00:57:27.390]Jeffrey Stevens: it's a very rich data set right we've got a lot of behavior to look at we've got all the fmri data to get through.
[00:57:33.210]Jeffrey Stevens: That we want to do these new analysis on brain to brain correlation so those movies, that you watch that I talked about where you watch yourself playing with the dog.
[00:57:42.240]Jeffrey Stevens: we're going to analyze that the activity over the five minutes we're gonna do the same thing with dog and we're going to ask is there a relationship between the oscillations that happen through the through the brain wise they watch will there be any kind of.
[00:57:56.430]Jeffrey Stevens: connection there that we can we can talk about when to take oxytocin scores and score you know correlate that with all our data and the questionnaires will all this stuff in and build our bio behavioral profile.
[00:58:10.170]Jeffrey Stevens: Alright future direction so here's the high, both through the years and, as the latest version.
[00:58:17.220]Jeffrey Stevens: robots are here to stay they're not going anywhere right they are becoming more and more part of our society right and you if you haven't interacted with a robot yet you will at some point right it's just it's not changing and there's this new field, I think it's actually pretty exciting.
[00:58:35.520]Jeffrey Stevens: Some of you will recognize the name a little bit, I want to tell you this data point first.
[00:58:42.060]Jeffrey Stevens: So if you ask humans right.
[00:58:45.630]Jeffrey Stevens: How they feel how comfortable, they are with a robot right, so you have it like human likeness industrial robot you know, like big arm putting you know, like caps on bottles.
[00:58:55.230]Jeffrey Stevens: yeah i'm not so comfortable with that Okay, but as the robot gets more human like right you get more comfortable right.
[00:59:05.640]Jeffrey Stevens: And then you get this big tip this last data point is basically like you can't tell the difference between the robot and a real person.
[00:59:14.070]Jeffrey Stevens: Right it's it's like the turing test right or maybe you know, maybe we'll be at the point where it's you know Blade Runner or something like that, who knows where we're going in the future, but.
[00:59:22.410]Jeffrey Stevens: right before we get there there's this uncanny Valley, so the robot is really close to a human.
[00:59:28.590]Jeffrey Stevens: But not a human and that people don't like that all right, that makes them feel weird.
[00:59:33.150]Jeffrey Stevens: Okay, so it's a, so this is an issue as we move forward potentially in the future with our robots right, so this is the new field before robotics you don't recognize the name being dog people that you are.
[00:59:44.460]Jeffrey Stevens: And what either robotics does is it it combines the principles of mythology in ecology and evolution, to look at the development of social robots.
[00:59:54.450]Jeffrey Stevens: Alright, so so as we kind of start communicating more with robots and putting them in our society we're going to be evolving with them.
[01:00:02.010]Jeffrey Stevens: Just like we evolve with dogs and doing through domestication the same thing is going to happen with robots potentially right we're going to build these relationships with them and what's the trade offs going to be as we move forward.
[01:00:16.890]Jeffrey Stevens: So why dogs well they're an excellent prototype for all this right all the work that everybody's doing with human, animal interactions we can learn from that.
[01:00:26.640]Jeffrey Stevens: and pull from that right to to create robots right to eventually hopefully avoid the uncanny valley right so as we move forward, so I think this is a really exciting future for us as we, as we go down the road.
[01:00:44.400]Jeffrey Stevens: So I want to thank my collaborators Dr frank kruger at George mason he's the one that got me into the robot dog stuff and working with him.
[01:00:52.200]Jeffrey Stevens: People at the MRI Center go be despond a colleague of mine and we we've been he does all the complex analysis for the neural imaging company our director at the.
[01:01:02.160]Jeffrey Stevens: MRI Center without him, none of this could ever happen because the the hours go way beyond what you what you can do and it's it's awesome that he works with us, of course, cannot perform and sciences.
[01:01:14.580]Jeffrey Stevens: directors, Dr Wagner and singletary angle and Dr loss roski and Dr quick mom and they're amazing trainers Bart Roger and amanda reeves my laboratory.
[01:01:24.840]Jeffrey Stevens: This is a Lucy allows roski myself and him to do a quick bomb macaques and Jordan Smith and there are tons of undergraduates that work with us over time and, lastly, I would like to thank these guys.
[01:01:39.750]Thank you for your time and attention.
[01:01:47.160]Jeffrey Stevens: Right, thank you, Dr katz come on up for your questions folks.
[01:01:56.670]I have two questions that allow.
[01:01:59.550]Jeffrey Stevens: Okay first question is might be dumb question, I know that handedness is important in human fmri is pop preference or handedness important dog fmri.
[01:02:08.700]Jeffrey Stevens: That is a good question I don't I don't know if anybody's even looked at that.
[01:02:14.580]Jeffrey Stevens: Certainly, for what we've looked at preferences why we've never done any correlations with our data, so we have that data we could do that so.
[01:02:22.380]Jeffrey Stevens: There might be some interesting asymmetries we did find some asymmetries with some of the emotional content to know is, in terms of contrast, and always map up with.
[01:02:31.230]Jeffrey Stevens: what's known about it, like positive being less hemispheres left hemisphere negative right but it'd be really interesting to kind of look at that and see if there's a relationship.
[01:02:41.040]Jeffrey Stevens: And second question I don't know if Adam a close he talked about it in that paper that you referenced do we have any evidence to suggest or not suggest that.
[01:02:48.780]Jeffrey Stevens: Humans would have that uncanny valley feeling with a dog robot as well as even we don't have any evidence of that it's hypothetical right but it's a but I don't see why not, I mean it should be the same similar principle but it's it's an open question.
[01:03:16.290]Jeffrey Stevens: So I have a question about your attachment finding so when you were showing the data, where the dog.
[01:03:23.550]Jeffrey Stevens: Pre bonding seemed to approach the unfamiliar person and then that increased after the bonding to me, I thought well that makes total sense within the.
[01:03:35.820]Jeffrey Stevens: attachment paradigm i'm thinking well if it's true that they oscillate between attachment seeking behavior.
[01:03:44.670]Jeffrey Stevens: Proximity seeking behavior versus exploration coming into the scene, they already you know see that person that they've had some exposure to.
[01:03:54.120]Jeffrey Stevens: So they go and seek out the stranger if they're secure and then that would increase over time after they've bonded but you seem to say well.
[01:04:03.540]Jeffrey Stevens: That you thought it may be, has more to do with the personality of the dog, so my question was or the working relationship of the dog or the type of dog.
[01:04:13.890]Jeffrey Stevens: So my question is did you separate this out by.
[01:04:18.090]Jeffrey Stevens: sort of any kind of classification of their attachment style like did you notice differences right no so that's a great question so first of all my.
[01:04:26.850]Jeffrey Stevens: My thinking was really just based on monique's paper where were there, it was a different type of effect so so that that's where that came from.
[01:04:35.250]Jeffrey Stevens: But that that's actually something i've been wanting to do is that we just haven't got there yet, look at their attachment style How does that actually relate to their seeking behavior.
[01:04:50.430]Jeffrey Stevens: I good question first of all thank you so much is also fascinating I love the implications so Actually, I have a question on like future directions implication kind of stuff so something that.
[01:04:59.880]Jeffrey Stevens: We found in our API study or I guess found but we've like kind of anecdotally seen is that therapy dogs can almost be more or less helpful.
[01:05:09.600]Jeffrey Stevens: Sometimes, depending on how they actually physically feel like we're attracted to something fluffy you can really dig your hands into versus opposite.
[01:05:17.220]Jeffrey Stevens: And that reminded me so much of like the attachment here, I was wondering, have you ever thought of like just adding like for covering or something a little softer to the robots and actually.
[01:05:27.930]Jeffrey Stevens: Possibly the attachment might actually shoot through the roof, because what I was noticing those videos is the ones that were working with the robots it was very command performance based.
[01:05:37.500]Jeffrey Stevens: Whereas at least just the videos that we saw today with the pet dogs, it was not command based it was just co time hanging out, I was wondering if you had any.
[01:05:46.260]Jeffrey Stevens: Things speak on that yeah I mean I think you're I think that's probably going to would happen, I think the physical characteristics are going to be really important to developing you know bond that's definitely one of the parameters that can be manipulated to potentially increase attachment.
[01:06:02.280]Jeffrey Stevens: So yeah I mean, I would say yeah that would be interesting to do.
[01:06:04.920]Jeffrey Stevens: it's you know we basically like all right let's just try this for the first time and start here and we'll just see how it comes out and then we'll start thinking about parameters that we want to manipulate in the future, but yeah That would be a good one.
[01:06:30.210]Jeffrey Stevens: Research, such as gender differences that had an effect on who the dog bonded with regardless of familiarity so, can you I didn't catch the first part of it, yes.
[01:06:41.070]Jeffrey Stevens: So in the aspect of attachment and bonding.
[01:06:44.970]Jeffrey Stevens: Were there any underlying factors you found within your research, such as gender differences within the humans.
[01:06:51.810]Jeffrey Stevens: That had an effect in who the dog bonded with regardless of the familiarity yeah so it's a it's a great question I didn't mention this, but we purposely only use female participants.
[01:07:02.850]Jeffrey Stevens: Right, because we were concerned that one couple of factors went into it, but there could be differences between males and females, and how they form this attachment.
[01:07:12.570]Jeffrey Stevens: And we just wanted to take that factor off the table for this particular experiment, so we just did female subjects.
[01:07:19.350]Jeffrey Stevens: is also a practical decision right so in part of it is with only 27 you're not going to have tons of power to kind of start looking at gender differences, for instance, so so that was the That was the reason and practically it's.
[01:07:32.070]Jeffrey Stevens: You know you're recruiting from the psychology pool, it was in we first started this over the summer time we were just like it was easier to get females so we're like let's just go with that and put that into our design, but it's definitely an important thing to look at in the future.
[01:07:50.640]Jeffrey Stevens: Maybe to follow up on that.
[01:07:54.180]Jeffrey Stevens: Did you Code, the interactions themselves right, so you gave instructions previously right but yeah did you Code, the interactions and have you seen anything yet so so we have, so we have all the playtime.
[01:08:07.980]Jeffrey Stevens: recorded every session we're going to build an ether gram of that and so that hasn't been completed yet, but that's for that reason Jeff is that we're going to do that, to look how that relates for sure.
[01:08:21.600]Jeffrey Stevens: Then we have a question from zoom they say thanks for the interesting talk, do you think it will be possible to overcome the uncanny valley anytime soon with a robot modeled after a dog.
[01:08:32.370]Jeffrey Stevens: And they say what i'm getting at is this perhaps attachment bonds could be formed faster with a robot modeled after a meetup species.
[01:08:39.450]Jeffrey Stevens: In that case we won't necessarily compare with the bonds we form with dogs, I think there is the potential for it to go faster with robots that it is that are not human like because.
[01:08:53.580]Jeffrey Stevens: With humans, you know you can't do things like oh let's expose them, you know a little babies to all these different types of variables right, and you know.
[01:09:01.710]Jeffrey Stevens: My concern unethical, but for robotic dogs, for instance, you can create them, however, you want to right, so you can actually potentially speed up that process.
[01:09:11.730]Jeffrey Stevens: and avoid the uncanny Valley, a lot sooner, like, for instance, taking the physical characteristics and improving on that right, you can get that.
[01:09:19.080]Jeffrey Stevens: you'll understand the the social interactions right, how is that different than a real dog right, and then you reprogram based on that and hopefully yeah I think it can go a lot faster.
[01:09:31.350]Jeffrey Stevens: Maybe a follow up to that do you is the uncanny valley focused on physical appearance, or is it behavioral how to So if you acting like it's all about it's all of it, so that it's not just physical appearance it's it's how they behave.
[01:09:47.640]Jeffrey Stevens: Great well, I think that wraps it up, so thank you please join me in thanking Dr katz first talk with you.
[01:09:58.080]Jeffrey Stevens: Excellent well we've wrapped up this be the the presentations for today so we've been.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Tags:
canine
dog
fMRI
human-animal interaction
robot

Comments

0 Comments

Dog, human, and robot bonding: Past, present, and future

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments