THURSDAY, 25 JULY 2024
Our brains contain billions of neurons. However, much like Lego, what truly underlies the power of our nervous system is the ability to use the same set of building blocks to create seemingly endless possibilities. Welcome to connectomics, the branch of neuroscience with the ambitious goal of finely mapping every connection, or synapse, of an entire nervous system (comprising the brain and spinal cord), of a specific region. The rationale is simple – in order to truly understand how brains produce thoughts, make decisions, and generate behaviour, we need an accurate map (connectome), outlining the highways along which information flows. Many high-profile proponents argue for the incredible value a human connectome would possess, but this remains a distant goal. With a rich history and a thriving present, will the future of connectomics be bright?THE FIRST STEPS
Connectomics launched onto the scene in the 1970s spearheaded by future Nobel Laureate Sydney Brenner’s group, based at the Laboratory of Molecular Biology (LMB) in Cambridge. They set out to identify a tractable model organism to begin to understand the links between nervous system structure and behaviour. The humble nematode C. elegans ticked all the boxes, with a nervous system of only around 300 neurons that didn’t vary from worm to worm. Nevertheless, the task stretched the technology of the time to its limits. The only tool capable of resolving synapses was electron microscopy (EM), which required the preparation of extremely thin sections of imaged samples. These had to undergo the painstaking process of stacking, alignment, and careful tracing of single neurons by hand to produce a 3D reconstruction. An intensive 15 year collaborative effort culminated in a landmark paper, published in 1985 and affectionately dubbed ‘The Mind of a Worm’. For the first time, a complete nervous system map of an organism was presented.
The insights gleaned from this groundbreaking resource have stood the test of time. The design principles identified were later shown to be conserved across vastly different scales. For example, groups of neurons called ‘rich clubs’ are particularly well connected, relaying signals between more local functional modules – much like a major transport hub like King’s Cross Station on the London Underground. Intriguingly, some circuit motifs even crop up in the architectures of today’s most sophisticated deep-learning tools.
LEARNING TO READ
What use is a map if it cannot be read? Brenner himself knew that structure would not be enough, but it would be needed to scaffold the next phase of research: linking to behaviour. Fortunately, C. elegans proved a suitable testing ground to begin connecting simple circuits to simple behaviours. The structure, location, and patterns of connectivity between subsets of neurons offered tantalising hints of their function. They studied mutant worms showing, for example, uncoordinated motion; peering closer showed that particular neurons were disrupted, likely causing their abnormalities. These theories could be verified by using a targeted laser to damage the same specific neurons in a normal (wild-type) worm to observe the same behavioural changes flagged in the mutant. For example, taking out the putative touch receptors led to worms ignoring the researchers jabbing them with an eyebrow hair, which would usually evoke an escape response. These pioneering studies led to further demonstrations in other invertebrates of how simple behaviours arise, such as the gill withdrawal reflex of the sea hare Aplysia.
Despite these successes, in many cases the connection between circuitry and behaviour proved too complex to read; the high degree of cross-connectivity flummoxed early researchers. The path had to be paved by advances in experimental techniques, including genetic constructs coding for proteins that read out neuron activity through light. In 2011, a Harvard group demonstrated an ‘optogenetic’ technique to express light-sensitive channels in defined neurons, conferring the ability to activate or shut down their activity with the flash of a light. This turned the worm into a puppet dancing to the experimenter’s tune, hijacking their motility to confirm the role of motor neurons as suggested by Brenner’s earlier mutants.
SCALING UP
Following the early C. elegans studies, the field stagnated due to a lack of technological progress, but revived in the mid-2000s at Janelia Research Campus with a focus on the fruit fly Drosophila melanogaster. Here, EM data was collected for brain regions by 2012, but the analysis and reconstruction took over a decade. In the waiting period, smaller projects explored particular circuits, such as those with putative roles in learning and memory. A major breakthrough came from teaming up with Google to develop artificial intelligence (AI)-assisted tools, delegating the laborious task of tracing to machines. This dramatically shortened the time required to interpret the data, but the error-ridden output still required manual proofreading. Very recently, Professors Marta Zlatic and Albert Cardona, and Dr Michael Winding, all at Cambridge, teamed up to publish the first ever complete synaptic connectome of an insect: larval Drosophila. With over 3000 neurons and 500,000 synapses, the resulting map highlighted features of connectivity such as extensive integration of sensory information in the brain. With a sophisticated behavioural repertoire, the mighty Drosophila experimental toolkit can now be unleashed to link these circuits functionally to different processes, such as learning and decision making.
BEYOND THE WIRES
As today’s technology showcases, not all communication follows wires. Even simple nervous systems have evolved neat tricks to bypass their structural connections. We now know that information flow is not confined to synaptic highways, with ‘extrasynaptic’ signalling offering a more flexible alternative. Neurons can converse through neuropeptides, a molecular language involving release of small proteins by neurons. They are then picked up by others in the neighbourhood that are kitted out with the right receptor - no physical coupling needed. Neuropeptides play vital roles in modulating neuron activity and regulating brain states, including sleep and arousal. A full connectome will therefore require an extrasynaptic overlay to catalogue this ‘volume transmission’. William Schaefer’s group at the LMB in Cambridge has just released a draft neuropeptidergic connectome for C. elegans, pieced together from gene expression data, anatomical information and peptide-receptor profiles. With a tenfold higher density than the synaptic connectome, they found important differences in the topology of the network, which were not unexpected given the ability to transcend the synaptic constraints. Much future research will focus on its interpretation.
THE ROAD AHEAD
The studies described may herald the dawn of a golden era for connectomics, with sights now set on the milestone of a whole mouse brain connectome (around 70 million neurons). This would be a massive step up in scale, with the raw imaging data required filling at least 500,000 laptops. A report by the Wellcome Trust synthesising the opinions of 50 worldwide experts concluded the monumental contribution such a resource would offer; the impact would cascade far beyond mere blue-sky science, with transformative consequences for healthcare, the economy and society. The timescale of delivering on this target, and how deep funders will need to reach into their pockets, remain to be determined. Collaboration, and the infrastructure to support it, will prove essential. The estimated human hours required to proofread the data currently stands at a whopping 100,000 years. A helping hand from AI-assisted technologies provides hope of reducing this burden, provided they are up to scratch in terms of accuracy and reliability. Overall, the task may stretch the connectomics community beyond any previous venture, but a truly great reward is at stake.
The road ahead for connectomics will likely be long and winding, with plenty of twists, turns and bumps along the way. Though a light shines brightly at the end of the tunnel, for those brave enough to walk in Brenner’s footsteps. Will the field deliver on his hopes? Only time will tell, but one thing is certain - neuroscientists of the modern day wait with bated breath.
Artwork by Grace Heslin.
