Online research with chil­dren with and with­out an intel­lec­tu­al disability

Kather­ine Ellis:
Hi, every­one. My name is Kather­ine Ellis. I’m a research fel­low in the Neu­rode­vel­op­men­tal Dis­or­ders Lab, which is led by Dr. Jo Moss at the Uni­ver­si­ty of Sur­rey. I’m going to be talk­ing today about how to run research online with a par­tic­u­lar focus on chil­dren with an intel­lec­tu­al dis­abil­i­ty using eye-track­ing studies.

Kather­ine Ellis:
A bit of back­ground about what we do in the lab. So we are inter­est­ed in genet­ic syn­dromes that are asso­ci­at­ed with an intel­lec­tu­al dis­abil­i­ty, but we’re more inter­est­ed in them because of their dis­tinct pro­files of autis­tic traits and these dif­fer between one anoth­er and also those with [broad 00:00:39] syn­dromic autism. So what we do is look at vari­abil­i­ty at the social cog­ni­tive lev­el across these groups, which might explain per­haps the dif­fer­ences in autis­tic traits, with the hope that this work is going to help to lead to the devel­op­ment of more indi­vid­u­al­ized sup­port for these groups.

Kather­ine Ellis:
So one of the big chal­lenges of work­ing with these pop­u­la­tions, which are already real­ly under­rep­re­sent­ed in research, is find­ing tasks that can give us a real­ly good deep phe­no­typ­ing, social cog­ni­tive, phe­no­typ­ing infor­ma­tion, but also reli­able and valid, but are appro­pri­ate for peo­ple with quite lim­it­ed lan­guage, atten­tion and motor skills. And this is where eye-track­ing has been real­ly use­ful. So we’ve got a cou­ple of eye-track­ing tasks that look at social cog­ni­tive abil­i­ties, such as today’s fol­low­ing task. And this works real­ly well in all groups because it’s a com­plete­ly pas­sive par­a­digm. They don’t have to do any par­tic­u­lar tasks, but we can get real­ly nice, detailed infor­ma­tion about what they’re look­ing at. And this helps us to under­stand how they might be pro­cess­ing the world around them with­out them hav­ing to tell us in language.

Kather­ine Ellis:
Pri­or to the pan­dem­ic, we were doing this face to face using a Tobii lap­top, and this was work­ing real­ly well. It was trans­portable. We could take it to peo­ple’s homes. We could sit on the floor with the kids. And we were find­ing that the par­a­digms that we were using were work­ing with the neu­ro-typ­i­cal con­trol groups, and we were start­ing to find some real­ly nice, inter­est­ing ear­ly online research in these groups. Obvi­ous­ly, with the pan­dem­ic, it’s just not eth­i­cal to do these sorts of pop­u­la­tions, so this is what gave us the push, as many peo­ple to work and see whether we could do this online. So we’ve been using the Goril­la web eye-track­ing zone, which, for those of you who don’t know, uti­lizes WebGaz­er. So this uses par­tic­i­pants’ own web cam­era to detect their face, and then uses this infor­ma­tion to give an esti­mate of where peo­ple are look­ing on the screen.

Kather­ine Ellis:
And what you might notice is that I haven’t just got that icon for Goril­la, I’ve also got an icon for Teams, and this is because for these pop­u­la­tions, being able to com­mu­ni­cate with par­ents is real­ly, real­ly crit­i­cal. So yes, in the­o­ry, we could just send links off to par­ents and say, “Give it go with your kids. See if you can get them to watch a video and do the cal­i­bra­tion.” But for these par­tic­u­lar pop­u­la­tions who have a lot of addi­tion­al needs, and often you need quite skilled researchers to get data face to face, when you’re mov­ing every­thing online.

Kather­ine Ellis:
Actu­al­ly, what our role as researchers has moved to is actu­al­ly being able to coach the par­ents to do our job essen­tial­ly, to be the researchers. And this is sort of the key point that I real­ly want to empha­size. So what we ask par­ents to do, if they’re lucky enough to have a sec­ond device, is to set that up with a Teams link, and then what we’ll do is coach them through set­ting up every­thing on their lap­top, come up with a plan before the child has even come into the room, to make sure that we try and make every­thing run smooth­ly as pos­si­ble, because the aim is to make sure that the child is sit­ting for as lim­it­ed amount as pos­si­ble. And we don’t want them to sit there for any longer than is necessary.

Kather­ine Ellis:
But it’s not just sort of dur­ing the test­ing ses­sion where we will do this, actu­al­ly com­mu­ni­ca­tion is real­ly vital through­out the whole research process. And it starts from per­haps your first recruit­ment call. And this is where you’re real­ly start­ing to estab­lish a part­ner­ship with that par­ent, and you estab­lished a strat­e­gy that you’re going to work with before you actu­al­ly have the actu­al test­ing ses­sion. Every child is dif­fer­ent, and it’s talk­ing through, “What’s going to moti­vate them? What’s going to make them feel the most com­fort­able? And what’s going to help the par­ent in that situation?”

Kather­ine Ellis:
Fol­low­ing for that, you want to pro­vide some resources over email, so this is where you might pro­vide some more detailed infor­ma­tion about things like inter­net speeds. But what’s real­ly been help­ful at this point has been send­ing par­ents an eye-track­ing dum­my link. So actu­al­ly they can have a go at the eye-track­ing them­selves, see how it works, see what does­n’t work. And this helps par­ents feel a lot less anx­ious about what hap­pens in the actu­al test, say, because they feel they under­stand what’s going to hap­pen, even if it is rel­a­tive­ly straightforward.

Kather­ine Ellis:
And then hope­ful­ly by that point, by the time that you get to the test­ing ses­sions, the par­ents feel sup­port­ive, you’re there for any tech issues, and it’s just ensur­ing that there’s a calm envi­ron­ment to make sure kids feel relaxed and par­ents feel relaxed, because if you have a stressed par­ent or stressed child, you’re not going to get good data.

Kather­ine Ellis:
And I just want­ed to see high­light here one of the biggest chal­lenges that we have, which is achiev­ing cal­i­bra­tion. And this is because I think in Goril­la, WebGaz­er is in beta mode at the moment. It’s not as for­giv­ing as per­haps the cal­i­bra­tion pro­ce­dure as in the Tobii lap­top. And this has been one of the major bar­ri­ers that have become, not just in child research, but I think in adult research as well. And I just want­ed to show you some of the infor­ma­tion that we send to par­ents, which we were giv­en from Sarah White’s lab, who very kind­ly passed this onto us, and just high­light all the num­ber of things that you have to think about before even get­ting start­ed to make sure that the sit­u­a­tion is set up as best as pos­si­ble, to make sure every­thing runs real­ly, real­ly smoothly.

Kather­ine Ellis:
So you might be think­ing at this point, “Gosh, this is a lot of effort. Is it worth it with these pop­u­la­tions?” And if it works, actu­al­ly it def­i­nite­ly is and has some advan­tages. And I think hope­ful­ly our work will lead to some reflec­tions on a cost ben­e­fit analy­sis of who and when to do this research with. So one of the real­ly good things about online research with these par­tic­u­lar groups is the amount of bur­den that it reduces for fam­i­lies who often have quite a lot of addi­tion­al bur­dens for car­ing with a child with addi­tion­al needs. The main thing being lack of trav­el, extra flex­i­bil­i­ty and less pres­sure to get every­thing done, con­sid­er­ing that the groups that we work with are often spread around the coun­try because they’re quite rare. And yes, there are chal­lenges, but there are ways of get­ting around it.

Kather­ine Ellis:
So of course, any sort of good cost ben­e­fit analy­sis is going to be depen­dent on the data. And we are get­ting some pos­i­tive results. We’re find­ing that actu­al­ly kids are able to get past that cal­i­bra­tion, and that’s a huge achieve­ment, I think. So a lot of the typ­i­cal­ly devel­op­ing kids are cal­i­brat­ing, and we’re get­ting use­ful infor­ma­tion out of them. And then for the syn­drome groups, actu­al­ly, about two thirds are man­ag­ing to cal­i­brate, and I’m absolute­ly thrilled about this because although it might not be as high as the face-to-face cal­i­bra­tion, that’s eight out of 12 kids who haven’t had to come to the lab and have addi­tion­al bur­dens. I think this is real­ly, real­ly great news.

Kather­ine Ellis:
And what I want to do is I’ve start­ed to draw out the descrip­tive char­ac­ter­is­tics of the syn­drome chil­dren who are being suc­cess­ful and who aren’t. And there’s no clear pic­ture at the moment because there’s a lot of over­lap in these char­ac­ter­is­tics. So my mes­sage is, “Just give it a go at this moment.”

Kather­ine Ellis:
And then, of course, we want to find out, “Is the data mean­ing­ful? Is it valid?” We don’t have the sam­ple size yet to say for cer­tain whether the par­a­digms that we are using are work­ing online, but we know that they are work­ing in adult pop­u­la­tions. And in terms of a sim­ple social video view­ing par­a­digm, we’re see­ing pat­terns of behav­ior which is sim­i­lar to pre­vi­ous find­ings, par­tic­u­lar­ly in the Frag­ile X group. So they are watch­ing the social scenes, they’re just doing same times with short­er episodes, and this is per­haps what we would expect with this group, so that’s [inaudi­ble 00:08:22]. So thank you for lis­ten­ing. Do let me know if you have any ques­tions, drop me an email. I’ve got heaps of sto­ries that I can’t share in eight minutes.

Sophie Scott:
Thank you very much. Thank you. Well, the ques­tion I have is about [inaudi­ble 00:08:41], which is about the sam­ples cal­i­bra­tion data, which did look real­ly inter­est­ing, but how often did you find that you need­ed to recal­i­brate? And how long were your ses­sions in total?

Kather­ine Ellis:
So our videos are real­ly short, so they were only about four to five min­utes, and we haven’t had any reports. I’ve watched them. We haven’t had any reports of actu­al­ly when they’ve got past the cal­i­bra­tion that they’ve had to recal­i­brate. That’s real­ly, real­ly promis­ing. So that’s a real­ly good sign, because I think there was a huge ques­tion of whether you’d be able to do any­thing like this with this population.

Sophie Scott:
Yeah.

Kather­ine Ellis:
And then, sor­ry, was there a sec­ond part of the question?

Sophie Scott:
Yeah. How long are your ses­sions in total?

Kather­ine Ellis:
So as I said, the actu­al video is real­ly, real­ly short. The range of ses­sions real­ly depends on the child. And as I said, the good thing is we can be flex­i­ble. So for some kids, 20 min­utes is the absolute lim­it, and that’s the range of behav­ioral and cog­ni­tive tasks. For some, they can do it a lit­tle bit longer. And you just have to be very flex­i­ble in terms of how long or how lit­tle you do. And one of the things that I did­n’t men­tion is some­times you find that par­ents tend to be a bit more con­cerned about whether kids do it or not. So as a research, what’s real­ly impor­tant is mak­ing sure when to jump in and say, “Do you know what? I think we’ve done a lot. Why don’t we try it anoth­er day?”

Sophie Scott:
Yeah. Well, thank you very much. And don’t for­get, you can con­tin­ue ask­ing ques­tions in the Q&A, and those be answered by the speak­ers like Kather­ine after they fin­ish talk­ing. Thank you very much, Katherine.