WEBVTT
00:00:09.140 --> 00:00:09.140
[Podcast jingle][ringtone] Hello?
00:00:09.140 --> 00:00:09.769
Oh, it is recording.
00:00:09.771 --> 00:00:10.699
I see the little figure.
00:00:10.730 --> 00:00:11.179
Okay, great.
00:00:11.359 --> 00:00:14.179
I will do my little spiel and then I'll introduce you.
00:00:14.300 --> 00:00:14.810
Nice.
00:00:14.900 --> 00:00:15.259
Okay.
00:00:15.619 --> 00:00:16.160
Here I go.
00:00:16.489 --> 00:00:21.230
Hi, I'm Sue from the Salvesen Mindroom Research Centre at the University of Edinburgh.
00:00:21.890 --> 00:00:32.990
And we're recording this little series of sort of, not quite a podcast, um, calls with developmental psychologists during the coronavirus lockdown.
00:00:33.380 --> 00:00:47.270
We thought that, um, because lots of parents are homeschooling their children because lots of practitioners are having to think about delivering services in different ways, um, because lots of brilliant professionals have had to be furloughed.
00:00:48.229 --> 00:00:56.390
Um, there's a lot of, um, need for some kind of intellectual input and there's maybe some interest in developmental psychology, particularly at this time.
00:00:56.689 --> 00:01:05.870
And so it seemed a good moment to, um, do these little chats with psychologists who are studying learning and development in children and young people.
00:01:06.140 --> 00:01:15.709
So today's Psychologic-PsychologiCALL is with Jo Van Herwegen from the UCL Institute of education.
00:01:15.799 --> 00:01:21.829
And she's going to talk to me about maths development in Williams syndrome and Down syndrome.
00:01:22.370 --> 00:01:25.159
Okay hello, Jo! How are you today?
00:01:25.549 --> 00:01:26.599
I'm good! Thank you, Sue!
00:01:28.040 --> 00:01:31.730
Good! So tell me, what did you discover in this bit of research we're talking about?
00:01:32.390 --> 00:01:53.900
So previous research in individuals that William syndrome, which is a rare genetic disorder, and also those with Down syndrome had suggested that both of these groups may be delayed in their mathematical development and that this delay might be caused by difficulties in making estimations between large non-symbolic amounts.
00:01:54.469 --> 00:02:04.790
So in the current study, what we examined was, um, where people with Down syndrome and Williams syndrome were looking during a non-symbolic estimation task using an eye-tracker.
00:02:05.450 --> 00:02:17.300
So an eye-tracker allows you to see where people are looking on a screen, and we compared the, um, the performance against typically developing children who had similar reasoning abilities, but were younger of course.
00:02:17.689 --> 00:02:31.430
And what we found was that, although the people with Williams syndrome and Down syndrome performed at a similar level to the typically developing controls, who were younger children, they scanned the stimuli in their task, um, differently.
00:02:31.759 --> 00:02:37.039
So, um, we looked at the looking behavior, how that changes over time.
00:02:37.340 --> 00:02:51.680
And we found that, uh, in people with Williams syndrome, um, they changed from starting to move their eyes only a few times and not scanning the stimuli properly, um, to then over time becoming better at it.
00:02:51.830 --> 00:02:56.240
So the older people with Williams syndrome performed much better than the younger children.
00:02:57.110 --> 00:03:01.930
And that's now informing us, you know, about how we can help with interventions for people.
00:03:03.150 --> 00:03:08.039
Um, so, um, I'm just gonna pick up on the"non-symbolic numbers".
00:03:08.040 --> 00:03:13.319
Can you just explain to me what the difference is between non-symbolic and symbolic amounts?
00:03:13.320 --> 00:03:14.819
I don't think I've come across that before.
00:03:15.629 --> 00:03:19.199
OK, so um, symbolic amounts, think about, you know, the actual digits.
00:03:19.349 --> 00:03:22.830
So seeing number seven, you know, that, that stands for seven items.
00:03:23.430 --> 00:03:35.159
Um, and that's obviously very important for mathematical development, but you may also have a mathematical development ever, you know, when you have ever added up numbers and with a calculator and at the end, you go like"that can't be right!"
00:03:35.650 --> 00:03:35.819
Right...
00:03:35.650 --> 00:03:38.490
So how come your brain can be better than a calculator?
00:03:38.789 --> 00:03:43.379
And that's because our brain is very good at estimating mathematical amounts.
00:03:43.650 --> 00:03:55.199
So just, if you see, you know, how many chairs are in one room compared to how many chairs are in another room and you can quickly guess one room's got more chairs than another, so that's your non-symbolic estimation abilities.
00:03:55.650 --> 00:03:55.949
Okay.
00:03:55.950 --> 00:03:57.990
So actual numbers of things.
00:03:58.319 --> 00:04:04.530
And so, um, so tell me a bit more about what you were showing on the screen with the eye tracker then?
00:04:06.229 --> 00:04:11.879
So we were literally just showing them blue dots and red dots, and the dots differed in the size.
00:04:11.881 --> 00:04:18.029
So, you know, they're all mixed up, large and small blue dots compared to large and small red dots.
00:04:18.660 --> 00:04:20.790
Um, but sometimes we made it quite easy for children.
00:04:20.790 --> 00:04:24.129
So sometimes we showed them maybe five dots versus 10 dots.
00:04:24.240 --> 00:04:25.980
And obviously that difference is very large.
00:04:26.009 --> 00:04:37.170
That ratio is 0.5 and then it's really easy to see that there are more blue dots, but if I show you six versus eight dots, then obviously it might be more difficult to say which color had more dots.
00:04:37.170 --> 00:04:39.269
So that's what we did with the participants.
00:04:39.271 --> 00:04:43.350
They saw lots of dots and every time they had to say whether they were more blue dots or more red dots.
00:04:43.769 --> 00:04:44.160
Okay.
00:04:44.610 --> 00:04:52.199
And so, so you were interested in how accurately they answer that question, but also where they're looking while they're working it out, right?
00:04:53.250 --> 00:04:53.550
That's correct.
00:04:53.550 --> 00:04:55.230
So that's exactly what we were doing.
00:04:55.350 --> 00:05:03.360
And what we found was that, you know, they're, they're quite good at seeing, you know, I mean, they're as good as younger, typically developing children.
00:05:03.360 --> 00:05:05.069
So we know they're a little bit delayed.
00:05:05.970 --> 00:05:12.000
Um, but in their looking, we could see that those with Williams syndrome didn't move their eyes so much.
00:05:12.120 --> 00:05:18.540
So they didn't scan these patterns of dots, as well as, uh, some of the other, the people with Down syndrome.
00:05:19.189 --> 00:05:27.000
Um, and obviously if you don't scan things enough, then obviously you're not going to process that information in the same way.
00:05:27.000 --> 00:05:40.589
So this is where you can see with an eye tracker that although there was no difference in their outcome performance, in their overall score, you can still see with an eye-tracker, how people performed the task and whether that might be different between different groups.
00:05:42.029 --> 00:05:42.240
Um...
00:05:42.029 --> 00:05:57.860
And, and so, so thinking about what you do with eye-tracking data, for maybe people who, who aren't familiar with the technique, can you describe this a little bit, you know, what the information from the eye-tracker looks like, and what you do with it in order to be able analyze it?
00:05:58.610 --> 00:05:59.180
So there's...
00:05:59.209 --> 00:05:59.509
There's...
00:05:59.540 --> 00:05:59.839
Well.
00:05:59.899 --> 00:06:00.139
If...
00:06:00.620 --> 00:06:17.660
It gives you back a really nice video, which actually takes a red dot, um, that's symbolizes your eyes where they are on the screen, and you can then see-, play a video and it shows you where the red dots moves over the screen, which is quite nice, but obviously that's the easy way of showing it to people.
00:06:17.661 --> 00:06:24.589
But in order to then see whether that looking is actually different between people, you have to do quite a lot of work with eye-tracking data.
00:06:24.860 --> 00:06:32.959
So what you do is you have to make areas of interest, so this is what-, you need to tell the software which places on the screen you're interested in.
00:06:32.961 --> 00:06:36.800
And obviously we'd asked those were the red dots and also the blue dots.
00:06:37.490 --> 00:06:48.920
Um, and then you need to, um, you need to calculate where, for example, people were looking at what particular point in time, and how long those looks were for.
00:06:49.310 --> 00:07:02.420
So, um, you tend to get quite a lot of data points when you do this, because you break the videos all down to milliseconds and, you know, very short looks, and then decide on what is a look and how long that look is for.
00:07:02.839 --> 00:07:19.879
So, um, if you have a lot of different participants and a lot of trials, and those of you who work with Excel know that Excel can handle an awful lot of data, but I have had studies with eye-tracking where actually Excel could not handle the amount of data anymore.
00:07:20.000 --> 00:07:21.500
So there you go.
00:07:21.769 --> 00:07:23.300
But this time we didn't break Excel!
00:07:25.209 --> 00:07:26.139
[Laughs] That's good to know.
00:07:27.250 --> 00:07:43.029
Um, and so, and so what you're finding is that, um, uh, it's all about how the eyes are moving around the screen and not about how much time they spend looking, or was there a difference in how quickly they could answer the question as well?
00:07:43.810 --> 00:07:51.550
So, so we measured how quickly they could, um, um, answer the question, and whether they got it correct or incorrect.
00:07:51.550 --> 00:07:53.860
So that's what typical studies would have done, right?
00:07:53.860 --> 00:07:56.980
So look at how long does it take a participant to respond?
00:07:57.250 --> 00:08:05.259
Do they get the answer correct or incorrect, but with eye tracking, you go one step further and you say, where were they looking and how did they make the decision?
00:08:05.269 --> 00:08:13.360
So for example, one of the things we looked at is, um, how many times did they look at a red dot versus how many times did they look at the blue dots, and was there a difference?
00:08:13.509 --> 00:08:17.860
Cause that tells you something about were they actually doing any comparisons between them?
00:08:18.339 --> 00:08:24.550
And did it take them longer, not necessarily in reaction time, but kind of going backwards and forwards?
00:08:25.029 --> 00:08:28.899
What you can also measure is the long-, whether they show any longer look.
00:08:28.901 --> 00:08:40.840
So it could be for example, that children who are just guessing or children who are actually not making any comparisons, they just look at one area very long, but not move their eyes around so much.
00:08:40.841 --> 00:08:43.179
So those are kind of indicators.
00:08:43.181 --> 00:08:48.250
So the longest looks, um, how many times they looked at one area versus another.
00:08:48.490 --> 00:08:53.860
It gives us an indication of their, um, scanning patterns, or how they approached the task.
00:08:54.639 --> 00:09:18.240
So I think a really interesting thing, a really interesting question in developmental psychology whenever you're working with groups who are, um, developing in a kind of unique way, is whether that development is, is a delay, so sort of broadly following the same path, but just a bit slower or, um, well, we often call it a deviance.
00:09:18.240 --> 00:09:31.259
It's not really a particularly pleasant word, but anyway, whether children are following their own unique, independent path, that's maybe like a bit of a branch line from the well-trodden road of the typical child.
00:09:31.769 --> 00:09:36.059
So what would you say for Williams syndrome and Down syndrome in terms of their maths?
00:09:36.120 --> 00:09:41.730
Are they trotting along down the same road, but a bit slowly, or are they on their own road?
00:09:42.409 --> 00:09:53.659
So if you just look at performance data without the eye-tracking, you would say that they're, they're following their own path, they're delayed, but that they're very similar.
00:09:53.660 --> 00:09:58.759
So that those with Williams syndrome and those with Down syndrome, that they follow a very similar pathway.
00:09:59.269 --> 00:10:10.519
When you then start looking at what is predicting that pathway, or how does that relate to their scanning abilities in this case, then you start to see the differences between the two groups.
00:10:10.850 --> 00:10:15.919
So what we've found was that we also did a study a long time ago with infants who were very young.
00:10:16.220 --> 00:10:36.240
And then the difference between the two groups was even larger, where we found that those with Down syndrome really found it difficult to concentrate and to actually have sustained looking at particular objects, compared to those with Williams syndrome who had very long looks and found it difficult to disengage their...
00:10:35.889 --> 00:10:41.149
their eye movements from one particular place and then move their eyes to another place.
00:10:41.570 --> 00:10:50.600
And what we now seen the data in the current study is that in William syndrome, that looking pattern is still predictive for their mathematical ability.
00:10:50.600 --> 00:11:03.200
So although they don't show this sustained, this um, these long looks anymore, these difficulties with disengaging their looks, but they're still in the, you can still see the correlation with their mathematical ability.
00:11:03.210 --> 00:11:17.929
So those who have more typical looks and are better at moving our eyes around, they're also much better at mathematical development compared to those who still have some longer looking patterns and find it difficult to disengage their eyes.
00:11:18.169 --> 00:11:21.350
They're the ones who should have a weaker mathematical abilities.
00:11:22.879 --> 00:11:23.059
Right...
00:11:22.879 --> 00:11:27.259
So what, what could we do with this information in terms of supporting maths development?
00:11:27.289 --> 00:11:39.049
Obviously it's a really important skill for future kind of autonomy and, you know, um, self-determination to be able to, um, use basic maths in everyday life, right?
00:11:39.259 --> 00:11:44.179
So do you think there are ways that we can present information that would be more helpful?
00:11:44.580 --> 00:11:47.960
What do you think it means for, for teachers, for example?
00:11:48.409 --> 00:11:49.460
So what, what comes out...
00:11:49.490 --> 00:11:56.470
Because this study was obviously part of a much larger study where we looked also at other factors that may predict mathematical ability.
00:11:56.471 --> 00:12:02.019
And this, with the eye- tracker, we just looked at attention and, and eye movements, whether they told us something.
00:12:01.900 --> 00:12:17.200
Um, so from the eye movements alone, actually what it shows us is that, um, yes, they might be looking slightly differently at a task, but these attention abilities don't seem to predict much for Down syndrome about how to train mathematical abilities.
00:12:16.660 --> 00:12:20.110
It's slightly different if you talk about for a young children.
00:12:20.980 --> 00:12:22.960
Um, so for the older children, we would say...
00:12:23.049 --> 00:12:34.929
For those with Down-, for those with Williams syndrome, they have quite good counting abilities, in that they can really r ecite the numbers, but they may not know what these numbers stand for because their estimation abilities aren't so great.
00:12:34.931 --> 00:12:41.139
So when they then do an addition and you say to them"how much is 2 plus 3?", and they will just say"9".
00:12:41.679 --> 00:12:49.539
And where our estimation abilities would say"ooh, that's a bit much" when you add 2 and 3, they don't have that kind of checking system.
00:12:49.809 --> 00:12:58.509
So giving them a better understanding, what we call"number sense", of how numbers relate to each other, would be beneficial for people with Williams syndrome.
00:12:58.870 --> 00:13:08.830
So our studies seem to suggest that in, in, in terms of how you would help people with Williams syndrome and Down syndrome, you probably have slightly different, um, intervention programs.
00:13:09.029 --> 00:13:09.389
Right.
00:13:09.779 --> 00:13:11.190
That's really fantastic.
00:13:11.549 --> 00:13:22.409
Um, it's lovely, isn't it, when you do a piece of kind of really, um, solid experimental research like this, but you can also, you know, draw something useful for the community out of it.
00:13:22.919 --> 00:13:32.909
Yeah, and I think this is where if you really think carefully about your hypothesis and how theory informs your research, you can still link it to practice.
00:13:33.690 --> 00:13:37.409
Um, you know, by, by thinking in the long term, what do we do?
00:13:37.410 --> 00:13:47.610
So this study was always planned with the idea of, if we do this study, we can then hopefully have clear predictions of which kind of practice or interventions should work.
00:13:48.309 --> 00:13:48.690
Mm Hmm.
00:13:49.559 --> 00:13:52.500
And so what do you want to do next in this area?
00:13:52.500 --> 00:13:53.970
Or what are you doing at the moment?
00:13:54.029 --> 00:13:56.850
Are you still working on maths development in these younger children?
00:13:57.960 --> 00:14:06.929
Yes, yes! So what we've done now is we've put in, um, um, you know, an application for funding cause that's where it all falls down to, right!...
00:14:07.679 --> 00:14:07.820
Yes!
00:14:07.679 --> 00:14:07.820
...
00:14:07.821 --> 00:14:13.919
to actually start doing, um, two interventions, um, that parents can do in the home.
00:14:14.940 --> 00:14:21.600
Um, and we're also helping parents, um, at home already during COVID-19, um, to help their children learn better.
00:14:21.600 --> 00:14:23.970
So we have the Maths at Home games at the moment.
00:14:24.269 --> 00:14:28.500
So every day we advertise a new game that parents can do with their children.
00:14:29.129 --> 00:14:32.850
Um, and you don't need any kind of specific tools or worksheets for it.
00:14:32.850 --> 00:14:35.519
It's all with things you can find around the house.
00:14:36.090 --> 00:14:58.309
Oh, that's amazing! I'll make sure we link to that in the, in the podcast description, cause that's, that's fantastic! Well done! Um, so the last question then is I'm thinking about, um, any early career researchers, um, students, PhD students, undergrads, who are kind of stuck at home and maybe feeling a bit anxious about...
00:14:57.889 --> 00:15:12.049
[laughs] how their career might develop from here through this, the mirky waters that we're in! So I just wondered if you had anything you wanted to share or say to them, any words of wisdom, um, for those people?
00:15:12.220 --> 00:15:15.100
Oh, I always find it so difficult! Cause I think, um...
00:15:15.100 --> 00:15:16.870
I think one...
00:15:17.289 --> 00:15:18.610
I have two, if I may.
00:15:18.750 --> 00:15:19.509
I'm gonna keep it short.
00:15:19.510 --> 00:15:31.450
One is, you know, to really start thinking, um, as I said earlier on like longitudinally so, so, you know, you need to break down if you want to get an intervention for people, there's, there's a lot of...
00:15:31.570 --> 00:15:40.269
Some basic research that needs to happen first before you kind of have a good feeling"this is what's going to help them best", right, from in terms of predictions.
00:15:40.899 --> 00:15:55.779
But this is also an important one when it comes about eye-tracking, because although it's very useful tool, in terms of knowing how people solve a task, it isn't as easy as just having your standard experiment and then just adding eye tracking to it.
00:15:56.230 --> 00:16:03.009
Basically what happens is when you do that, you probably are going to get in a lot of difficulties with analyzing your data.
00:16:03.340 --> 00:16:05.830
And I'm speaking here[laughs] out of experience in that um...
00:16:05.740 --> 00:16:26.019
[Laughs] You know, the Tobi, the Tobi software we've used i s very easy for simple experiments, but because we had to also integrate it with iPrime studio to show our experiments, we, u m, because of t he software updates the two couldn't talk to each other anymore.
00:16:26.020 --> 00:16:30.190
And so we had to, we ended up with having to do a lot of the data coding by hand.
00:16:30.610 --> 00:16:36.429
So I would say, do you know what if I knew Matlab much better, it w ould h ave been much easier.
00:16:36.610 --> 00:16:48.820
So if you're stuck at home as an early career researcher, I think this is the perfect time to start learning some new programming languages, whether it's, you know, or s oftwares whether it's Python or Matlab or...
00:16:49.059 --> 00:16:54.519
Because that's where the future in research really lies, it's knowing all these different systems much better.
00:16:55.059 --> 00:16:57.840
So that's what I would advise t hem t o do.
00:16:58.809 --> 00:17:00.519
Yeah, I, I really agree.
00:17:00.580 --> 00:17:11.019
I actually feel it's, it's one of the biggest things that I wish I could do for myself is learn to code properly and I worry that I've just kind of missed the boat and I'll never find the time.
00:17:12.009 --> 00:17:12.609
Um, but...
00:17:12.789 --> 00:17:12.849
Yeah...
00:17:13.299 --> 00:17:26.380
I've tried to start coding in Python five times, but every time, you know, there's a bigger project or I don't, you know, I get stuck at a particular point and by the time I've had the ability to pick it up again, you know, I need to start from scratch again.
00:17:26.381 --> 00:17:36.819
So I think if you're now stuck at home for a few solid weeks, this is a skill on your CV that is really worth a lot in terms of money and opportunities, I think.
00:17:37.539 --> 00:17:44.740
And that other thing of course, that your advice highlights is how much, you know, research just doesn't happen in an ideal world.
00:17:44.740 --> 00:17:52.769
And I think sometimes we're trying to present out to the public as if everything that happened was exactly what we had planned to happen.
00:17:52.050 --> 00:18:03.269
[Laughs] Um, but actually, you know, like everyone we're subject to these challenges, technical challenges, you know, your point about the software getting updated is just so resonant.
00:18:03.299 --> 00:18:19.980
You know, especially if you're doing longitudinal studies or studies that are open for a long time because you're recruiting, uh, a hard to reach sample these, you know, you're really vulnerable to these things and it can be pretty stressful, so well done for overcoming that Jo I'm impressed.
00:18:21.059 --> 00:18:22.200
Thanks very much Sue.
00:18:23.069 --> 00:18:27.329
So I think we will draw to a close, thank you so much for giving me your time.
00:18:28.170 --> 00:18:36.299
Um, for anyone listening, you'll be able to find out more about Jo's work by following links on the podcast page at ed.ac.uk/salvesen-research.
00:18:40.200 --> 00:18:46.079
We'll link to Joe's paper and her profile and the resources that she mentioned for maths learning as well.
00:18:46.980 --> 00:18:48.299
Thank you so much, Jo!
00:18:49.109 --> 00:18:50.490
Thank you! Have a good day!
00:18:50.849 --> 00:18:51.329
You too! Bye!
00:18:51.140 --> 00:18:51.329
Bye![ringtone]
00:18:57.319 --> 00:19:00.259
Okay we did it! I thought that went quite smoothly![Podcast jingle]
00:00:09.140 --> 00:00:09.140
[Podcast jingle][ringtone] Hello?
00:00:09.140 --> 00:00:09.769
Oh, it is recording.
00:00:09.771 --> 00:00:10.699
I see the little figure.
00:00:10.730 --> 00:00:11.179
Okay, great.
00:00:11.359 --> 00:00:14.179
I will do my little spiel and then I'll introduce you.
00:00:14.300 --> 00:00:14.810
Nice.
00:00:14.900 --> 00:00:15.259
Okay.
00:00:15.619 --> 00:00:16.160
Here I go.
00:00:16.489 --> 00:00:21.230
Hi, I'm Sue from the Salvesen Mindroom Research Centre at the University of Edinburgh.
00:00:21.890 --> 00:00:32.990
And we're recording this little series of sort of, not quite a podcast, um, calls with developmental psychologists during the coronavirus lockdown.
00:00:33.380 --> 00:00:47.270
We thought that, um, because lots of parents are homeschooling their children because lots of practitioners are having to think about delivering services in different ways, um, because lots of brilliant professionals have had to be furloughed.
00:00:48.229 --> 00:00:56.390
Um, there's a lot of, um, need for some kind of intellectual input and there's maybe some interest in developmental psychology, particularly at this time.
00:00:56.689 --> 00:01:05.870
And so it seemed a good moment to, um, do these little chats with psychologists who are studying learning and development in children and young people.
00:01:06.140 --> 00:01:15.709
So today's Psychologic-PsychologiCALL is with Jo Van Herwegen from the UCL Institute of education.
00:01:15.799 --> 00:01:21.829
And she's going to talk to me about maths development in Williams syndrome and Down syndrome.
00:01:22.370 --> 00:01:25.159
Okay hello, Jo! How are you today?
00:01:25.549 --> 00:01:26.599
I'm good! Thank you, Sue!
00:01:28.040 --> 00:01:31.730
Good! So tell me, what did you discover in this bit of research we're talking about?
00:01:32.390 --> 00:01:53.900
So previous research in individuals that William syndrome, which is a rare genetic disorder, and also those with Down syndrome had suggested that both of these groups may be delayed in their mathematical development and that this delay might be caused by difficulties in making estimations between large non-symbolic amounts.
00:01:54.469 --> 00:02:04.790
So in the current study, what we examined was, um, where people with Down syndrome and Williams syndrome were looking during a non-symbolic estimation task using an eye-tracker.
00:02:05.450 --> 00:02:17.300
So an eye-tracker allows you to see where people are looking on a screen, and we compared the, um, the performance against typically developing children who had similar reasoning abilities, but were younger of course.
00:02:17.689 --> 00:02:31.430
And what we found was that, although the people with Williams syndrome and Down syndrome performed at a similar level to the typically developing controls, who were younger children, they scanned the stimuli in their task, um, differently.
00:02:31.759 --> 00:02:37.039
So, um, we looked at the looking behavior, how that changes over time.
00:02:37.340 --> 00:02:51.680
And we found that, uh, in people with Williams syndrome, um, they changed from starting to move their eyes only a few times and not scanning the stimuli properly, um, to then over time becoming better at it.
00:02:51.830 --> 00:02:56.240
So the older people with Williams syndrome performed much better than the younger children.
00:02:57.110 --> 00:03:01.930
And that's now informing us, you know, about how we can help with interventions for people.
00:03:03.150 --> 00:03:08.039
Um, so, um, I'm just gonna pick up on the"non-symbolic numbers".
00:03:08.040 --> 00:03:13.319
Can you just explain to me what the difference is between non-symbolic and symbolic amounts?
00:03:13.320 --> 00:03:14.819
I don't think I've come across that before.
00:03:15.629 --> 00:03:19.199
OK, so um, symbolic amounts, think about, you know, the actual digits.
00:03:19.349 --> 00:03:22.830
So seeing number seven, you know, that, that stands for seven items.
00:03:23.430 --> 00:03:35.159
Um, and that's obviously very important for mathematical development, but you may also have a mathematical development ever, you know, when you have ever added up numbers and with a calculator and at the end, you go like"that can't be right!"
00:03:35.650 --> 00:03:35.819
Right...
00:03:35.650 --> 00:03:38.490
So how come your brain can be better than a calculator?
00:03:38.789 --> 00:03:43.379
And that's because our brain is very good at estimating mathematical amounts.
00:03:43.650 --> 00:03:55.199
So just, if you see, you know, how many chairs are in one room compared to how many chairs are in another room and you can quickly guess one room's got more chairs than another, so that's your non-symbolic estimation abilities.
00:03:55.650 --> 00:03:55.949
Okay.
00:03:55.950 --> 00:03:57.990
So actual numbers of things.
00:03:58.319 --> 00:04:04.530
And so, um, so tell me a bit more about what you were showing on the screen with the eye tracker then?
00:04:06.229 --> 00:04:11.879
So we were literally just showing them blue dots and red dots, and the dots differed in the size.
00:04:11.881 --> 00:04:18.029
So, you know, they're all mixed up, large and small blue dots compared to large and small red dots.
00:04:18.660 --> 00:04:20.790
Um, but sometimes we made it quite easy for children.
00:04:20.790 --> 00:04:24.129
So sometimes we showed them maybe five dots versus 10 dots.
00:04:24.240 --> 00:04:25.980
And obviously that difference is very large.
00:04:26.009 --> 00:04:37.170
That ratio is 0.5 and then it's really easy to see that there are more blue dots, but if I show you six versus eight dots, then obviously it might be more difficult to say which color had more dots.
00:04:37.170 --> 00:04:39.269
So that's what we did with the participants.
00:04:39.271 --> 00:04:43.350
They saw lots of dots and every time they had to say whether they were more blue dots or more red dots.
00:04:43.769 --> 00:04:44.160
Okay.
00:04:44.610 --> 00:04:52.199
And so, so you were interested in how accurately they answer that question, but also where they're looking while they're working it out, right?
00:04:53.250 --> 00:04:53.550
That's correct.
00:04:53.550 --> 00:04:55.230
So that's exactly what we were doing.
00:04:55.350 --> 00:05:03.360
And what we found was that, you know, they're, they're quite good at seeing, you know, I mean, they're as good as younger, typically developing children.
00:05:03.360 --> 00:05:05.069
So we know they're a little bit delayed.
00:05:05.970 --> 00:05:12.000
Um, but in their looking, we could see that those with Williams syndrome didn't move their eyes so much.
00:05:12.120 --> 00:05:18.540
So they didn't scan these patterns of dots, as well as, uh, some of the other, the people with Down syndrome.
00:05:19.189 --> 00:05:27.000
Um, and obviously if you don't scan things enough, then obviously you're not going to process that information in the same way.
00:05:27.000 --> 00:05:40.589
So this is where you can see with an eye tracker that although there was no difference in their outcome performance, in their overall score, you can still see with an eye-tracker, how people performed the task and whether that might be different between different groups.
00:05:42.029 --> 00:05:42.240
Um...
00:05:42.029 --> 00:05:57.860
And, and so, so thinking about what you do with eye-tracking data, for maybe people who, who aren't familiar with the technique, can you describe this a little bit, you know, what the information from the eye-tracker looks like, and what you do with it in order to be able analyze it?
00:05:58.610 --> 00:05:59.180
So there's...
00:05:59.209 --> 00:05:59.509
There's...
00:05:59.540 --> 00:05:59.839
Well.
00:05:59.899 --> 00:06:00.139
If...
00:06:00.620 --> 00:06:17.660
It gives you back a really nice video, which actually takes a red dot, um, that's symbolizes your eyes where they are on the screen, and you can then see-, play a video and it shows you where the red dots moves over the screen, which is quite nice, but obviously that's the easy way of showing it to people.
00:06:17.661 --> 00:06:24.589
But in order to then see whether that looking is actually different between people, you have to do quite a lot of work with eye-tracking data.
00:06:24.860 --> 00:06:32.959
So what you do is you have to make areas of interest, so this is what-, you need to tell the software which places on the screen you're interested in.
00:06:32.961 --> 00:06:36.800
And obviously we'd asked those were the red dots and also the blue dots.
00:06:37.490 --> 00:06:48.920
Um, and then you need to, um, you need to calculate where, for example, people were looking at what particular point in time, and how long those looks were for.
00:06:49.310 --> 00:07:02.420
So, um, you tend to get quite a lot of data points when you do this, because you break the videos all down to milliseconds and, you know, very short looks, and then decide on what is a look and how long that look is for.
00:07:02.839 --> 00:07:19.879
So, um, if you have a lot of different participants and a lot of trials, and those of you who work with Excel know that Excel can handle an awful lot of data, but I have had studies with eye-tracking where actually Excel could not handle the amount of data anymore.
00:07:20.000 --> 00:07:21.500
So there you go.
00:07:21.769 --> 00:07:23.300
But this time we didn't break Excel!
00:07:25.209 --> 00:07:26.139
[Laughs] That's good to know.
00:07:27.250 --> 00:07:43.029
Um, and so, and so what you're finding is that, um, uh, it's all about how the eyes are moving around the screen and not about how much time they spend looking, or was there a difference in how quickly they could answer the question as well?
00:07:43.810 --> 00:07:51.550
So, so we measured how quickly they could, um, um, answer the question, and whether they got it correct or incorrect.
00:07:51.550 --> 00:07:53.860
So that's what typical studies would have done, right?
00:07:53.860 --> 00:07:56.980
So look at how long does it take a participant to respond?
00:07:57.250 --> 00:08:05.259
Do they get the answer correct or incorrect, but with eye tracking, you go one step further and you say, where were they looking and how did they make the decision?
00:08:05.269 --> 00:08:13.360
So for example, one of the things we looked at is, um, how many times did they look at a red dot versus how many times did they look at the blue dots, and was there a difference?
00:08:13.509 --> 00:08:17.860
Cause that tells you something about were they actually doing any comparisons between them?
00:08:18.339 --> 00:08:24.550
And did it take them longer, not necessarily in reaction time, but kind of going backwards and forwards?
00:08:25.029 --> 00:08:28.899
What you can also measure is the long-, whether they show any longer look.
00:08:28.901 --> 00:08:40.840
So it could be for example, that children who are just guessing or children who are actually not making any comparisons, they just look at one area very long, but not move their eyes around so much.
00:08:40.841 --> 00:08:43.179
So those are kind of indicators.
00:08:43.181 --> 00:08:48.250
So the longest looks, um, how many times they looked at one area versus another.
00:08:48.490 --> 00:08:53.860
It gives us an indication of their, um, scanning patterns, or how they approached the task.
00:08:54.639 --> 00:09:18.240
So I think a really interesting thing, a really interesting question in developmental psychology whenever you're working with groups who are, um, developing in a kind of unique way, is whether that development is, is a delay, so sort of broadly following the same path, but just a bit slower or, um, well, we often call it a deviance.
00:09:18.240 --> 00:09:31.259
It's not really a particularly pleasant word, but anyway, whether children are following their own unique, independent path, that's maybe like a bit of a branch line from the well-trodden road of the typical child.
00:09:31.769 --> 00:09:36.059
So what would you say for Williams syndrome and Down syndrome in terms of their maths?
00:09:36.120 --> 00:09:41.730
Are they trotting along down the same road, but a bit slowly, or are they on their own road?
00:09:42.409 --> 00:09:53.659
So if you just look at performance data without the eye-tracking, you would say that they're, they're following their own path, they're delayed, but that they're very similar.
00:09:53.660 --> 00:09:58.759
So that those with Williams syndrome and those with Down syndrome, that they follow a very similar pathway.
00:09:59.269 --> 00:10:10.519
When you then start looking at what is predicting that pathway, or how does that relate to their scanning abilities in this case, then you start to see the differences between the two groups.
00:10:10.850 --> 00:10:15.919
So what we've found was that we also did a study a long time ago with infants who were very young.
00:10:16.220 --> 00:10:36.240
And then the difference between the two groups was even larger, where we found that those with Down syndrome really found it difficult to concentrate and to actually have sustained looking at particular objects, compared to those with Williams syndrome who had very long looks and found it difficult to disengage their...
00:10:35.889 --> 00:10:41.149
their eye movements from one particular place and then move their eyes to another place.
00:10:41.570 --> 00:10:50.600
And what we now seen the data in the current study is that in William syndrome, that looking pattern is still predictive for their mathematical ability.
00:10:50.600 --> 00:11:03.200
So although they don't show this sustained, this um, these long looks anymore, these difficulties with disengaging their looks, but they're still in the, you can still see the correlation with their mathematical ability.
00:11:03.210 --> 00:11:17.929
So those who have more typical looks and are better at moving our eyes around, they're also much better at mathematical development compared to those who still have some longer looking patterns and find it difficult to disengage their eyes.
00:11:18.169 --> 00:11:21.350
They're the ones who should have a weaker mathematical abilities.
00:11:22.879 --> 00:11:23.059
Right...
00:11:22.879 --> 00:11:27.259
So what, what could we do with this information in terms of supporting maths development?
00:11:27.289 --> 00:11:39.049
Obviously it's a really important skill for future kind of autonomy and, you know, um, self-determination to be able to, um, use basic maths in everyday life, right?
00:11:39.259 --> 00:11:44.179
So do you think there are ways that we can present information that would be more helpful?
00:11:44.580 --> 00:11:47.960
What do you think it means for, for teachers, for example?
00:11:48.409 --> 00:11:49.460
So what, what comes out...
00:11:49.490 --> 00:11:56.470
Because this study was obviously part of a much larger study where we looked also at other factors that may predict mathematical ability.
00:11:56.471 --> 00:12:02.019
And this, with the eye- tracker, we just looked at attention and, and eye movements, whether they told us something.
00:12:01.900 --> 00:12:17.200
Um, so from the eye movements alone, actually what it shows us is that, um, yes, they might be looking slightly differently at a task, but these attention abilities don't seem to predict much for Down syndrome about how to train mathematical abilities.
00:12:16.660 --> 00:12:20.110
It's slightly different if you talk about for a young children.
00:12:20.980 --> 00:12:22.960
Um, so for the older children, we would say...
00:12:23.049 --> 00:12:34.929
For those with Down-, for those with Williams syndrome, they have quite good counting abilities, in that they can really r ecite the numbers, but they may not know what these numbers stand for because their estimation abilities aren't so great.
00:12:34.931 --> 00:12:41.139
So when they then do an addition and you say to them"how much is 2 plus 3?", and they will just say"9".
00:12:41.679 --> 00:12:49.539
And where our estimation abilities would say"ooh, that's a bit much" when you add 2 and 3, they don't have that kind of checking system.
00:12:49.809 --> 00:12:58.509
So giving them a better understanding, what we call"number sense", of how numbers relate to each other, would be beneficial for people with Williams syndrome.
00:12:58.870 --> 00:13:08.830
So our studies seem to suggest that in, in, in terms of how you would help people with Williams syndrome and Down syndrome, you probably have slightly different, um, intervention programs.
00:13:09.029 --> 00:13:09.389
Right.
00:13:09.779 --> 00:13:11.190
That's really fantastic.
00:13:11.549 --> 00:13:22.409
Um, it's lovely, isn't it, when you do a piece of kind of really, um, solid experimental research like this, but you can also, you know, draw something useful for the community out of it.
00:13:22.919 --> 00:13:32.909
Yeah, and I think this is where if you really think carefully about your hypothesis and how theory informs your research, you can still link it to practice.
00:13:33.690 --> 00:13:37.409
Um, you know, by, by thinking in the long term, what do we do?
00:13:37.410 --> 00:13:47.610
So this study was always planned with the idea of, if we do this study, we can then hopefully have clear predictions of which kind of practice or interventions should work.
00:13:48.309 --> 00:13:48.690
Mm Hmm.
00:13:49.559 --> 00:13:52.500
And so what do you want to do next in this area?
00:13:52.500 --> 00:13:53.970
Or what are you doing at the moment?
00:13:54.029 --> 00:13:56.850
Are you still working on maths development in these younger children?
00:13:57.960 --> 00:14:06.929
Yes, yes! So what we've done now is we've put in, um, um, you know, an application for funding cause that's where it all falls down to, right!...
00:14:07.679 --> 00:14:07.820
Yes!
00:14:07.679 --> 00:14:07.820
...
00:14:07.821 --> 00:14:13.919
to actually start doing, um, two interventions, um, that parents can do in the home.
00:14:14.940 --> 00:14:21.600
Um, and we're also helping parents, um, at home already during COVID-19, um, to help their children learn better.
00:14:21.600 --> 00:14:23.970
So we have the Maths at Home games at the moment.
00:14:24.269 --> 00:14:28.500
So every day we advertise a new game that parents can do with their children.
00:14:29.129 --> 00:14:32.850
Um, and you don't need any kind of specific tools or worksheets for it.
00:14:32.850 --> 00:14:35.519
It's all with things you can find around the house.
00:14:36.090 --> 00:14:58.309
Oh, that's amazing! I'll make sure we link to that in the, in the podcast description, cause that's, that's fantastic! Well done! Um, so the last question then is I'm thinking about, um, any early career researchers, um, students, PhD students, undergrads, who are kind of stuck at home and maybe feeling a bit anxious about...
00:14:57.889 --> 00:15:12.049
[laughs] how their career might develop from here through this, the mirky waters that we're in! So I just wondered if you had anything you wanted to share or say to them, any words of wisdom, um, for those people?
00:15:12.220 --> 00:15:15.100
Oh, I always find it so difficult! Cause I think, um...
00:15:15.100 --> 00:15:16.870
I think one...
00:15:17.289 --> 00:15:18.610
I have two, if I may.
00:15:18.750 --> 00:15:19.509
I'm gonna keep it short.
00:15:19.510 --> 00:15:31.450
One is, you know, to really start thinking, um, as I said earlier on like longitudinally so, so, you know, you need to break down if you want to get an intervention for people, there's, there's a lot of...
00:15:31.570 --> 00:15:40.269
Some basic research that needs to happen first before you kind of have a good feeling"this is what's going to help them best", right, from in terms of predictions.
00:15:40.899 --> 00:15:55.779
But this is also an important one when it comes about eye-tracking, because although it's very useful tool, in terms of knowing how people solve a task, it isn't as easy as just having your standard experiment and then just adding eye tracking to it.
00:15:56.230 --> 00:16:03.009
Basically what happens is when you do that, you probably are going to get in a lot of difficulties with analyzing your data.
00:16:03.340 --> 00:16:05.830
And I'm speaking here[laughs] out of experience in that um...
00:16:05.740 --> 00:16:26.019
[Laughs] You know, the Tobi, the Tobi software we've used i s very easy for simple experiments, but because we had to also integrate it with iPrime studio to show our experiments, we, u m, because of t he software updates the two couldn't talk to each other anymore.
00:16:26.020 --> 00:16:30.190
And so we had to, we ended up with having to do a lot of the data coding by hand.
00:16:30.610 --> 00:16:36.429
So I would say, do you know what if I knew Matlab much better, it w ould h ave been much easier.
00:16:36.610 --> 00:16:48.820
So if you're stuck at home as an early career researcher, I think this is the perfect time to start learning some new programming languages, whether it's, you know, or s oftwares whether it's Python or Matlab or...
00:16:49.059 --> 00:16:54.519
Because that's where the future in research really lies, it's knowing all these different systems much better.
00:16:55.059 --> 00:16:57.840
So that's what I would advise t hem t o do.
00:16:58.809 --> 00:17:00.519
Yeah, I, I really agree.
00:17:00.580 --> 00:17:11.019
I actually feel it's, it's one of the biggest things that I wish I could do for myself is learn to code properly and I worry that I've just kind of missed the boat and I'll never find the time.
00:17:12.009 --> 00:17:12.609
Um, but...
00:17:12.789 --> 00:17:12.849
Yeah...
00:17:13.299 --> 00:17:26.380
I've tried to start coding in Python five times, but every time, you know, there's a bigger project or I don't, you know, I get stuck at a particular point and by the time I've had the ability to pick it up again, you know, I need to start from scratch again.
00:17:26.381 --> 00:17:36.819
So I think if you're now stuck at home for a few solid weeks, this is a skill on your CV that is really worth a lot in terms of money and opportunities, I think.
00:17:37.539 --> 00:17:44.740
And that other thing of course, that your advice highlights is how much, you know, research just doesn't happen in an ideal world.
00:17:44.740 --> 00:17:52.769
And I think sometimes we're trying to present out to the public as if everything that happened was exactly what we had planned to happen.
00:17:52.050 --> 00:18:03.269
[Laughs] Um, but actually, you know, like everyone we're subject to these challenges, technical challenges, you know, your point about the software getting updated is just so resonant.
00:18:03.299 --> 00:18:19.980
You know, especially if you're doing longitudinal studies or studies that are open for a long time because you're recruiting, uh, a hard to reach sample these, you know, you're really vulnerable to these things and it can be pretty stressful, so well done for overcoming that Jo I'm impressed.
00:18:21.059 --> 00:18:22.200
Thanks very much Sue.
00:18:23.069 --> 00:18:27.329
So I think we will draw to a close, thank you so much for giving me your time.
00:18:28.170 --> 00:18:36.299
Um, for anyone listening, you'll be able to find out more about Jo's work by following links on the podcast page at ed.ac.uk/salvesen-research.
00:18:40.200 --> 00:18:46.079
We'll link to Joe's paper and her profile and the resources that she mentioned for maths learning as well.
00:18:46.980 --> 00:18:48.299
Thank you so much, Jo!
00:18:49.109 --> 00:18:50.490
Thank you! Have a good day!
00:18:50.849 --> 00:18:51.329
You too! Bye!
00:18:51.140 --> 00:18:51.329
Bye![ringtone]
00:18:57.319 --> 00:19:00.259
Okay we did it! I thought that went quite smoothly![Podcast jingle]
