The Brain Connectivity of Anxiety in Preschool-Aged Children
Your hands start to sweat. Your breathing increases. The room is getting smaller and smaller. For many of us, this is a common feeling we get called anxiety. But what if you felt this every day, not at the age of 15, 20, or 35, but instead at the age of 5? For 500,000 children ages 3–5, this is the unfortunate reality. Why?
Introduction to Connectivity
So, in order to understand why this happens, we need to first understand what connectivity is specifically in the brain. Have you ever played connect the dots before? If you have you would know that the goal of the game is to get a connection of four of the same colored markers in a row.
Now, what if I told you that connectivity in the brain was actually pretty similar to that? Even though connectivity in the brain isn’t a game the goal is to make roads of connections. Every experience you have created a new neural network which is a connection in your brain telling you why something happened or reminding you why something happened.
Just like in the game of connecting the dots each of these connections build on with the other connections in order to create a whole system. And this is connectivity in the brain.
Now to go a little bit deeper there are two main different types of connectivity; structural and functional.
Functional Connectivity
Let’s start with functional connectivity. Functional connectivity in the brain is defined as the temporal coincidence of spatially different neurophysical events. Ok, I know that seems like a big definition but let me break it down for you.
Basically, functional connectivity is when two different parts of your brain work together to do something. Think of it like being able to see the stars in the night sky but when you go to take a picture of them with your phone they no longer show up.
Functional connectivity is not a difference with the fibers connecting the neural networks but instead something about how the neural networks interact.
Structural Connectivity
Next, we’re going to look at structural connectivity. Unlike functional connectivity structural connectivity is looking at the physical fibers that make up the connections in your brain.
Basically, think of structural connectivity like examining the roots of a tree in order to understand why it's dying.
For example, researchers have found a biomarker in the brain linked to schizophrenia. This biomarker is a change in structural connectivity.
Now in order to get the best idea of a condition, we need to look at both structural and functional connectivity, which we do in this experiment.
Understanding the Dataset
Of course, the first thing we need to do when preparing to look at connectivity is to examine the data set. This data set was gathered from open neuro.org.
The intent of this data set was to examine whether social anxiety or social phobia of preschool-age children resulted in dysregulation of the amygdala-prefrontal cortex. Essentially, the goal was to determine whether anxiety in young kids resulted in them having mood or psychiatric stress disorders in the future or currently.
In order to test this, a group of children were tested over a five year period, starting at age two to five, and then test it again between ages 6 to 9 and a half to determine the result of their anxiety disorder on their neural responses to angry or sad faces.
This data was collected in an fMRI or a functional MRI. A functional MRI is a specific type of MRI that allows us to see the connectivity in the brain as it is happening.
Now that we understand the data set, we can start looking at what it means.
Using CONN Toolbox
So, to look at the dataset, I used a program called CONN Toolbox. CONN toolbox is a MATLAB program that allows us to perform analysis on fMRI data to observe connectivity! It’s kind of like a magnifying glass with built-in machine learning. 🤯
CONN Toolbox runs analysis in 4 different steps; setup, denoising, analyses, and results. By running the data through these 4 steps, we can get accurate data in easy to see forms, such as a connectivity matrix!
Our next step is to start the setup!
Setting Up the Dataset
Now, we are ready to prepare the dataset for analysis. The first thing we need to do is set up the session. We do this so that we can run all of our fMRI data with accuracy in the program!
To set up the session, we first enter the number of subjects in our dataset, which in this case is 45. Next, we need to enter the number of sessions, or the number of times the data in the set was taken for each subject. Once we do that, we can move on to entering the data!
Remember when we talked about the different types of connectivity? Well, here is where that becomes really important! The first type of data we want to put into the program is structural data.
Structural data is going to have 1 brain in the picture, with bright white veins looking like things in the brain.
Once we have identified the file with the structural data, we then match up the files to the subjects (which should be labeled in the big file you downloaded with the dataset) and move on to functional data!
Now, the functional dataset is going to look slightly different than the structural dataset, because as you remember, it is measuring the activity in the brain instead of the structure of the brain.
The functional dataset is going to look more blurry, and there are going to be 2 brains. This dataset is the one that kind of looks like a blobfish! 🐟🐟🐟
Once you match the functional data with the subject, just like with the structural dataset, you can click done and let the setup run! This part takes about 5 minutes for each subject, so be prepared to wait! ⏲
Denoising: A Bunch of Graphs
You probably just saw the word denoising and started thinking, “What the heck am I looking at?” Honestly, I felt the same way at first too! Denoising is just reducing the noise in your image.
Now, you’re probably wondering why there is sound in an image, but I am here to tell you that it is not actually noise! 🤯 Noise in a picture is just an alteration of brightness and color that creates a grainy look to it. And, the reason we take it out is that it can make our picture blurry and hard to analyze. I told you it wasn’t that bad!
In our program, denoising helps us average out signals, and create a clearer picture!
Denoising is also really helpful when finding connectivity! By clearing out the outliers in our set, we can get an accurate idea of connectivity.
Now that we have denoised the dataset, we can go to analyzing it!
Analysis
Before we jump ahead into the analysis, it is important to know that the analysis is the first level. This means that it is going to be in graphs or pictures that aren’t necessarily finalized. Also, if we had more than one session, we could see the data from each session here.
The analysis is important because it allows us to understand the dataset at a higher level!
The first type of analysis we are going to see is an SBC map. An SBC map stands for seed-based connectivity map, which basically means it’s a map that computes the connectivity (or action) between a point of the brain and another part of the brain.
The computation is going to be represented in the colors of the brain. For example, in the figure below, we see the map of the reference point (0,0,17) and the other point, (91, 127, 90).
We can also see the overall connectivity in a smaller picture, like the one below. The cooler the color, like dark blue, the less connectivity it has with the desired point, and vice versa!
In the next type of analysis, we are going to see a similarity matrix. If you read my article, Connectomics: An Overview, you remember that a connectivity matrix is a graph that shows the roadwork of connectivity!
A similarity matrix is a kind of like a connectivity matrix, but instead of a map, we see a graph! In the middle, there is a dark line that is the point of most similarity in the dataset. As we go out, the colors become lighter as they are less similar.
Similarity matrices help us identify clusters of connectivity in our dataset, like at point (84,-70,18) in subject 36!
Now that we have examined our analyses, we can move onto the results and interpreting the dataset! 😁
Results
So, what are we even looking at in results if we already did an analysis? Analysis, is the overall picture of each subject, separately. It allows us to perfect the dataset to get the best results! 😍 But in results, we can see the information from all the subjects combined. Basically, it is the final product of the dataset!
The first result we are going to see is another similarity matrix! This similarity matrix is going to show us the specific parts of the brain that have the most connectivity.
A few of the clusters from our dataset include; the Cereblunum, the Pallidum and Putamen, the Central Opecular Cortex, and the Gyrus.
Another matrix we can see is a hierarchal clustering map, which looks like this:
By looking at an individual point on the brain, for example, the left side of the amygdala. This allows us to evaluate connections quickly and easily!
Next, we are going to compute the graph theory results, which are going to show us the connectivity matrix, also known as the road map! When we do this, we get a brain with a ton of red dots and lines all over it. Pretty awesome, right?!
Now that we have this matrix, we can select specific parts to see the connectivity of it. Since this dataset is meant to observe the amygdala-prefrontal cortex, that’s where we start!
Here are the connections between the amygdala, thalamus, pallidum, putamen, and hippocampus, all of which are in the prefrontal cortex:
As you can see, the highest area of connectivity is between the right side of the pallium and the left side of the amygdala, which have the boldest line in between them. We also see high connectivity in the thalamus. Isn’t it awesome all the things we can see from just a few dots and lines?!
We can also see the connections in this form inside the brain!
This form allows us to look at the clusters of the connections, instead of just the individual connections.
So now that we have all this information, what do we do with it?
Understanding the Results
Does preschooled-aged social anxiety affect the connectivity between the frontal cortex and the amygdala? This is the question the dataset was meant to answer. So does it? Well, the first thing we are going to want to look at our connectivity maps between the prefrontal cortex and the amygdala.
As we can see, the connections between the amygdala and the hippocampus are very limited. In a normal child, there should be strong connections between the amygdala and the hippocampus, as the connection helps regulate emotions. The lack of connections would result in difficulty regulating and measuring emotions, as expected.
We can also see that the cluster graph shows limited cluster connections in areas in the prefrontal cortex, although it does show a stronger connection to the hippocampus than seen in the connectivity matrix. This could also be a sign of reduced connectivity that would lead to reduced emotional regulation in the child.
Finally, we are going to look at the similairty graph. In the blue circle is our admygla connections, and as you can see, it is not very red. Compared to other regions, like the cerebellum, it has extremely limited connections, which again stregthens the claim that preschool-aged anxiety reduces the connectivity in the admyglda.
What do these results show us? That are claim is true. Preschool-aged anxiety reduces the connectivity in the admyglda and the prefront cortex, which can lead to trouble regulating emotions.
So unfortunatley for those poor little kids, right now we can’t help them. But, hopefully in the future with connectivity, we can reduce the risk of developing a mood disorder because of anxiety or maybe even eliminate it.
TL;DR
Now that we are done, let’s summarize and wrap it up!
- Functional connectivity is the actions between different areas in the brain
- Structral connectivy is the physical fiber connections in the brain
- Our dataset showed the connectivity in preschool-aged children with anxiety
- CONN toolbox is a great tool for analyzing connectivity
- Using CONN toolbox, we confirmed that the dataset prediction was correct
- We evaltated how the graphs and data showed the preschool-aged children have reduced connectivty in the admygdala and prefrontal cortex, which can lead to mood disorders
If you’ve made it this far, thank you! I am a 15-year-old who is interested in regenerative medicine, biocomputing, and public health. If you want to see me continue to grow and 10X myself, sign up for my newsletter here!
