By MW Pleijzier
The previous post in this series introduced the field of connectomics and how it reflects the holistic approaches used by other subfields of biology. Connectomics details the architecture of information flow within highly connected brain networks, enabling neuroscientists to paint a picture of how these complex systems operate. The first complete connectome was of the nematode Caenorhabditis elegans, by John White and Sydney Brenner at the Laboratory of Molecular Biology in Cambridge1.
However, [#3] White and Brenner was back in the 80s, how does post-millennium Connectomics compare?
Modern connectomics has diversified its spatial scales of investigation (microscale, mesoscale and macroscale) because of the diversity of methods available to neuroscientists. The C. elegans connectome was the first complete connectome ever published at the microscopic level. We will continue examining microscale connectomics before discussing mesoscale and macroscale methods. Microscopic connectomics has advanced in three ways: data acquisition, data processing and data analysis. Whilst the modus operandi of microscopic connectomics remains largely preserved: slice the brain, image parts of individual slices, realign images (registration) and reconstruct the neurons in 3D, the methods available have evolved since White & Brenner, particularly for the generation of tomographs (i.e. “slicing”).
For microscale data acquisition, we can choose from multiple EM methodologies such as transmission electron microscopy (TEM), scanning electron microscopy (SEM), scanning transmission electron microscopy (STEM) and focused ion-beam scanning electron microscopy (FIB-SEM) which are extraordinary in their level of spatial resolution. The EM sectioning process itself has also become more of an automated process through the use of machines which reduce the risk of damage to the sections themselves; enabling unattended, multi-day imaging which increases the speed of data accumulation2. Furthermore, using machines to perform the alignment process enables more conserved use of the imaging hardware, reducing the error-correcting time during the registration process2.
As in all –omic fields, computer advancements have increased the speed of data processing – the tracing out of neurons. Brenner’s vision of computer-assisted tracing is now very much a reality in connectomics3. In 2012, the connectome of the sexually-dimorphic posterior region of the male C. elegans was published4. Whilst the data-set had been constructed before, the use of PC-software called ‘Elegance’5 made this the first full neuronal area to be detailed with unprecedented precision and speed through the use of computers. The reconstruction process for this region took two years to complete to publishing standard [Emmons, personal communication] and reconstruction is now the most arduous aspect of microscopic connectomics.
The reverse-engineering of the posterior area of the C. elegans’ nervous system used modern network analyses. The C. elegans’ posterior connectome paper, titled ‘The Connectome of a Decision-Making Network’, described the flow of information mathematically which did not feature in the White & Brenner paper. By using graph theory, Jerrell et al. (2012) were able to describe features of the neural network such as how most of the information flow within the system passed via feedforward monosynaptic pathways, that sensory neurons were heavily reciprocally connected, feedforward loops characterised the interneurons and that 55% of the input to the muscle cells came from sensory receptors largely located in the head of C. elegans (figure 2)4. Such priceless details are necessary for ‘connectionists’ to investigate the relationship between structure and function, between brain networks and behaviour.
Together, these advances enabled microscopic connectomics to shift from work on a simple system of 302/386 neurons, to much larger, more complex nervous systems; bringing us to our next question.
#4 Has Connectomics been performed on other organisms besides C. elegans?
Brenner pioneered the use of C. elegans as a model organism for biology in general and with many genetic studies being performed on the nematode as a result, C. elegans was an obvious choice for studying connectivity. Every cell within the nematode had its developmental origin documented6 and certain genes were identified to control certain behavioural phenotypes. However, the information and understanding produced by the C. elegans connectome was extremely limited and many studies remain to be performed on C. elegans7–9. This dissuaded funding for similar, large scale, long-term connectomics projects for quite some time10. Since then, and on par with White & Brenner’s choice of model organism, complete connectomes have been completed for sea squirt Ciona intestinalis11 and the worm Platynereis dumerilii12. However, sub-volume connectomic studies have been performed on many interesting neuronal regions to include the zebra finch (Taeniopygia guttata) cortex13, the zebrafish (Danio rerio) olfactory bulb14, the adult fruit-fly (Drosophila melanogaster) medulla15, the mouse (Mus musculus) neo-cortex16 and the mouse retina17, the cat cortex18 (and cortico-thalamic pathway19) and the macaque visual cortex20. With these studies came technical developments resulting in a major landmark for the field last year. This was the publication of the entire connectome of newly hatched (L1-stage) larval D. melanogaster21. (Many of these connectomic datasets can be viewed at neurodata.io.For a timeline of connectomic studies, see figure 1 in Learning from connectomics on the fly).
D. melanogaster is another key model organism used in the biological sciences. During the 1970s, Seymour Benzer (figure 3) and colleagues at the California Institute for Technology pioneered the use of Drosophila for uncovering the biological basis of behaviours, effectively establishing the field of neurogenetics (although, Drosophila was spearheaded as a biological model organism by Thomas Hunt Morgan in the early 20th century)22. By analysing alterations in behaviour caused by specific genetic mutations, they were able to determine which genes are implicated in a particular behaviour, much in the same way as Brenner discovered certain behavioural phenotypes in the nematode. Now, Drosophila is easily manipulated through genetic tools which can fluorescently label specific neuronal regions implicated in a behaviour (figure 4), or block the activity of certain neurons in vivo23,24. These tools are necessary for connectomics studies being performed on Drosophila, as the behaviour of the fruit fly is not only more complex than C. elegans, but its nervous system at the larval stage is ~33X larger and ~330X larger in the adult. These increases in neuronal complexity bring distinct behavioural repertoires both in larval and adult Drosophila. To unravel the biological basis of these behaviours and solve tractable biological problems, a diverse, flexible toolset is required. Such problems that can be proposed and tackled include: How is sensory information routed to a particular higher-order processing centre? How does synaptic organisation contribute to dendritic computation? As Pavlovian conditioning can be performed in flies, what is the circuit basis of learning and memory?
An invaluable tool is a microscale map of the adult Drosophila brain. Here at the Drosophila connectomics lab in Cambridge, our mission is to reconstruct neurons of the adult Drosophila brain involved in olfactory-memory processing. We take advantage of the largest EM data-set ever created by collaborating with the Bock lab at Janelia Research Campus in the United States. This EM dataset contains ~21 million images that occupy a total of ~106 terabytes of disk memory2! In order to make sense of all of these images, we use software called CATMAID25,26 to trace out the neurons of the Drosophila brain. However, we are not blindly picking neurons at random to gain a comprehensive map of the Drosophila brain. Instead, we have sub-teams which specialise in tracing out certain regions of interest. For example, one team focuses on the mushroom body so that we can discover the circuit basis of learning and memory. Another traces neurons within the lateral horn, a brain region which mediates innate behaviour in response to olfactory cues and other sensory stimuli.
For a ‘connectionist’ to have a sense of which part of the brain we are operating in, we take advantage of previous, mesoscopic connectomics studies. These studies fluorescently labelled specific neurons, which are then viewed under high-power light microscopes to determine where neuronal tracts project to within the brain, at micrometre or sub-micrometre resolution27. However, light microscopy is limited in spatial resolution and there is an inability to conclusively determine the presence of synapses, and therefore to determine which neurons specifically connect to one another. Mesoscopic light-level studies can be considered ‘projectomes’ rather than true connectomes, but they are valuable discoveries regardless and indispensable to microscopic connectomics work27. Mesoconnectomics has also had necessary technical developments, such as rendering entire organs transparent for visualisation under a microscope (for a full discussion see Light microscopy mapping of connections in the intact brain). Projectomes (or mesoconnectomes) have been created for many different organisms and different brain regions; the mouse cortico-striatal pathway28, the olfactory and optic tracts of zebrafish29. Additionally, projectomes have been generated of the entire Drosophila brain and can be viewed interactively at www.virtualflybrain.org and www.flycircuit.tw30. These studies utilise fluorescent proteins to label specific regions or groups of neurons within a brain and visualise where they project to both in the short and long-range31. However, this light-level data is limited in its ability to determine absolute connectivity (the individual synapses).
Returning to Sporns’ original definition (see the previous post), connectomics work is also performed on humans, at a macroscale level. The initial tomographs are generated through non-invasive techniques, such as functional magnetic resonance imaging (fMRI) and diffusor tensor imaging (DTI). These methods capture the gross structural network features in humans and other primates whilst combining them with functional data in the form of the blood oxygenation level dependent (BOLD) fMRI response. Macroscale connectomics however has two main caveats; spatial and temporal resolution. fMRI and DTI image neuronal tracts at millimetre resolution and therefore cannot image the individual synapses. The passage of time of between in vivo neuronal activity and the BOLD response is also too large for them to be direct correlates of one another (see https://www.ndcn.ox.ac.uk/divisions/fmrib/what-is-fmri/how-is-fmri-used). Large animal, macroscale connectomics capture the tidal flow of information. These tidal flows are useful for determining which brain regions are affected by connectopathies, such as Autism Spectrum Disorder, schizophrenia and depression in humans.
However, microscopic connectomics, such as that performed in the Drosophila connectome project, details the architecture supporting the individual waves of an entire brain (figure 5) and their particular route, enabling a high-resolution description of information flow through specific synapses when combined with microscopic functional methodologies (e.g. two-photon calcium imaging). Functional methods can also provide activity wiring diagrams in various nervous systems; from the crab Cancer borealis’ stomatogastric ganglion32, to the mouse amygdala33. But functional analyses should not be treated independently from structural motifs – functional maps only paint half of the picture. Once the wiring topology is known, then it can be built upon with functional data. Furthermore, neurotransmitter release at synapses which trigger action potentials in downstream neurons often ‘spillover’, inducing activity in neurons which wouldn’t be considered the intended target34, and neurons can have multiple neurotransmitter type35. By building a comprehensive network architecture image, the subtleties of synaptic transmission can be fully investigated within a holistic context. In a very broad metaphor, action potentials are the vocabulary of neuronal communication. The organisation of neurons through synapses provide neuronal grammar. Microconnectomics is therefore vital for understanding neuronal communication and its manifestation as behaviour.
Present day connectomics has an array of tools available to investigate brain connectivity. Each connectomic approach has respective, valuable, questions it can ask and answer. Each connectomics approach has respective limitations and caveats; whether technical, theoretical or financial. Understanding where these limitations come from paves the way for technical developments and an increased understanding of nervous systems from across the animal kingdom.
- White, J. G., Southgate, E., Thomson, J. N. & Brenner, S. The Structure of the Nervous System of the Nematode Caenorhabditis Elegans. 314, 1–340 (1986).
- Zheng, Z. et al. A Complete Electron Microscopy Volume Of The Brain Of Adult Drosophila melanogaster. bioRxiv (2017). doi:10.1101/140905
- 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).
- Jarrell, T. A. et al. The Connectome of a Decision-Making Neural Network. Science (80-. ). 337, 437–444 (2012).
- Xu, M. et al. Computer Assisted Assembly of Connectomes from Electron Micrographs: Application to Caenorhabditis elegans. PLoS One 8, 1–6 (2013).
- Sulston, J. E. & Horvitz, H. R. Post-embryonic cell lineages of the nematode, Caenorhabditis elegans. Dev. Biol. 56, 110–156 (1977).
- Varshney, L. R., Chen, B. L., Paniagua, E., Hall, D. H. & Chklovskii, D. B. Structural properties of the Caenorhabditis elegans neuronal network. PLoS Comput. Biol. 7, (2011).
- Towlson, E. K., Vertes, P. E., Ahnert, S. E., Schafer, W. R. & Bullmore, E. T. The Rich Club of the C. elegans Neuronal Connectome. J. Neurosci. 33, 6380–6387 (2013).
- Azulay, A., Itskovits, E. & Zaslaver, A. The C. elegans Connectome Consists of Homogenous Circuits with Defined Functional Roles. PLoS Comput. Biol. 12, 1–16 (2016).
- Schlegel, P., Costa, M. & Jefferis, G. S. X. E. Learning from connectomics on the fly. Curr. Opin. Insect Sci. 1–10 (2017). doi:10.1016/j.cois.2017.09.011
- Ryan, K., Lu, Z. & Meinertzhagen, I. A. The CNS connectome of a tadpole larva of Ciona intestinalis (L.) highlights sidedness in the brain of a chordate sibling. Elife 5, 1–34 (2016).
- Randel, N. et al. Neuronal connectome of a sensory-motor circuit for visual navigation. Elife 3, 1–22 (2014).
- Kornfeld, J. et al. EM connectomics reveals axonal target variation in a sequence-generating network. Elife 6, 1–20 (2017).
- Wanner, A. A., Genoud, C., Masudi, T., Siksou, L. & Friedrich, R. W. Dense EM-based reconstruction of the interglomerular projectome in the zebrafish olfactory bulb. Nat. Neurosci. 19, (2016).
- Takemura, S. ya et al. The comprehensive connectome of a neural substrate for ‘ON’ motion detection in Drosophila. Elife 6, 1–16 (2017).
- Kasthuri, N. et al. Saturated Reconstruction of a Volume of Neocortex. Cell 162, 648–661 (2015).
- Helmstaedter, M. et al. Connectomic reconstruction of the inner plexiform layer in the mouse retina. Nature 500, 168–174 (2013).
- Young, P. & Scannell, J. W. The connectional organization cat cerebral of neural systems in the cortex. Curr. Biol. 3, 191–200 (1993).
- Scannell, J. W., Burns, G. A. P. C., Hilgetag, C. C., O’Neil, M. A. & Young, M. P. The connectional organization of the cortico-thalamic system of the cat. Cereb. Cortex 9, 277–299 (1999).
- Felleman, D. J. & Van Essen, D. C. Distributed hierachical processing in the primate cerebral cortex. Cereb. Cortex 1, 1–47 (1991).
- Ohyama, T. et al. A multilevel multimodal circuit enhances action selection in Drosophila. Nature 520, 633–639 (2015).
- Bonini, N. M. A tribute to Seymour Benzer, 1921–2007. Genetics 180, 1265–1273 (2008).
- Duffy, J. B. GAL4 system indrosophila: A fly geneticist’s swiss army knife. Genesis 34, 1–15 (2002).
- Owald, D., Lin, S. & Waddell, S. Light, heat, action: neural control of fruit fly behaviour. Philos. Trans. R. Soc. B Biol. Sci. 370, 20140211 (2015).
- Schneider-Mizell, C. M. et al. Quantitative neuroanatomy for connectomics in Drosophila. Elife 5, 1–36 (2016).
- Saalfeld, S., Cardona, A., Hartenstein, V. & Tomančák, P. CATMAID: Collaborative annotation toolkit for massive amounts of image data. Bioinformatics 25, 1984–1986 (2009).
- Meinertzhagen, I. A. Connectome studies on Drosophila: a short perspective on a tiny brain. J. Neurogenet. 30, 62–68 (2016).
- Hintiryan, H. et al. The mouse cortico-striatal projectome. Nat. Neurosci. 19, 1100–1114 (2016).
- Robles, E. The power of projectomes: genetic mosaic labeling in the larval zebrafish brain reveals organizing principles of sensory circuits. J. Neurogenet. 31, 61–69 (2017).
- Chiang, A. S. et al. Three-dimensional reconstruction of brain-wide wiring networks in drosophila at single-cell resolution. Curr. Biol. 21, 1–11 (2011).
- Shih, C.-T. et al. Connectomics-Based Analysis of Information Flow in the Drosophila Brain. Curr. Biol. 25, 1249–1258 (2015).
- Marder, E. & Bucher, D. Understanding Circuit Dynamics Using the Stomatogastric Nervous System of Lobsters and Crabs. Annu. Rev. Physiol. 69, 291–316 (2007).
- Hou, W.-H. et al. Wiring Specificity and Synaptic Diversity in the Mouse Lateral Central Amygdala. J. Neurosci. 36, 4549–4563 (2016).
- Morgan, J. L. & Lichtman, J. W. Why not Connectomics? 10, 494–500 (2014).
- Croset, V., Treiber, C. D. & Waddell, S. Cellular diversity in the Drosophila midbrain revealed by single-cell transcriptomics. bioRxiv (2017). doi:https://doi.org/10.1101/237818
- Harris, W. A. Seymour Benzer 1921–2007 The Man Who Took Us from Genes to Behaviour. PLoS Biol. 6, e41 (2008).
- Jenett, A. et al. A GAL4-Driver Line Resource for Drosophila Neurobiology. Cell Rep. 2, 991–1001 (2012).
- Bouzaiane, E., Trannoy, S., Scheunemann, L., Plaçais, P. Y. & Preat, T. Two Independent Mushroom Body Output Circuits Retrieve the Six Discrete Components of Drosophila Aversive Memory. Cell Rep. 11, 1280–1292 (2015).