Explore the brain in 3D at high resolution, identify cell processes, and recognize synapses
WEBKNOSSOS is a great tool to visualize and annotate images of the brain, and it’s free! We believe it can be especially helpful for teachers and professors who want to show their students features of the brain interactively and in context, using volume electron microscopy images.
For that, we have designed two exercises: the first exercise focuses on recognizing cell processes in EM data, while the second exercise involves identifying synapses and exploring neurons and synapses in 3D. The dataset used for these exercises represents a portion of layer 4 of the mouse cortex. It was acquired by Motta et al. (Science, 2019) and reconstructed by scalable minds with AI.
EXERCISE 1
Learn to identify and recognize different features of the brain on electron microscopy images
Exploring structures of the brain
First, open this link: https://webknossos.org/links/gubV8GGnBYw7agqi. You are now seeing the WEBKNOSSOS interface. Have a look at this image to familiarize yourself with the user interface.
In the middle, you’ll find the viewports (XY, YZ, XZ, and 3D) to observe microscopy data and annotations. On the left are the image layers (microscopy, segmentation, skeleton, etc.). On the right are tabs such as Info, Segments (your annotated segments), and Skeleton (your skeleton trees). At the top is the toolbar with tools for annotating your data (brush, eraser, etc.).
In the right-side panel, make sure you are in the Skeleton tab. You will see different groups (folders) and items within them. Each group corresponds to a structure, such as mitochondria, nucleus, myelin, vesicles, etc.
Start by expanding a group, for example “Mitochondria”, by clicking on the “+” symbol near the group name (see video below). Clicking on an entry will automatically navigate the viewports with the electron microscopy data to the location of the node. Zoom in and out using ‘I/O’ or ‘Ctrl/Cmd + mouse wheel’. You are now observing a mitochondrion in EM data. Use ‘D/F’ or the mouse wheel to move through the Z direction and observe the organelles in their 3D context. Click through all the entries to familiarize yourself with these structures and how they appear in EM data. The following structures have been annotated:
nucleus
mitochondria
myelin
vesicles
endoplasmic reticulum
In the Skeleton tab, expanding the group “Mitochondria”, selecting entries, zooming and scrolling through the EM slices with the keyboard shortcuts I/O and D/F.
Annotating structures of the brain
Now, it’s your turn! Click “Copy to my account” and create a free account. Take some time to get familiar with the WEBKNOSSOS interface. Find more information on how to use WEBKNOSSOS here.
You are now going to place more nodes on similar structures. To do this, select the Skeleton tool and activate the Single Tree mode (see the videos below). Then, go to the respective feature groups and annotate more of the features on the EM images. When you are done, send a link to your teacher so they can review your work.
Selecting the Skeleton tool, activating “Single Tree Mode”, and annotating a mitochondria.
Annotating further mitochondria and renaming one entry in the skeleton tab.
EXCERCISE 2
Learn to identify synapses and explore the synaptic partners in 3D
Open this annotation and look at the EM data. In this annotation, the neurons have been reconstructed automatically with machine learning by scalable minds. Each cell has a different color. You can toggle the Segmentation Layer in the left-side panel (or use the keyboard shortcut ‘3’) and explore the available settings.
Learn to identify synapses
To identify a synapse in EM data, start by looking for a cluster of synaptic vesicles, which are small, round structures that store neurotransmitters. If you observe the vesicles docking at a contact surface between two neurons, it indicates a potential synapse. Another clue is the appearance of a darker area attached to the post-synaptic membrane, known as the post-synaptic density.
On this image, extracted from a Mouse dataset acquired with SBEM by S. Loomba et al. (Science. 23 June 2022 10.1126/science.abo0924), you can observe these features.
Synapse on EM data.
Now, go to your WEBKNOSSOS annotation and click “Copy to my account”. Click on the 3 synapses that are labeled in the skeleton tab. Explore them in their contexts and try to understand why they have been identified as synapses. Toggle the “Segmentation layer” to have a better look by seeing the raw EM.
Going to the skeleton tab, clicking on the first synapse, exploring it in its context, and turning the Segmentation layer off.
Look at the synaptic partners in 3D
Once you have examined the EM data and understood it, ensure the segmentation layer is turned on and navigate to the Segments tab. Right-click on the segments to load the meshes for the first synapse. In the 3D viewport, press and hold the right mouse button to rotate the view and observe the synapse from all angles. Repeat this process for the other two synapses.
Loading the meshes for the synaptic partners in the segments tab.
Annotate synapses
Now, it’s your turn to annotate synapses! Go back to the skeleton tab, select the skeleton tool and create a new tree (see video below). Move around in the EM data, with the segmentation layer turned off for better visibility. Click on a location to place a node, pointing to what you think is a synapse. Create a new tree, and repeat for other synapses. Rename then, and copy and send the link to your teacher!
Going back to the skeleton tab, turning the segmentation layer off, selecting the skeleton tool, creating a new tree. You are set to start annotating!
Finding a synapse on the EM and placing a node at its location. Renaming the node.
We hope these exercises have helped you get more familiar with structures of the brain on EM data!
Credits: Motta et al. Science, 2019, Dr. med. Holger Jastrow, EM-Atlas.