WEBVTT

00:00:00.000 --> 00:00:00.870 align:middle line:90%
Hello.

00:00:00.870 --> 00:00:01.710 align:middle line:90%
Hello.

00:00:01.710 --> 00:00:03.090 align:middle line:90%
How are you?

00:00:03.090 --> 00:00:04.590 align:middle line:90%
Fine.

00:00:04.590 --> 00:00:08.100 align:middle line:84%
So when you're doing
pupillometry studies,

00:00:08.100 --> 00:00:11.640 align:middle line:84%
do you also often use
gaze as a measure?

00:00:11.640 --> 00:00:15.450 align:middle line:84%
Yes, I do first of all, when
you're doing pupillometry,

00:00:15.450 --> 00:00:17.470 align:middle line:90%
you're also doing eye tracking.

00:00:17.470 --> 00:00:20.790 align:middle line:84%
So you have that
data automatically.

00:00:20.790 --> 00:00:24.660 align:middle line:84%
And it's very useful
to look at it.

00:00:24.660 --> 00:00:27.910 align:middle line:84%
And what kinds of
reasons is it useful?

00:00:27.910 --> 00:00:31.090 align:middle line:90%
Well, a couple of reasons.

00:00:31.090 --> 00:00:38.250 align:middle line:84%
One reason is that in most
pupillometry studies, people,

00:00:38.250 --> 00:00:42.180 align:middle line:84%
participants are required
to maintain fixation.

00:00:42.180 --> 00:00:43.920 align:middle line:84%
Then if you have
eye-tracking data,

00:00:43.920 --> 00:00:47.610 align:middle line:90%
you can actually validate that.

00:00:47.610 --> 00:00:52.320 align:middle line:84%
You can make sure that people
were maintaining fixation.

00:00:52.320 --> 00:00:56.470 align:middle line:84%
Also, there are
many studies that

00:00:56.470 --> 00:01:00.700 align:middle line:84%
want people to look at something
in a very spontaneous way,

00:01:00.700 --> 00:01:03.580 align:middle line:84%
for example, you're
showing a photograph.

00:01:03.580 --> 00:01:12.310 align:middle line:84%
So in that case, you
cannot control completely

00:01:12.310 --> 00:01:15.730 align:middle line:84%
the luminance condition because
every photograph will have

00:01:15.730 --> 00:01:18.370 align:middle line:84%
parts that are
darker or brighter.

00:01:18.370 --> 00:01:24.340 align:middle line:84%
Then by checking where people
are looking on the photograph,

00:01:24.340 --> 00:01:30.280 align:middle line:84%
you may know, for example,
what could be causing the pupil

00:01:30.280 --> 00:01:34.930 align:middle line:84%
to change, whether it is
the content of the image

00:01:34.930 --> 00:01:39.100 align:middle line:84%
or the differences in
luminance in the image?

00:01:39.100 --> 00:01:40.240 align:middle line:90%
Mm-hmm.

00:01:40.240 --> 00:01:43.360 align:middle line:84%
And do you have an
example of this?

00:01:43.360 --> 00:01:44.920 align:middle line:90%
Yes.

00:01:44.920 --> 00:01:48.520 align:middle line:84%
As a matter of fact, it
was very important in order

00:01:48.520 --> 00:01:52.540 align:middle line:90%
to understand this illusion.

00:01:52.540 --> 00:01:57.580 align:middle line:84%
This is called the face
race lightness illusion.

00:01:57.580 --> 00:02:00.850 align:middle line:84%
So you see there are
two faces side by side.

00:02:00.850 --> 00:02:01.600 align:middle line:90%
Yeah.

00:02:01.600 --> 00:02:03.910 align:middle line:90%
One looks European.

00:02:03.910 --> 00:02:05.990 align:middle line:90%
One looks African.

00:02:05.990 --> 00:02:07.810 align:middle line:84%
Now, you look at
these two images,

00:02:07.810 --> 00:02:13.270 align:middle line:84%
and you tell me which of
the two faces looks darker?

00:02:13.270 --> 00:02:15.310 align:middle line:84%
To me, the African
face looks darker.

00:02:15.310 --> 00:02:17.150 align:middle line:90%
The African looks dark.

00:02:17.150 --> 00:02:22.910 align:middle line:84%
However, each face has the
same luminance than the other.

00:02:22.910 --> 00:02:26.030 align:middle line:90%
So that's why it's an illusion.

00:02:26.030 --> 00:02:29.890 align:middle line:84%
Now, there may be several
explanations for this,

00:02:29.890 --> 00:02:33.070 align:middle line:84%
but what we thought
is that people

00:02:33.070 --> 00:02:38.570 align:middle line:84%
are likely to look at each
face a bit differently.

00:02:38.570 --> 00:02:41.370 align:middle line:84%
What do you mean
looking differently?

00:02:41.370 --> 00:02:44.450 align:middle line:84%
For example, they might
look at the European face

00:02:44.450 --> 00:02:49.010 align:middle line:84%
in a way that will change
their pupil or the African face

00:02:49.010 --> 00:02:52.560 align:middle line:84%
in a way that would change the
pupil in another direction.

00:02:52.560 --> 00:02:56.900 align:middle line:84%
So if you look at the
images a bit closely,

00:02:56.900 --> 00:03:00.710 align:middle line:84%
you see that the nose
of the European face

00:03:00.710 --> 00:03:04.980 align:middle line:84%
is a bit dark compared to
the nose of the African face.

00:03:04.980 --> 00:03:07.970 align:middle line:84%
If you look at the eyes,
the eyes of the African face

00:03:07.970 --> 00:03:11.930 align:middle line:84%
are a bit brighter than the
eyes of the European face.

00:03:11.930 --> 00:03:17.030 align:middle line:84%
This means that, for
example, if participants

00:03:17.030 --> 00:03:22.250 align:middle line:84%
would tend to look more at
the nose of the European face,

00:03:22.250 --> 00:03:29.870 align:middle line:84%
they would be a little bit more
stimulated by the darker parts

00:03:29.870 --> 00:03:30.800 align:middle line:90%
of the image.

00:03:30.800 --> 00:03:32.510 align:middle line:84%
Whereas if they
tend to look more

00:03:32.510 --> 00:03:34.940 align:middle line:84%
at the eyes of the
African face, there

00:03:34.940 --> 00:03:38.930 align:middle line:84%
would be stimulated more by
the brighter parts of the face,

00:03:38.930 --> 00:03:41.790 align:middle line:84%
and that's exactly
what happened.

00:03:41.790 --> 00:03:43.610 align:middle line:84%
And so for this
study, you looked,

00:03:43.610 --> 00:03:46.190 align:middle line:84%
then used gaze to see where
on the face they were looking?

00:03:46.190 --> 00:03:51.680 align:middle line:84%
Yes, what you have to do in
order to make this efficient

00:03:51.680 --> 00:03:55.200 align:middle line:84%
is to do an area of
interest analysis.

00:03:55.200 --> 00:04:02.250 align:middle line:84%
So what you do is you parse
the images by face parts,

00:04:02.250 --> 00:04:07.040 align:middle line:84%
and then you can compute
the percentage of time

00:04:07.040 --> 00:04:11.690 align:middle line:84%
that the eyes are actually
spending within each

00:04:11.690 --> 00:04:13.490 align:middle line:90%
of these area of interest.

00:04:13.490 --> 00:04:19.100 align:middle line:84%
And then you get a sort
of a statistical frequency

00:04:19.100 --> 00:04:21.470 align:middle line:84%
of looking at parts
of the face that will

00:04:21.470 --> 00:04:25.580 align:middle line:90%
tend to be lighter or darker.

00:04:25.580 --> 00:04:27.780 align:middle line:90%
And these are the results.

00:04:27.780 --> 00:04:31.430 align:middle line:84%
You can see that there
is a bit more tendency

00:04:31.430 --> 00:04:36.620 align:middle line:84%
to look at the nose
for the European face

00:04:36.620 --> 00:04:41.800 align:middle line:84%
and to the eyes for
the African face.

00:04:41.800 --> 00:04:44.950 align:middle line:84%
So this explains
the illusion then?

00:04:44.950 --> 00:04:51.130 align:middle line:84%
Yes, potentially, it can because
it does show that behaviour

00:04:51.130 --> 00:04:56.050 align:middle line:84%
will increase the likelihood
to see the image with a large

00:04:56.050 --> 00:04:58.300 align:middle line:90%
pupil or with a small pupil.

00:04:58.300 --> 00:05:01.510 align:middle line:84%
If you see one image
with the larger pupil,

00:05:01.510 --> 00:05:02.710 align:middle line:90%
it will look brighter.

00:05:02.710 --> 00:05:06.400 align:middle line:84%
If you see the other image
with a slightly smaller pupil,

00:05:06.400 --> 00:05:09.170 align:middle line:90%
it will look darker.

00:05:09.170 --> 00:05:11.870 align:middle line:84%
Well, this is an interesting
way of combining these two

00:05:11.870 --> 00:05:12.650 align:middle line:90%
measures.

00:05:12.650 --> 00:05:15.250 align:middle line:90%
Yes, it's just an example.

00:05:15.250 --> 00:05:26.000 align:middle line:90%