Network Statistics of the Whole-Brain Connectome of Drosophila

Albert Lin, Runzhe Yang, Sven Dorkenwald, Arie Matsliah, Amy Sterling, Philipp Schlegel, Szi-chieh Yu, Claire McKellar, Marta Costa, Katharina Eichler, Alexander Shakeel Bates, Nils Eckstein, Jan Funke, Gregory S.x.e. Jefferis, Mala MurthybioRxiv 2023 · DOI

TL;DR

Researchers have analyzed the complete whole-brain connectome of the fruit fly Drosophila melanogaster, revealing a highly interconnected network with a prominent "rich club" organization. This detailed map of neural connections provides a foundation for understanding how brain structure supports computation and information flow, identifying key neuron types that may act as signal integrators or broadcasters.

What Problem Does This Paper Address?

Understanding how the complex wiring of a brain gives rise to its functions is a fundamental challenge in neuroscience. While previous studies have mapped smaller neural circuits or used indirect methods like fMRI for larger brains, the complete, synapse-resolution wiring diagram of an entire brain has been elusive. This paper leverages the newly completed FlyWire connectome of Drosophila melanogaster – the largest such dataset to date – to perform a comprehensive network analysis, aiming to uncover fundamental principles of brain organization and connectivity that might be conserved across species.

Methods

The study employs a range of network analysis techniques applied to the FlyWire connectome. Key methods include:

  • Network Motif Analysis: Quantifying the prevalence and strength of small recurring patterns of connectivity (2- and 3-node motifs) and comparing them to null models (Erdős-Rényi and Configuration models) to identify over- and under-represented structures.
  • Rich Club Analysis: Identifying "rich club" neurons – those with high degrees that are preferentially connected to each other – and defining thresholds for these neurons.
  • Broadcaster/Integrator Identification: Categorizing neurons within the rich club based on the balance of their inputs and outputs to identify potential signal broadcasters (more outputs than inputs) and integrators (more inputs than outputs).
  • Spectral Analysis: Using random walk models (forward and reverse) to identify "attractor" and "repeller" neurons, which represent key nodes in information flow.
  • Neuropil-Specific Analysis: Examining connectivity patterns within and between the 78 anatomically defined brain regions (neuropils) of the fly brain.
  • Neurotransmitter Analysis: Integrating predicted neurotransmitter types with connectivity data to understand the chemical underpinnings of network motifs and neuron roles.

Datasets

The primary dataset used is the FlyWire connectome of an adult female Drosophila melanogaster. This dataset comprises:

  • 127,978 neurons
  • 2,613,129 thresholded connections (defined by at least 5 synapses)
  • Synapse-level neurotransmitter predictions.

This dataset represents the first complete whole-brain connectome at synapse resolution for an adult fly.

Key Insights and Ideas

  • Rich Club Dominance: The fly brain exhibits a remarkably large "rich club" – approximately 30% of neurons are highly connected and preferentially interconnected, suggesting a robust and efficient network architecture.
  • Broadcasters and Integrators: Within the rich club, distinct populations of neurons were identified as potential broadcasters (high out-degree) and integrators (high in-degree), hinting at specialized roles in information processing.
  • Neuropil Specialization: Connectivity patterns, including reciprocity and motif prevalence, vary significantly across different brain regions (neuropils), indicating functional specialization at the regional level.
  • Over-representation of Recurrent Motifs: Unlike simple feedforward structures, recurrent motifs (like 3-unicycles) are over-represented, suggesting complex feedback and local processing loops are crucial for brain function.
  • Connectivity Bridges Bottlenecks: The rich club organization appears to compensate for anatomical bottlenecks, such as the connection between hemispheres or between the central brain and optic lobes, by maintaining short path lengths across the entire brain.

Key Results

  • The fly brain connectome is highly interconnected, forming a giant strongly connected component (SCC) containing 93.3% of neurons and a giant weakly connected component (WCC) containing 98.8% of neurons.
  • The rich club regime, defined by a total degree threshold of 37, encompasses 40,218 neurons (approx. 30% of the connectome), with a connection probability 5.4 times higher than the overall brain.
  • The study identified 676 broadcasters and 638 integrators within the rich club.
  • Reciprocal connections are significantly over-represented compared to random null models, with a connection reciprocity probability of 0.138.
  • The clustering coefficient is 0.0463, also higher than expected by chance, indicating a non-random, highly connected structure.
  • Three-node motifs, particularly recurrent ones, are over-represented, while feedforward motifs are under-represented.

Limitations and Future Work

  • Neurotransmitter Prediction Accuracy: While high (94% per neuron), neurotransmitter predictions are not perfect and may contain errors, potentially affecting interpretations. Manual correction was applied for Kenyon cells.
  • Synapse Thresholding: The choice of a 5-synapse threshold for defining connections is conservative and may underestimate the true connectivity. However, key statistics were shown to be robust to this threshold.
  • Co-transmission and Extrasynaptic Communication: The study assumes Dale's Law (single neurotransmitter per neuron) and does not account for co-transmission or non-synaptic communication, which could influence information flow.
  • Functional Validation: The identified neuron populations (broadcasters, integrators, attractors, repellers, NSRNs) are based on structural connectivity. Future work will involve experimental validation of their functional roles.
  • Larger Network Motifs: The analysis focused on 2- and 3-node motifs; exploring larger, more complex motifs remains an avenue for future research.
  • Inter-individual Variability: While the study provides a baseline, understanding the extent of variability in connectomes between individual flies requires more datasets.

The authors emphasize that the FlyWire dataset and the derived neuron lists are shared resources that will serve as a foundation for future experimental and theoretical work, enabling detailed circuit profiling and hypothesis generation.

Summary updated

Abstract

Abstract Brains comprise complex networks of neurons and connections. Network analysis applied to the wiring diagrams of brains can offer insights into how brains support computations and regulate information flow. The completion of the first whole-brain connectome of an adult Drosophila , the largest connectome to date, containing 130,000 neurons and millions of connections, offers an unprecedented opportunity to analyze its network properties and topological features. To gain insights into local connectivity, we computed the prevalence of two- and three-node network motifs, examined their strengths and neurotransmitter compositions, and compared these topological metrics with wiring diagrams of other animals. We discovered that the network of the fly brain displays rich club organization, with a large population (30% percent of the connectome) of highly connected neurons. We identified subsets of rich club neurons that may serve as integrators or broadcasters of signals. Finally, we examined subnetworks based on 78 anatomically defined brain regions or neuropils. These data products are shared within the FlyWire Codex and will serve as a foundation for models and experiments exploring the relationship between neural activity and anatomical structure.
BibTeX
@misc{lin2023network,
  author = {Albert Lin and Runzhe Yang and Sven Dorkenwald and Arie Matsliah and Amy Sterling and Philipp Schlegel and Szi-chieh Yu and Claire McKellar and Marta Costa and Katharina Eichler and Alexander Shakeel Bates and Nils Eckstein and Jan Funke and Gregory S.X.E. Jefferis and Mala Murthy},
  title = {{Network Statistics of the Whole-Brain Connectome of <i>Drosophila</i>}},
  year = {2023},
  doi = {10.1101/2023.07.29.551086},
  howpublished = {bioRxiv},
}

Loading related content...

Figure