Understanding Event-Related Potentials in Cognitive Psychology: Key Methods and Findings

Hello and welcome to the course basics of experimental design for cognitive psych. ology. I am Dr. Ark Whmer from

the department of cognitive science at IT Kpur. We're starting the week eight of the course where we are continuing to

discuss about different methodologies that utilize experimental research designs. In this week, I will begin with

discussing about event related potentials in today's lecture and a couple of more lectures.

Now, the ERP technique or the event related potentials technique, it provides a powerful method for exploring

the human mind and brain. The technique involves the measurement of electrical activity in the brain contingent to a

presented stimulus or an event and it provides an index of the temporal processing of that stimulus in the

brain. So that is basically why it is considered very important. Whenever you are presenting a stimulus to the brain,

you want to know what is happening post presentation of that stimulus or in some cases how is the brain preparing for

receiving a particular kind of stimulus uh acting on it while the stimulus is still presented and what happens even

after the stimulus is gone. So all the temporal aspects of cognitive processing are best studied using the event related

potentials technique. Now uh before we uh dive further let's have a brief historical review of how

does this technique develop. So the ERP methodology it seems to have been discovered in the early 1930s and has

now progressed over ensuing uh 100 plus years almost. Uh, one of the first interesting events

was in 1929 when Hans Burgger reported a remarkable set of experiments wherein he showed the that the electrical activity

of the human brain could be measured by place placing an electrode on the scalp amplifying that signal and plotting

changes in the voltage over time. So that is basically the first time people discovered that electronic activity in

the brain can actually be measured and this is uh the contribution of Hans Burgger. This electrical activity that

you could record from electrode on the scalp was referred to as an electroenphiloggram.

Electroenphoggram. Over the following years, uh the EEG or the electrophiloggram proved to be very

useful in both scientific and clinical applications. However, if you look at it, the EEG is a very coarse measure of

brain activity and it cannot be used in its raw form to measure the highly specific neural processes and that are

studied in cognitive neuroscience. The kind of questions you want to ask. So there is always electrical activity in

the brain and it's a summation of a bunch of processes that are going on. But if you want to know what is

happening in the brain contingent to a specific stimulus, then you will you cannot work with the raw EEG form. you

will have to devise new methods. You have to devise a more uh you know uh sharper way of analyzing this data which

is basically what we're going to talk about. As I saying uh EEG represents a mixedup

conglomeration of dozens of different neural sources of activity. It is difficult to isolate individual

neurocognitive processes from the electrophiloggram within the EG. For example, there are

neural responses associated with specific sensory, cognitive, and motor events. And it may be possible to

extract these responses from the overall EEG by means of a simple averaging technique. These specific responses that

you can gain or that you can isolate which are contingent to specific uh actions or specific events referred to

as event related potentials. that is these are the electric potentials that are related to specific events or

specific stimuli which the brain might be processing. Now, one of the first early or one of

the earliest ERP recordings from humans was first reported by Pauline Hallowell Davis. Pauline and Halo Davis in the

years 1935 and 36 when the researchers were able to isolate clear ERPs on single trials during periods in which

EEG was relatively stable in a quason state. The first ERP waveforms were published by Galamos and sheets in 1962.

Now it seems there was a bit of a uh stoppage in ERP research during the years of the world war and it took some

time to come up. Uh so the modern era of ERP research actually begins in the year 1964 when Walter and colleagues reported

the first cognitive ERP component and ERP component that was contingent to uh you know neural processing. That

component is referred to as the contingent negative variation or the CNV component. Now, how did they come about

this or how did they discover this? On each trial in their study, subjects were presented with a warning signal just

like let's say a click sound followed by 100 or followed 500 or 1,000 milliseconds later by a target stimulus.

So, there is a warning signal and uh after that the participant is expecting a stimulus and that target stimulus uh

happens around 5, 500 or 1,000 milliseconds later. In the absence of a task, both both the warning signal and

the target elicited uh the sort of sensory ERP response that would be expected for these stimula. However, if

the subjects were required to uh press a button on detecting the target, a large negative voltage was observed at the

frontal electrode sites uh during the period that separated the warning signal and the target. So between the time the

warning signal has come and the target has not yet appeared and the participants are expecting a target

stimulus to appear there is this negative voltage which is being observed in the frontal areas of the brain in the

frontal electrode sides. This negative voltage was referred to as the contingent negative variation and it was

taken to reflect the subject's preparation of the oncoming target. So the subject is preparing to sort of deal

with this oncoming target and process it. So this new finding encouraged many

researchers to begin exploring cognitive components because now it showed a window that the electrical activity in

the brain which we were calling the electrophiloggram or the EEG can actually be uh analyzed well to slice

out specific electrical events in the brain that are tied to specific events that are tied to specific cognitive

operations that the brain does. So another important step in the same direction was the discovery of the P3

component by Sutton and colleagues and uh here the researchers created a situation in which the subject could not

predict whether the next stimulus would be auditory or it would be visual and they found that the stimulus uh elicited

a large positive component that peaked around 300 milliseconds post the presentation of the stimulus. This

positive component 300 that peaks 300 millconds after the presentation of the similars was named as the P3 component

or the P300 component. This component was found to be much reduced when the conditions were such that the subjects

could predict the modality of the stimulus and the difference in brain responses was attributed to the

difference in the predictability or or the unpredictability of the stimulus. So if the stimulus was unpredictable then

the uh P3 component was slightly larger as opposed to when the uh stimulus was predictable that it was going to come in

an auditory modality and the subject is able to predict that. Now these two uh were the first uh you know cognitive

components that could be isolated from the EEG uh recordings and after these and similar findings researchers

basically uh lapped up the ERP research paradigm through the 1970s and in ' 80s and they look took took upon that to

investigate the general exploration of questions that are uh you know required in cognitive neuroscience.

Now let us explore some examples of how the ERP methodology has been used to investigate cognitive processes. Look at

this. First is the classic oddball uh paradigm. I'm sure everybody knows that in a classic oddball paradigm a

particular stimulus is presented again and again and again and uh very rarely another stimulus is presented uh you

know which the participant has to detect and respond to. Okay. So uh in the classic oddball experiment, subjects

viewed sequences containing it of 80% X's and 20% O's and they had to press a particular button for X's and another

button for O. Each letter was presented on the computer monitor for 100 milliseconds followed by a400

milliseconds blank interstimulus interval. So a stimulus will come and then there's a gap of 1400 millconds

another stimulus will come most of the time that is 80% of the time the stimula will be X's and sometimes it will be O's

when the participant detects an X it presents a particular button when a participant detects an O it presents a

it presses a particular button now when the subjects were engaged in performing this task the researchers recorded the

EEG from several electrodes embedded in the electro electrode cap that was placed on their skull

The EG from each site was amplified by a factor of 20,000 and then converted into digital form for storage on the

digitization computer. So you're basically getting the signal and then storing it for further analysis.

Whenever a stimulus was presented, event codes were sent from the stimulation computer to the EEG digitization

computer where they stored it along with the EEG data. So you're basically tagging or marking the EEG data with

respect to what event is going to happen. Now in each recording session, the

researchers viewed the EEG on the digitization computer, but the stimulus elicited ERP responses were typically

too small to discern within the much larger EEG. EEG was recorded at one electro site from one of the subjects

over a period of 9 seconds. The EEG waveform was first recorded from the PZ electrode site which is on the midline

of the parietital loes where the P3 component was supposed found to be the largest. On a closer examination, it was

cleared there was a consistent pattern in the EEG response to each stimulus which became even clearer when the

continuous voltage was converted into a discrete set of samples for storage on uh the computer. So when the signal was

analyzed further it started becoming clear that the there was contingent electrical activity in the brain with

respect to the stimula. Now the EG was being recorded concurrently from about 20 electrodes placed in the 1020 system

according to the uh protocol set by the American NCilographic society and the system names each electrode using one or

two letters to indicate the general brain region. Say for example FNP for frontal pole, F for frontal, C for

central, P for parietal and O for oxipital and T for temporal. So these are basically how the names of these

components are kept or names of these electrodes are kept actually and a number to indicate the hemisphere uh and

a distance from the midline. So for example, a lowerase Z is is used to represent the number zero which uh

basically indicates that the electrode was placed along the midline of the scalp. All right. So thus f_sub_3 lies

over the frontal cortex to the left of the midline whereas fz lies over the frontal cortex on the midline whereas

f_sub_4 lies on the frontal cortex to the right of the midline. So this is again just some convention that we ought

to be aware of. This is basically the setup. You can see that the uh subject uh an electrode cap is placed on the

subject. These are the active electrodes. There is this reference electrodes. There is the stimulation

computer and monitor. All of these feed into the uh go through the filters and amplifier go to the digitization

computer where the analysis is taking place. These are the different electrode sites uh you know the 20 of them from

which the data is being observed. Here you can see we have zoomed in this region of of the particular signal and

this is basically how that uh signal sort of comes across. You can see this is the average of 80 x's and this is the

average of 20 O's. You can see that there is a difference in the peaks that is observed contingent to X's and

contingent to O's. Now at the end of each recording session, the researchers were performing a simple signal

averaging procedure to extract the ERP waveform elicited by the X's and the ERP waveform that was elicited by the O. The

basic idea that uh was that recording uh eg contains the brain responses to the stimulus plus also other activity that

may be unrelated to the stimulus which basically therefore needs to be extracted to yield a consistent response

by averaging all the trials which had x's and all the trials which had O's. To achieve this, the researchers extracted

the segment of the eg around each x and o and lined up these eg segments with respect to the event codes that mark the

onset of each stimulus. starting from when the stimulus was supposed to come till when the stimulus

has passed they carved out those segments and then they analyze those segments. So after this these were

simply averaged together uh uh you know over the single trial EG waveforms creating one average ERP waveform for

the X's and one average ERP waveform for the O's at each electrode site. So this is what you can see at each electro site

you get particular uh you know waveform for X's and another particular waveform for the O's.

So for the at the uh the voltage at 24 millconds in the average X waveform was computed by taking the voltage that was

measured 24 milliseconds after each X stimulus and averaging all of these voltages together. So this is just you

know a bit of a processing uh you know convention on how this data is uh cleaned. Any brain activity that was

consistently elicited by the stimulus at that time will basically still in the average. However any voltages that were

unrelated to the stimulus will be negative on some trials and will be positive on some trials and therefore

they will c cancel out during the averaging. The resulting average waveform basically will be consisting of

a sequence of positive and negative voltage deflections which we can call peaks and waves or components. And you

can see here that these are named separately. So you have uh P1, you have P1, N1, P2, N2 and P3. So these are the

peaks. This is a first positive peak, second positive peak, third positive peak and the first negative peak and the

second negative peak. This is basically by convention how these waveforms are actually labeled. Alternatively, the

number may indicate the latency of the peak in milliseconds. So sometimes you will also come across ERP components

that have the number which tells the latency or the time period of when that peak is observed. So here we see P1, N1,

P2, N2 and so on. But you can also have a component such as the N170 which basically says that this is a negative

peak which peaks at negative component which peaks at around 170 mconds post stimulus. All right.

Now components may also be given uh paradigm or function based name such as sometimes there are components which are

referred to as the error related negativity or the no go into which is basically observed on the no-go trials

in go no-go experiments. The sequence of ERPs uh peaks basically reflects the flow of information through the brain

the temporality basically. So what how does the brain react first and how does it react second. So all of these peaks

P1 P2 P3 first positive response second. So in in uh chronology P1 will come first, P2 comes second and P3 comes

later. So the sequence of ERP peaks basically reflects the flow of information throughout the brain and the

voltage at each time. uh when the VRP ERP waveform reflects the brain activity that was going on at that precise moment

in time. Now in the experiment that we have just seen the infrequent O stimula elicit a much larger uh P3 wave than the

frequent X uh stimulus a finding that has been replicated with thousands of previous oddball experiments. So you can

look here the infrequent O basically leads to a much larger positive peak which goes over 10. The rare O's

actually uh reflect a much larger P3 component than the very frequent X's. So this is some uh the finding this is a

critical finding in the oddball paradigm and this is something that has been replicated and uh repeated across

several experiments. More specifically, the P3 wave was also at the uh was also largest at the PZ

electrode uh but could be seen at all 20 electrode. So it was observed at all 20 electrodes but it was found to be

largest at the PZ electrode. The P1 wave in contrast was the largest at the lateral occipital uh electrode sites and

was not present in the frontal site. So this is again uh you can differentiate between which are the peaks, where are

they appearing and what time courses they are uh peaking or not peaking. Now here you can see that each ERP component

has a distinctive scalp distribution that reflects the location of the patch of cortex in which it was originally

distributed. So each uh uh ERP component basically has a proper distribution that basically tells us the location of the

scalp where it is coming actually from where it is uh originated from. Now this is about the oddball paradigm. Let's

look at another uh very important ERP paradigm. This these are just examples that we are taking to illustrate how the

ERPs are collected and how they are able to tell something about the brain. Now another uh very interesting experiment

is there which yields a specific N170 component that is a face related component typically peing around 170

milliseconds after stimulus onset and is largest over the vententral areas of the behind areas of the visual cortex. Now,

in a typical uh experiment, in a typical N170 paradigm, photographs of faces and other types of non-face objects such as

houses or cars are flashed on a computer monitor while the subjects are passively viewing the stimula. Okay. The X

presents the time relative to the stimulus onset and the Y presents the magnitude of the neural response. You

can see here the x is the time here and the uh you know uh y-axis is presenting the response in in microvolts that is

contingent to a particular response. Okay. In the scalp map is shown the shading indicates the voltage measured

at each electrode side during the time period of the N 170 response. So you can see here uh there there are these

different shading. They basically tell us uh the amount of voltage that is experienced at different sites when this

particular uh thing when this particular component is isolated. You can see this one is the N 170. It is a negative peak

happening at around 170 milliseconds. Peaking at around 170 millconds. You can see here uh and this is uh happening in

response to face stimula but not in response to any other kinds of stimula. So this is notable because it is larger

when elicit when the eliciting stimulus is a face compared to when the stimulus is a non-face object such as a car or a

house or anything. uh the difference between faces and non-face objects begins approximately it starts at around

150 millconds after the onset of the stimulus and peaks at around uh 170 mconds and then sort of recedes back

which basically would allow us to conclude that the human brain is able to distinguish uh distinguish between faces

and other objects within a very short period of 150 milliseconds. The scalp distribution helps us to know that this

is the same component that is observed in other similar studies of uh the N170 and it suggests that the N170 generator

basically lies somewhere in the visual cortex in the back part in the vententral part of the visual cortex.

Now several researchers have used this N170 uh paradigm and they have basically used it to address interesting questions

about how faces are processed in the brain. For example, some studies have asked whether face processing is

automatic uh by testing whether uh face elicited N170 is smaller when the faces are ignored. So uh if the face is there

and you're not consciously processing it, does the N170 appear again? So that is basically what some people wanted to

look at and the results of these experiments actually indicated that phase processing is at least partially

automatic but it can be modulated under circum some circumstances. So when the PE faces are being uh actively attended

then the nature of the peak versus when the faces are partially ignored is slightly different from each other.

Other studies have used the N170 to ask whether faces are processed in a specialized face module or whether the

same neural processing is also used when PE when people process other sorts of complex uh stimula uh for which they

have extensive expertise. For example, so expertise. So some people have linked the N170 to the role of expertise as

well and therefore very interesting experiments have happened uh and it has been shown that bird experts exhibit an

enhanced one N170 in response to birds. Dog experts exhibit an enhanced N170 in response to dogs and fingerprint experts

exhibit uh an enhanced N170 in response to fingerprints. So there must be some role in 170 with regards to our

expertise in processing faces. That is why it basically is elicited by faces. So that is all uh that I wanted to talk

about in this lecture. We have just initiated a bit on uh uh you know uh ERPs. In the next two lectures I'll take

this uh uh discussion a little bit further. Thank you.