I love maps. There’s something about a really detailed, meticulously made map that is wonderful to look at.
To take an off topic example (well, the blog is called Scientific Wanderlust after all…) – William Smith’s geological map of the UK is one of my favourite maps. Not only is it incredibly detailed, but it’s also astonishingly beautiful.
Maps of this kind are made from large amounts of carefully collected data and are valuable resources for the scientific community.
Now to get back to the topic in hand, the field of connectomics aims to make detailed maps of the connections between different parts of a brain. These ‘connectomes’ are again built upon massive datasets and take years to collect. They are a valuable resource for the Neuroscience community, allowing us to see the major roads and tiny winding paths that allow signals to move through our brains.
As an added bonus, many of these connectomes are also really beautiful maps:
[Nature Video have show some great images of a partial mouse brain reconstruction – here]
But how do we make them? What can we get out of them? And why should we care?
A matter of scale
Like conventional maps, brain maps can be made at various scales and resolutions. The appropriate resolution depends on what kind of data you want to collect, and also what sort of brain you are trying to map.
For example, current efforts to make a human connectome use MRI (Magnetic Resonance Imaging) to look at the broad structure of the brain. This allows them to visualise large bundles of nerve cells (neurons) and examine which regions of the brain are active at the same time. This allows them to infer a broad picture of how the brain is connected.
Going to higher resolution techniques, a number of studies have used light microscopy to study how much finer bundles of neurons interact. For example, the Allen Institute for Brain Science created a map of the mouse brain by using viruses to transfer GFP (Green Fluorescent Protein) into neurons in different parts of the brain. With fluorescence microscopy, they could then take images of these glowing cells to see where they extend in the brain. [Available online as ‘The Allen Mouse Brain Connectivity Atlas’]
Now to the highest resolution technique – Electron Microscopy. At this resolution individual neurons can be visualised. It is also the only technique that allows us to systematically map all the sites where neurons connect to each other (known as synapses). Hence, it allows us to build the most intricate maps of brain connectivity. Studies in a variety of organisms have used this technique – including on roundworms, fruit flies and mice!
You may wonder why we don’t use Electron Microscopy for every study, since it has the highest resolution? Well, each technique for making a connectome has its own advantages and disadvantages. For Electron Microscopy, the advantage of high resolution comes at the cost of speed. This kind of data takes years to collect and many more years to interpret, and so is only feasible for very small brains.
For example, I’m currently working in the Drosophila Connectomics Group to aid in making a microscale connectome of the fruit fly brain. Once finished, this connectome would be the largest collected to date at this resolution!
Making a Map
Most microscale connectomes are made in a similar way. First, the brain is sliced into thousands of very thin slices, which are then imaged at high resolution with an electron microscope. Each of these individual images are then stitched back together in a computer to make a complete 3D volume of the brain. This data collection step can take a number of years, but it is easy when compared to the next step – reconstruction!
Once all the images have been collected, we then need to follow and label each individual neuron through this dataset. It’s sort of like going from a satellite image of the Earth, to a Google Maps like representation where each road, railway and house is annotated on top. Only once this is complete, will we have a usable map of what every neuron in the brain looks like and also how they all connect to each other.
This annotation step takes much longer than the original data collection. At present, some of this annotation can be done automatically with computers, but due to the very fine and winding nature of neuron branches, this is still quite inaccurate. Hence, much of this work must be done manually by teams of people who analyse these images and mark where the different neurons are.
The C. elegans Connectome
To take an example, C. Elegans (a roundworm) was the first organism to have its entire connectome completed at this level of resolution. This tiny (about 1mm long) transparent worm has been used in a wide range of research.
In 1969, John White began his work on reconstructing the locations and connections between the 302 neurons of the worm. Much of this work was done by hand by John White and Eileen Southgate, working on 12×16 inch prints of electron microscopy images of sections collected by Nichol Thomson. Painstakingly, they went through every printed image marking each of the 302 neurons. Finally, in 1986, the complete map was published under the running head of ‘The mind of a worm’. It was the first synaptic resolution map of an entire nervous system.
There’s a great quote from John White in his personal account of the project – “I have often wondered how dry stone walls are built. Once, over several days, I observed this process and came to realise that the key factor was time. Fortunately the MRC allowed us time; piece-by-piece, our stone wall was assembled over the course of some 13 years”.
Technology has improved since these times, and computers make this process a lot faster. But still, it is painstaking, time consuming work – to make a connectome of this kind you need the patience to build it up slowly, one brick at a time.
This all sounds like a lot of effort to go through – why bother? Since electron microscopy is the only technique that allows us to map connections between individual neurons, it is the only way to build a single cell resolution map of how a brain is wired. Thinking of the brain as a complex circuit, these wiring diagrams can help us to understand how information is processed by the brain, and how this results in different complex behaviours.
Starting with small brains like the worm or the fruit fly allows us to use this technique to study how an entire brain is wired. As technology and the speed of reconstruction improves, we can gradually work up to mapping ever more complex brains.
Still, even at this highest resolution, a connectome cannot tell use everything about how a brain actually works. Thinking of the brain as a circuit again, we can think of the connectome as the wiring diagram showing how different components of the circuit are wired together. Yet, without an understanding of what each component does, and the dynamics of the circuit over time, we still cannot fully understand it. Experimental data can tell us about the properties of the neurons that appear in these connectomes – e.g. does the connection activate or inhibit? How strongly? Which chemicals are released at each of the connection sites? What are the behavioural consequences of changing or removing a particular connection?
These ‘connectomes’ are a valuable resource for the Neuroscience community, but they only become really powerful when they are combined with other experimental work. As with most areas in science, we can only progress when we look at lots of different kinds of data together.
Connectomes, at all their various scales and resolutions, provide valuable insight into how a brain is wired together. As technology improves we can create these brain maps at ever higher resolutions, and at much faster speeds! These maps provide a resource that can inform experimental work – allowing us to collaborate to move closer to a full understanding of how a brain works.
Thanks to the rest of the Drosophila Connectomics Group – many of the diagrams used here were created for our group’s presentation at the Cambridge BRAINFest
References / suggestions for further reading
Open Connectome Project – https://www.openconnectomeproject.org/ – free online access to a wide range of connectome datasets. e.g. you can scroll around in the C. elegans nervous system by clicking on ‘Images’ on this page
Examples of Connectome Projects:
Human Connectome Project – https://www.neuroscienceblueprint.nih.gov/connectome/
The Allen Mouse Brain Connectivity Atlas – Oh, S. W. et al. A mesoscale connectome of the mouse brain. Nature 508, 207–214 (2014). doi:10.1038/nature13186
White J.G. Getting into the mind of a worm—a personal view (June 25, 2013), WormBook, ed. The C. elegans Research Community, WormBook, doi/10.1895/wormbook.1.158.1, http://www.wormbook.org – a personal account from John White about the creation of the C. elegans connectome
Emmons, S. W. The beginning of connectomics: a commentary on White et al. (1986) ‘The structure of the nervous system of the nematode Caenorhabditis elegans’. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 370, 20140309- (2015). DOI: 10.1098/rstb.2014.0309 – a great article discussing how the C. elegans connectome was created and its wider context
Mouse retina – Helmstaedter, M. et al. Connectomic reconstruction of the inner plexiform layer in the mouse retina. Nature 500, 168–174 (2013). doi:10.1038/nature12346
Fragment of mouse brain – Kasthuri, N. et al. Saturated Reconstruction of a Volume of Neocortex. Cell 162, 648–661 (2015). DOI: http://dx.doi.org/10.1016/j.cell.2015.06.054
General Connectomics reviews:
Dance, A. Connectomes make the map. Nature 526, 147–149 (2015) doi:10.1038/526147a – an excellent, accessible overview of the connectomics field
Swanson, L. W. & Lichtman, J. W. From Cajal to Connectome and Beyond. Annu. Rev. Neurosci. 39, 197–216 (2016). https://doi.org/10.1146/annurev-neuro-071714-033954
Helmstaedter, M. Cellular-resolution connectomics: challenges of dense neural circuit reconstruction. Nat. Methods 10, 501–7 (2013) doi:10.1038/nmeth.2476
Lichtman, J. W., Pfister, H. & Shavit, N. The big data challenges of connectomics. Nat. Neurosci. 17, 1448–54 (2014) doi:10.1038/nn.3837