Understanding Neuroimaging Methods in Cognitive Psychology Research

Hello and welcome to the course basics of experimental design for cognitive psychology. I am Dr. Arma from the

department of cognitive science at ID Kpur. We are in the final week of the course where we are discussing different

methodologies that have been used for experimental research and the last three lectures we talked about ERP technique

and in the last two lectures of this week we are going to talk about neuromaging.

Now while the methods which basically mainly rely on the measurement of either structure or the functioning of the

brain they have their own merit. More recently, cognitive neuroscientists have been excited about methods which provide

a joint measurement of the structural analysis and the functioning of the brain. So, which areas of the brain are

uh working and uh completing a task and where these areas are. So, both temporal and uh uh spatial resolution of whatever

goes on in the brain while the brain is processing a particular stimulus is of interest to a lot of researchers. These

methods are capable of continuously measuring physiological changes in the brain as the function of person's

perceptions, thoughts, feelings and actions. So for example, when an individual is processing particular

stimula or you've given him a task to think about, imagine something or the person is by themselves uh you know

thinking and engaging in certain kinds of thought processes, there are uh contingent physiological changes that

are going on in the brain. So the methods that we're going to talk about in the last two lectures are the methods

that basically can map these physiological changes in the brain to the cognitive activity or to the mental

processes that are taking on in the person of the in the brain of the person. These methods are uh very common

say for example the posetron emission tomography that is the PET method and the functional magnetic resonance

imaging or the fMRI method. We'll talk about these in some detail. Together these methods have been found capable of

detecting metabolic changes in the brain while uh the participant is actually performing cognitive tasks. So you can

give a person a task for uh for recognizing particular faces recognizing uh you know which letter a particular

word starts from mentally uh you know uh verbalizing certain words all of those kinds of things. While the brain

obviously when it is making these actions certain physiological changes are going on in the brain. So when these

uh physiological changes are going on in the brain and these methods are used, they enable the researchers to identify

the very specific regions of the brain that are involved in performing these mental uh operations invoked in these

different cognitive tasks. So PET and fMRI basically they measure metabolic changes correlated with neural activity

uh although it is h it happens in a in a bit of an indirect manner. Okay. So for example, how when an area of the brain

is active, it uses up more oxygen and glucose by directing more blood flow towards itself as opposed to the less

active parts of the brain. So when a particular area in the brain is active, it is working then it basically directs

more oxygenated blood towards itself taking it away from the other less active areas of the brain. Both PET and

fMRI can measure these relative changes in the blood flow. So this is called RCBF.

uh and hence it is able to these uh techniques are able to identify the areas of the brain that are involved in

performing specific tasks. So for example, if you've given somebody a task of recognizing faces, you can expect

that there will be more uh blood flow towards the uh you know vententral temporal areas of the brain which where

the fusififor gyus lies and that will give you a hint of okay this is the area of the brain that is involved in

performing this particular task. Now, positon emission tomography. Let's uh look at this in some detail. Uh PET

activation studies are able to measure local variations in the cerebral blood flow that are correlated with mental

activity. A radioactive substance is introduced into the bloodstream. Radiation emitting from this radioactive

tracer can then be monitored by the PET instrument. So to perform this test a radioactive isotope uh which is uh

obviously not harmful for the individuals is inserted into the bloodstream and this along with the

blood flow it enters the brain and whenever the brain is involved in particular activity and the certain

areas of the brain start receiving blood this radioactive tracer reaches those areas and we can basically trace that

these areas of the brain are active by tracing the radioactive emittance of this tracer. So the injected radioactive

isotope uh I'll just describe it in more detail. The injected radioactive isotopes rapidly decay by emitting a

posetron from their atomic nuclei. When a posetron collides with an electron, two photons or gamma rays are created.

The two photons then move in opposite directions at the speed of light passing unimpeded through the brain tissue, the

skull and the scalp. The PT scanner determines where the collision took place. As these traces are in the blood,

uh a reconstructed image shows the distribution of the blood flow and where there is more blood flow, there is the

possibility of more radiation. The PT studies can actually measure relative but not absolute metabolic

activity. In a PT study, the radioactive tracer is injected at least two times. once during the experimental condition

and it charts the blood flow during the experimental condition and another time during a control condition where again

the blood flow is charted and then the two are compared. The results are reported as a change in what is called

the regional cerebral blood flow as a comparison between the experimental and the control conditions. Now the PT

scanners are actually capable of resolving metabolic activity to regions of high resolution up to 5 to 10 mm of

tissue and it is therefore sufficient to identify both cortical and subcortical brain areas which were involved in a

particular task or which were invoked during particular mental functions. Now moving on to functional magnetic

resonance imaging or fMRI. In case of fMRI, the imaging uh basically utilizes the magnetic properties of the

deoxxygenated form of uh hemoglobin that is called deoxyhemoglobin. The fMRI detectors measure the ratio of

oxygenated to deoxxygenated hemoglobin referred to as the blood oxygen level dependent response or the bold effect.

FMRA results are reported as an increase in the ratio of oxygenated to deoxxygenated hemoglobin. So the idea is

when a brain is working when it is performing some mental functions there will be more oxygen it pushs more

oxygenated blood towards itself and the deoxxygenated blood reduces. So the ratio of you know the oxygenated blood

to the deoxxygenated blood in the active area of the brain will be more. As an area of the brain becomes active the

amount of blood flow being directed to that area is increased causing a change in this ratio. As I was just explaining,

fMRI studies basically measure the time course of this change. How does this change uh happen gradually uh which

happens over a period of few seconds almost 16 seconds uh it takes uh and peaks around 6 to 10 seconds later uh

when the brain is involved say for example when a brain was processing a given stimulus. So fMRI can be used to

obtain an indirect measure of neuronal activity by measuring changes in the uh blood flow. It is more advantageous as

compared to PET on several accounts because MRI scanners are much less expensive and they are relatively easier

to maintain. It also does not involve the injection of uh you know a radioactive tracer which basically saves

any expenses of handling such a substance and is noninvasive and you'll be you know finding subjects more easily

for fMRI studies than for PET studies. Now both these methods have some limitations. Before I move on to a

particular experimental design uh protocol with fMRI, let's broadly uh you know uh survey the limitations of these

methods as well. Now both of these methodologies actually have poor temporal resolution. Remember in the

previous lectures we've talked about the strength of ERP method which is the great temporal resolution it has. It can

give you moment to moment time to time analysis of how the brain works contingent to a given stimulus. However,

these neuroming methods have actually very poor temporal resolution. And so neither of these two methods are capable

of providing us with a chronology of mental operations or how the brain is supposedly processing a given stimulus

or uh performing a particular cognitive operation. The data obtained from both of these methods are massive. There's

huge data that is generated and several patterns of differences do emerge between the experimental and control

sessions which basically make it very difficult to interpret the data. See it is mostly correlational and not causal.

So we'll talk about this also when we're talking about experimental design in fMRI.

Also it is difficult to make concrete differences about each of the activated areas functional contributions from the

neural imaging data. uh while it is possible that during a particular mental function area A, B and C are activated.

Uh because the temporal resolution is is not very good, it might not be possible to know the exact chronology. Also, it

might not be uh you know easy to be able to uh infer a causal role of either area A or B or C in the menal operation. So

while you can expect to understand that the regional cerebral blood flow increase increased or the uh ratio of

oxygenated to deoxxygenated blood increase in specific areas there is no way that we can deduce causality and can

say that oh this particular area is definitely the area that is involved in this specific mental operation. So that

is a bit of a drawback of these neuroming techniques because several areas may be activated during a

particular task but their relative contributions will not be very clear. Uh related or sort of a building on these

there is this technique called making brain graphs. Now we have mentioned earlier that it is important to look at

the functioning of the whole brain. The overall brain how it is acting how it is interacting with a given stimulus as

well as a network of connected neurons. Okay. Now one of the methods that could help uh neuroscientists to understand

connections of the brain and the patterns of flow uh of flow of information. Brain graph basically is a

visual model of the connections between within some parts of the nervous system. So how does information flow from this

part of the nervous system to that that part of this part of the brain to that other part? This model is typically made

up of nodes which are individual neural elements and edges which are the connections between these different

elements. So you can think of this is that particular this is a particular area of the brain this another is a

particular area of the brain and they're connected through these edges. The geometric relationships of the nodes and

the edges define the graph and they provide a visualization of brain organization. So that becomes very

useful. Neuroscientists can actually work on this and they can create certain brain graphs by putting together data

from several of these neuroiming methods PET, fMRI, diffusion tensor imaging and so on. The selected data set determines

the edges and nodes and the edges of the network. So which are the prominent areas how are they connected and so on.

For the human brain given its billions of neurons the nodes and the edges of a graph can be taken to represent

anatomical or functionally defined units. So you can basically say okay I am charting the brain as it is

anatomically structured or as it is functionally structured. Which area receives the input first and then gives

to the other area and so on. In this manner what can happen is that researchers can differentiate between

nodes that act as hubs sharing links with many neighboring nodes and nodes that act as connectors providing links

to many more distinct cultures uh distinct uh clusters. Beyond simply showing the edges, a brain graph can

also depict the relative strength and the waiting of those edges. Basically showing which connections are more

important and which connections are less important in this uh you know hypothetical sort of map of the brain.

Brain graphs have actually been found useful uh to compare results from different experimental methods. For

instance, graphs based on anatomical messages, anatomical methods such as diffusion tensor imaging can be compared

with graphs that are obtained based on fMRI and you can basically see how structurally and functionally the

particular hypothesis that you have tested can actually uh you know compare against each other. Brain graphs also

provide us with the means to visualize the organizational properties of the different neural networks. For example,

the network for vision or the network for memory or the network for uh spatial location and so on. It is therefore it

will give you a very good idea of how information flow uh flows from one part or the other and how it is organized in

terms of the brain. The final method that is uh you know uh extremely important deriving mainly from

these neuro imaging meth methods is that of computational modeling. So computer-based models of human brain

have also been created to stimulate the processes of the brain and the connectivity between its neurons.

Simulations are designed to mimic behavior and the mental operations supporting that behavior. A computer

typically is provided with an input then it must perform similar mental functions to create a behavior. And basically what

the these people are doing is they are uh drawing comparisons between simulated behavior and actual human behavior. In

order for the computer to be able to successfully mimic human behavior, the modeler will need to specify how the

information is represented in this computational model. How it is going to be transformed within the program. This

is typically made possible through concrete estimates or hypothesis about how the brain really uh performs these

functions. The success and the failure of these different models typically yields insights about the strengths and

weaknesses of the theories and this is a very good method of testing hypothesis about how does the brain really

function. So this is all about uh neuro imaging methods that I wanted to say and in the

next lecture I will talk to you about uh design issues related to a neuro imaging technique that is fMRI. Thank you.