The plotting Package

This package uses several other python packages to make it easy to visualise our brains.

Let's start by importing a brain:

A general piece of advice

In the plotting.histograms and plotting.matrices modules, when you define the output_file argument in the functions, they will return None. So, when you don't define the output_file argument, all the plotting functions in these two modules return a pair with the matplotlib.figure.Figure object and the Axes object. In the next examples you will see that we will get the Figure object so we close it.

The plotting.histograms module

Maybrain makes it easy to plot some histograms by using the python package Matplotlib.

For example, with plot_weight_distribution() you can plot the distribution of the weights in the Brain.G's edges. In case you want to pass some arguments to hist() (matplotlib's function used to plot the histogram under the hood), plot_weight_distribution() accepts any argument (kwargs) that will directly be passed to hist().

The plotting.matrices module

When you have many brains, you can average all the adjMats and have an averaged strength matrix. Maybrain can calculate and plot that matrix with the plot_avg_matrix() function. You just have to pass a dictionary with all the brains.

In our next example we will create a dictionary with just two repeated brains, which practically means nothing, but shows how the function works. The image is very small here in the notebook but if you choose to write the plot to a pdf file (eg. pass an argument output_file="file.pdf"), maybrain will create a pdf file and you can zoom it. You will thus be able to see that the image has the information of the anatomical labels of each node (on the top and on the left). These labels are ordered first by their hemisphere in the brain, then by the lobe and then by the anatomical label. If you want, you have parameters to set up a matrix for brains with a size different of 500x500.

You also have the option to plot the adjMat of a single brain, with the function plot_strength_matrix(). The parameters you can pass to this function are similar to the previous one, with the difference that you can define a title to the plot.

The plotting.connectome module

Maybrain as a wrapper over nilearn in order to visualise brains.

Let's start by applying a simple threshold of just 200 edges for easier visualisation:

The basic usage of Maybrain's plot_connectome function is the following:

You can see that by default nilearn plot is probably not the most adequate for our specific needs. Maybrain's plot_connectome allows customisation as you were using nilearn's directly by providing a kwargs argument. In the following example you can see how it is possible to make our figure bigger, tell nilearn to show the colorbar on the right, and making the letters bigger by calling annotate on the returned display from nilearn.

If you want to see the nodes only without edges, there is a boolean argument for that, only_nodes:

You can make the colours of the nodes different according to the attributes for each. In the beginning of this notebook we loaded information about lobes and hemispheres, so let's visualise it.

In the next example we say that we want to colourise nodes according to the property hemisphere, and make different colours for when that property has the alue R or L in the node's attribute. You can see that the plot has some nodes with a different colour (blue). That happens when the value in the specified property it's not equal no any of the values passed in nodes_attributes.

Note that some nodes labelled as right might not exactly be in the right hemisphere because the way these labels were created was based on the anatomical labels. As some anatomical labels cover the two hemispheres, sometimes the hemisphere information is not exaclty correct for the respective nodes.

If you have more attributes to distinguish, more colours will be created:

Sometimes you want to plot the nodes with different sizes according to a specific property it might have. In that case, you just need to use the argument node_size_property. Here it is an example where we import for each node a property with its degree, and as a consequence the size of each node will vary depending on its degree:

If you don't like to the range of sizes you are seeing, you can define them by yourself. node_size_max will be the size of the node with the biggest property (in our example the degree), and node_size_min will be the size of the node with the smallest property (in our example the degree).