4.

2 Case: Kirk et al + Ramsøy EEG, metode = EEG

Hello, this is the 2 studies and a method talking about Emotions and Feelings.
So, we will talk to the topic emotions and feelings, two studies on that, and we
will go through a method. This method would be, would be ED or
electroencephalogram. Which is the method that we're using, electrodes
positioned on people's head. What I'm about to talk about first is the study by
and colleagues Martin Scolle and are a part of my collaborators in Copenhagen
as well. They did a study some time ago now, it's 2009 in neuro-image. And the
purpose of the study was to see, first of all whether they could affect the way in
which people's preference for abstract art could be affected by mere contextual
information. And the other part of that study was to study the effects that had,
the causes that had on brain activation. So this was an fMRI study a study
where they had participants coming into the scanner and, and their task was to
look at abstract art paintings and rate their preference for the art. Now the
cover story here, was that the researchers have been able to, to produce
images or to get images from a prestigious art, art gallery in the vicinity of
Copenhagen called Louisiana. And the problem was that they didn't get
enough images. So they had to produce some images themselves from, either
from the web or make the abstract art themselves, from unknown artists or
from themselves. So the cover story here was that either your looking at art
that is famous or art that is unfamous. Of course all of these paintings, all the
abstract art paintings were unknown and nothing was from a prestigious art,
art gallery. That was just a cover story to, to see if that affected people's
preference. And as you can see in the middle bar chart here, you can see that if
people believed that the abstract art was from a gallery, from the abstract art
gallery, than their preference was significantly higher than if they thought that
it was made on the computer or gotten through the internet. So, it's much
higher. The thing that they also found was that using fMRI to scan people's
brain at the same time. They discovered that the part of the brain that is shown
here in, as white dots on the right side, is the medial orbitofrontal cortex. So
the medial side of the bottom part of the, of the frontal cortex, the medial of the
frontal cortex. Strong activation in this part of the brain was related to people's
preference, and they found out the gallery, of what people, when people believe
that, they were looking at African art from the gallery. It had a much stronger
activation level, in this part of the brain, than compared to if they believe that it
was just made on the computer or gotten from the web. What this study shows
us is that our preference, first of all can be modulated by context, contextual
information. And the other if insight we get from this is that this seems to take
place in this part of the brain that we are knowing as the autofrontal cortex, the
medial, the middle side of that. The part of this study also demonstrates that
this medial part of the autofrontal cortex is also well known to be a part of the
brain that is generally responding to preference. So people's enjoyment of
food. Of people's faces, for example, of music, of odors and so forth. We see
that stronger activation in this part of the brain correlates very significantly
with people's preferences. So, this goes in, right into that mystery. The
interesting thing is, of course, that this seems to be a neural, call it, of liking, if
you like. And another study that was done by myself and that is currently
under publication but it has been published and, and presented at conferences,
is what we can say, call the neural predictors of choice or even, even more
than that, we can call it the. Neural predictors of willingness to pay. In this
study, we had people look at particular products and after looking at products.
And after looking for particular product for several seconds, they made up their
minds how much they were willing to pay for the product. They were imbued
with money before going into the lab and going through the test. So they had
money they could spend or save. And then, some of those choices were going
to be realized. What we were looking at was using EEG, and at this part of, this
study was actually using a what we're going to call a lightweight or low-
resolution EEG. It has some challenges when it comes to certain resolutions
and certain kinds of measurements. But for this purpose we wanted to try this
out with a, with this kind of tool to see if that can provide some good insights
and information and even predictors of what people are going to choose. What
we were looking at was the so called frontal asymmetry score. We have alluded
to this early in the course, and we'll describe it briefly. Basically is that when
we see strong activation in the left frontal cortex as opposed to the right frontal
cortex we see this to be highly related, highly predictive of approach
behaviors. The inverse pattern, the right vers, right bigger than left activation
seems to be related to avoidance behavior. Although, I have to say that, that
relationship is currently being studied a bit more. But at least this frontal
asymmetry has something like 20 to 25 years on it's back in terms of scientific
credentials. From psychiatry to neurology to neuroscience and so forth. And it
seems to be a pretty robust phenomenon that might, as far as we understand,
might relate to some deeper structures of the brain. The brain stem regions
that has some asymmetry e, engagements when it comes to the sympathetic
versus the parasympathetic nervous system. So using this asymmetry score,
we were able to look at particular frequencies. We were particularly interested
in the gamma range, the high frequency range. And what we were interested in
was whether the asymmetry score in the gamma range could be used to
predict what people are going to do. In particular, whether the gamma frontal
asymmetry during product viewing alone, and several seconds before people
actually made up their minds, could predict what people are actually going to
purchase. And how much they were willing to pay for that. What we found was
a pretty signficant stat statistical sig, significance effect, in that the stronger
left versus right symmetry score during product viewing, the more likely
people were going to buy, but also that more people were going to be willing to
pay for the product. So this demonstrates that using a relatively low resolution
EEG system, can allow us to make certain predictions and, and even
understanding certain motivational processes in consumer choice. Now this
motivation asymmetry score, or the motivation score is currently being used
by multiple companies, by multiple academic institutions and we're trying to
both understand the basic phenomenon, but also to what extent it can be used
to predict behavior. When we're talking about EEG. We know that it is called
electroencephalography or the electro, electroencephalogram. We know that
when we are measuring brain activation using electrodes put outside the scalp,
we can see several discharges. We can see the fluctuations as we see here on
the top left corner, for example. Several small and large fluctuations. We know
also, unfortunately that many of those fluctuations may be due to openings, to
teeth clenching, to talking and so forth. It's just a muscle activation of the, of
the of the head. so, that needs to be sorted out. That needs to be taken away
from the signal. When that is done, we have a better approximation to
understand the neural discharges, the brain signals so to speak. And as we've
seen earlier in this course what the brain actually does is also both can be
chemical, have chemical communications and also have electrical discharges.
It has been found that the EEG signal mainly stems from what we call the
dendritical activations, the dendrites, the receiving ends of the nervous system
or the nerve the neuron per se. That suggests that what we are measuring
mostly is the incoming signal to a region. If we look at different discharges at
the bottom here, for example, we can easily pick up things like epileptic
seizures, for example. We can see that as the erratic movement here a couple
of times here on all channels. Every single line there, is an electrode that is put
outside on the scalp. Of course there are ways to do intra-cranial EEG, such as
putting electrodes inside the skull during surgery, for example. But of course,
this is not feasible for consumer research. So, we'll have to look at the EEG the
extracranial measurements. Putting electrodes on people's head. Currently, the
most used method is, is using gel. So, a conductive gel that conducts the
electricity from the skull onto electrodes. But there are certain several dev,
developments in terms of. Using dry electrodes so-called dry electrodes that
allow us to not use gel any more, just basically put the equipment on people's
head. So this is a, an illustration using you know, a high definition grid of
different electrodes on the scalp of the person. What you see is that it's, it's
highly dense, it's pretty dense. I think this is a 128 channel system, or a 64
channel system. What you can see is that from the raw signal at the bottom,
the b version here, you can that there are several fluctuations across all the
channels here. What we often can see, some people do, some researchers are
doing, and some commercial companies are doing is using what we, what is
called the event potent event related potentials or evoke potentials. What that
means is that, we have a highly systematized way of showing stimuli. So we
present an image at one particular time we repeat the same image or a similar
image several times to generate a, an average response profile from that
signal, from that stimulus. So, as you can see, in the orange here in the b this
is the time points where we are showing let's say a particular, a particular
image. And, what you can see then is that you have eight or nine different time
points where you have shown that. What you then do is to make the average
signal over time the first second or so. And you get a, what we could call, a
response profile. And this profile you can see on the top right corner in the, in
the c, str, figure here. What this allows us to do is to look at the responses of
different peaks and valleys in terms of the, the changes in the discharges. And
there are several there's a whole literature on showing what those typically
mean. We know that at approximately 400 milliseconds after the onset of the
stimulus, there is typically some kind of semantic and understanding of the
stimulus itself. So changes in words or misplaced words might actually lead to
a deflection of the, what's called the, the N400 signal. And finally, a, another
way of doing EEG is to look at different changes in frequency. As you can see
at the bottom right, there are ways in which you can look at the high frequency
or the low frequency, or the middle frequency. The low frequency being the
delta, for example, is the slow wave. You typically see the delta activation very,
really high when people are in deep sleep. As opposed to that you see that the
beta and the gamma tends to be this the they are always the high frequency
bands. And they tend to be related to high engagement of of brain activation.
And also, you know co, communication across different areas of the brain as
well. So integrative activation we could call it. This is a a, a 2D map put on the
scalp o, of a of a, kind of a, a template scal, scalp showing the spatial
distribution of the signal as well. And today there are even ways that you can
reconstruct the EEG; signal relatively well so you can, you can pinpoint where
in the brain, especially in the cortical mantle, but also some deeper regions.
You can reconstruct the signal to find out not only grossly where you are, but
also pinpoint certain structures where the activation has been has been most
active. There are two ways to do an EEG, at least. One way is to have the lab
setting where you have people entering a lab. They are mounted with EEG
equipment. This is something you can do where you, they sit in front of a
screen. You can present them videos and ads, and print ads. They can browse
webpages, they can look at TV, drama for example. So this is kind of a
stationary setting. And the other effect can be in a so-called mobile setting. So
here we have a, an example from a lab study sort of an in store setting where
we these are just role data looking at the the eye tracking results on the left
here, and you can also see the raw brain signals coming from the EEG headset
while people are walking inside the store. Of course, you need to control for
certain things such as walking, which produces a lot of artifacts in the data.
There are ways to correct for that that are becoming available these days. So
this allows now more and more researchers to not only test in highly
constricted lab environments, but also to test people inside store
environments, in amusement parks, while they're even running. They can be
tested in augmented in virtual reality setups. And there is just today more and
more ways in which you can do these kinds of studies. So today EEG is in
many ways, becoming a, a an increasingly more interesting solution for the
both the academic, but also the commercial neuromarketer and consumer
neuroscientists. As I mentioned, in the stationary setup, it's possible to do a
test of of TV drama, and using the EEG you can generate certain responses.
You can look at the changes in workload, for example. If people are working,
trying to basically understand very much the, hold on to the information. Which
is related to working memory performance, for example. You could look at, as I
mentioned, the motivation score, the, the changes in approach and avoidance
behavior. You can also look at changes in arousal, for example. And of course,
there are several ways in which you can use the EEG to, to assess different
kinds of responses, both to specific events, but also over time. As we can see
here, this is a, a response from a number of people, I think we were testing 30
or 40 people, in a lab setting testing people's responses to TV drama. On the
top you can see the eye tracking results. The heat map showing where people
are mainly looking. And at the bottom you can see the aggregate scores over
time for motivation. So this is a symmetry score for cognitive load which is
highly related to, or is using the so called frontal So a activation in the frontal
lobes which is related to working memory. And what's called the emotional
engagement which is an arousal index which is highly related to people's
people relation, their changes in respiration and pulse for example. So these
are ways that you can use in which you can EEG, that are becoming more and
more available for you to use. I would say, of course there are no such things
as just a turn key solution for this. It's not easy to just get started, and pick a
solution off the shelf. But it's becoming more and more available and more,
and more easy to start at least gather the data. And then of course you need to
have expertise to correct the data, to set up the study correctly. To pre process
the data so else it's called, to clean the data from artifacts and so forth. And
then of course you need to know how to do the analysis. But it is becoming
more and more available. The pricing is become much more slower than it's
been during the past decade. So we've seen that EEG is becoming really one of
those good tools to use for the neuromarketer, the commercial neuromarketer,
but also for academics who also want to study the brain responses during
consumer behavior. And, finally just let me run up with a recent study by and
colleagues, that was produced this summer. They are taking a different look at
the EEG response. They were looking at how synchronized a group, a small
group of people they only tested something like 16 people. And how
synchronized how coherent they their brain responses were to specific
episodes of this was the the TV series Walking Dead, but also two
commercials. What they are able to do is to generate a, a index of group
coherencies. So, if you can imagine that if everyone's looking at a horror movie
and all of a sudden there is a, a surprising event and everybody's jumping in
their chair. Now this, this indicates that everyone is, is is basically responding
in very much the same way. So that is a highly coherent response in the group.
As opposed to when people are looking all over the place and people are
responding very differently people lose attention and things, you start thinking
about other things, then there's a low coherency in the group. By looking at the
coherency in the EEG, the researchers were able to have a strong correlation, a
high correlation with, that also means prediction, of the number of tweets that
people had to the episode, that people were looking at. But also, the so called
Neilson ratings. The ratings from one of those major market research
companies that track peop, how many people are actually looking at the
episodes and how well they are actually following that episode. And also the
TV Ads, so it's the powerfull performance of the Ad. So that basically means
that just by look at a small group of 16 people, and looking at how coherently
they are responding, it possible to now predict with high accuracy. the, the
cultural effect, the cultural response to those ads.