NeuroTrailblazers banner

Datasets & Case Studies Journal Papers

Landmark connectomics papers that defined the field — each presenting a major dataset or reconstruction. Each paper includes summaries at three expertise levels.

1. White et al. (1986) — The Structure of the Nervous System of C. elegans

Citation: White JG, Southgate E, Thomson JN, Brenner S. The structure of the nervous system of the nematode Caenorhabditis elegans. Philosophical Transactions of the Royal Society of London B. 1986;314(1165):1-340. DOI: 10.1098/rstb.1986.0056

Tags: case-studies:C-elegans case-studies:whole-brain case-studies:dense-reconstruction imaging:serial-section imaging:TEM connectomics:graph-theory methodology:ground-truth

Summaries

Beginner: This is the paper that started connectomics. Published in 1986, it describes the complete wiring diagram of the roundworm C. elegans — all 302 neurons and roughly 5,000 chemical synapses, traced by hand from thousands of electron microscope images. It took over a decade to complete. Every subsequent connectomics project traces its intellectual lineage back to this work. If you read one paper in this field, this is it.

Intermediate: White et al. produced the first (and for decades, the only) complete nervous system wiring diagram. Using serial section TEM of multiple animals, they reconstructed all neurons, identified their synaptic connections, and classified the connections as chemical synapses or gap junctions. Key contributions include: a complete parts list (302 neurons, named and classified), a complete wiring diagram (adjacency matrix), and identification of circuit motifs underlying known behaviors (locomotion, feeding, sensory processing). The methodology — serial section TEM with manual tracing — remained the dominant approach for 25 years.

Advanced: This 340-page paper remains a model of thoroughness. Methodologically, it established practices still used today: multiple animals for verification, systematic documentation of every neuron and connection, and explicit accounting of reconstruction confidence. Limitations that became apparent later include: reliance on a small number of animals (introducing individual variability), missed connections subsequently identified by Cook et al. (2019), and the absence of quantitative synapse size data. The paper’s classification of neuron types by morphology and connectivity anticipated modern computational approaches. The complete dataset, reanalyzed with modern graph theory tools, continues to yield new biological insights.

Key figures: Fig. 1-5 (nervous system overview), Appendix tables (complete connectivity matrices)

Discussion prompts:

Related content: C. elegans revisited, Connectome history

2. Kasthuri et al. (2015) — Saturated Reconstruction of a Volume of Neocortex

Citation: Kasthuri N, Hayworth KJ, Berger DR, Schalek RL, Conchello JA, Knowles-Barley S, et al. Saturated reconstruction of a volume of neocortex. Cell. 2015;163(3):633-647. DOI: 10.1016/j.cell.2015.06.054

Tags: case-studies:mouse case-studies:dense-reconstruction imaging:ATUM imaging:serial-section neuroanatomy:synapse neuroanatomy:dendrite methodology:ground-truth

Summaries

Beginner: This paper mapped every single neuron, synapse, and glial cell in a tiny cube of mouse brain tissue. “Saturated” means nothing was left out — every object was identified and traced. The cube was tiny (about the width of a hair) but incredibly complex, containing thousands of synaptic connections. This was one of the first demonstrations that dense wiring maps could be created for mammalian brain tissue.

Intermediate: Kasthuri et al. performed dense reconstruction of a 1,500 μm³ volume of mouse somatosensory cortex using ATUM-SEM. All neuronal and glial profiles were segmented, all synapses were identified, and connectivity was fully mapped. Key findings: axonal boutons synapse on diverse targets rather than concentrating on single partners; synapse size varies over two orders of magnitude; and even in this small volume, non-random connectivity patterns are evident. The dataset became an important benchmark for segmentation algorithm development.

Advanced: This paper established the concept of “saturated reconstruction” — the explicit goal of identifying every object and connection within a volume, as opposed to sparse tracing of selected neurons. The methodological contribution extends beyond the biology: the paper defines completeness criteria, reports inter-annotator agreement, and provides the dataset as a benchmark. The volume size limitation (1,500 μm³, roughly one dendritic field) means that few neurons have complete arbors within the volume, raising questions about how to interpret truncated connectivity patterns. The finding that connectivity is non-random even at this scale motivated larger reconstruction efforts.

Key figures: Fig. 1 (saturated volume), Fig. 3 (synapse size distribution), Fig. 5 (connectivity matrix), Fig. 7 (wiring diagram)

Discussion prompts:

Related content: MICrONS visual cortex, Synapse classification

3. MICrONS Consortium (2021) — Functional Connectomics Spanning Multiple Areas of Mouse Visual Cortex

Citation: MICrONS Consortium, Bae JA, Baptiste M, Bodor AL, Brittan D, Buchanan J, et al. Functional connectomics spanning multiple areas of mouse visual cortex. bioRxiv. 2021. DOI: 10.1101/2021.07.28.454025

Tags: case-studies:MICrONS case-studies:mouse case-studies:visual-cortex case-studies:dense-reconstruction neuroai:structure-function imaging:ATUM imaging:serial-section connectomics:graph-theory

Summaries

Beginner: The MICrONS project is one of the most ambitious connectomics efforts to date. Scientists first recorded the activity of tens of thousands of neurons in a living mouse as it watched visual stimuli, then removed that same piece of brain and mapped all the connections using electron microscopy. This creates a unique dataset where you can see both what each neuron does (its function) and who it connects to (its structure) — enabling direct tests of how wiring relates to function.

Intermediate: The MICrONS consortium produced a mm³ volume of mouse visual cortex with matched functional imaging (two-photon calcium imaging of ~75,000 neurons) and structural reconstruction (serial section EM with ~80,000 reconstructed neurons and ~524 million synapses). The functional-structural co-registration enables direct testing of structure-function hypotheses: are functionally similar neurons preferentially connected? The dataset spans multiple visual areas (V1, LM, AL) and all cortical layers, providing the most comprehensive functional connectomics resource for any mammalian brain region.

Advanced: MICrONS represents the convergence of multiple technical advances: large-scale two-photon imaging with single-cell resolution, petascale EM acquisition, FFN-based segmentation, and CAVE-based proofreading. The functional-structural matching required solving a challenging cross-modal registration problem (in vivo optical to ex vivo EM). Key caveats: calcium imaging provides noisy functional characterization (compared with electrophysiology), the EM volume boundaries truncate many neurons, and proofreading completeness varies across the volume. The structure-function correlations reported are modest in magnitude (~0.1-0.2), consistent with the view that connectivity reflects multiple organizing principles beyond functional similarity.

Key figures: Fig. 1 (experimental pipeline), Fig. 2 (EM volume overview), Fig. 5 (structure-function correlations)

Discussion prompts:

Related content: MICrONS visual cortex, NeuroAI bridge

4. Shapson-Coe et al. (2024) — A Petavoxel Fragment of Human Cerebral Cortex

Citation: Shapson-Coe A, Januszewski M, Berger DR, Pope A, Wu Y, Blakely T, et al. A connectomic study of a petascale fragment of human cerebral cortex. Science. 2024;384(6696):eadk4858. DOI: 10.1126/science.adk4858

Tags: case-studies:H01 case-studies:human case-studies:dense-reconstruction imaging:serial-section imaging:SEM neuroanatomy:synapse neuroanatomy:dendrite cell-types:pyramidal-cell

Summaries

Beginner: This paper presents the first large-scale wiring map from a human brain. A small piece of temporal cortex (about 1 cubic centimeter) was removed during epilepsy surgery and mapped at nanometer resolution, revealing about 50,000 neurons and 130 million synapses. The human tissue showed several features never seen in mouse or fly brains, including unusually large neurons, rare multinucleated cells, and unique wiring patterns, suggesting that human brain organization differs from model organisms in important ways.

Intermediate: Shapson-Coe et al. present the H01 dataset: a petavoxel (~1.4 PB) serial section SEM volume of resected human temporal cortex at 4 nm lateral resolution. The reconstruction reveals ~50,000 neurons with ~130 million synapses across cortical layers 1-6. Key findings include: human-specific morphological features (large L3 pyramidal neurons with extensive basal dendrites), unusual connectivity patterns (axonal whorls, multinucleated neurons), and quantitative differences from rodent cortex (higher spine density, larger synapses). The dataset comes from a surgical specimen (epilepsy patient), raising questions about typicality.

Advanced: H01 is important as the first dense human connectomics dataset, but interpretation requires careful consideration of several factors: (1) the tissue source (epilepsy surgery specimen from a 45-year-old female) means some features may be pathological rather than typical; (2) fixation was delayed compared with perfusion-fixed mouse tissue, affecting ultrastructural preservation; (3) temporal cortex may differ from primary sensory cortex in connectivity patterns. The paper’s most impactful contribution may be the catalog of human-specific features that were unknown from model organisms: the “deep axo-axonic” connections, the axonal whorls of unknown function, and the multinucleated neurons. These features challenge the assumption that mouse cortex is a sufficient model for human brain organization.

Key figures: Fig. 1 (volume overview), Fig. 2 (human-specific morphology), Fig. 4 (connectivity patterns), Fig. 6 (pathological features)

Discussion prompts:

Related content: H01 human cortex, Soma ultrastructure

5. Dorkenwald et al. (2024) — Neuronal Wiring Diagram of an Adult Brain (FlyWire)

Citation: Dorkenwald S, Matsliah A, Sterling AR, Schlegel P, Yu SC, McKellar CE, et al. Neuronal wiring diagram of an adult brain. Nature. 2024;634:124-138. DOI: 10.1038/s41586-024-07558-y

Tags: case-studies:FlyWire case-studies:Drosophila case-studies:whole-brain case-studies:dense-reconstruction proofreading:crowd-sourced-proofreading infrastructure:CAVE connectomics:graph-theory

Summaries

Beginner: In 2024, scientists published the first complete wiring diagram of an adult animal brain — the fruit fly Drosophila melanogaster, with about 139,000 neurons and 50 million synaptic connections. This was accomplished through a massive collaborative effort: hundreds of researchers worldwide used a web-based tool to check and correct the computer-generated neuron shapes. The result is a complete “parts list and wiring manual” for a brain, enabling systematic study of how brain circuits produce behavior.

Intermediate: Dorkenwald et al. present the FlyWire connectome: a complete, proofread wiring diagram of the adult Drosophila brain. Built from the FAFB EM dataset (Zheng et al., 2018), the connectome was generated through FFN segmentation followed by community-scale proofreading via CAVE. Key features: ~139,000 neurons and ~50 million synapses; collaborative proofreading by 300+ contributors; cell type annotation for all neurons; freely available data and analysis tools. Initial analyses reveal brain-wide circuit architecture, including processing hierarchies, bilateral symmetry, and cell-type-specific wiring rules.

Advanced: FlyWire is the first whole-brain connectome at single-synapse resolution for an organism with complex behavior, making it a milestone comparable to White et al. (1986) for C. elegans. Key methodological contributions: (1) the crowd-sourced proofreading model that scaled human effort beyond what any single lab could achieve; (2) the CAVE infrastructure that enabled concurrent editing by hundreds of users; (3) the quality metrics framework that quantifies proofreading completeness. The connectome’s completeness is not absolute: small-caliber neurites in the anisotropic EM volume have higher error rates, and some neuron types were proofread more thoroughly than others. Companion papers analyze cell types (Schlegel et al.), neuropeptide signaling (Dacks et al.), and specific circuits (multiple papers in the same Nature issue).

Key figures: Fig. 1 (whole-brain overview), Fig. 2 (reconstruction statistics), Fig. 3 (cell type diversity), Fig. 4 (circuit examples)

Discussion prompts:

Related content: FlyWire whole-brain, Connectome history

6. Zheng et al. (2018) — A Complete Electron Microscopy Volume of the Brain of Adult Drosophila

Citation: Zheng Z, Lauritzen JS, Perlman E, Robinson CG, Nichols M, Milkie D, et al. A complete electron microscopy volume of the brain of adult Drosophila melanogaster. Cell. 2018;174(3):730-743.e22. DOI: 10.1016/j.cell.2018.06.019

Tags: case-studies:Drosophila case-studies:FAFB case-studies:whole-brain imaging:ATUM imaging:serial-section infrastructure:pipeline infrastructure:alignment

Summaries

Beginner: Before you can map a brain’s wiring, you need to image the entire brain at high enough resolution to see individual connections. This paper describes how scientists imaged a complete adult fruit fly brain — producing about 21 million images that were stitched together into one continuous 3D picture. The imaging alone took years and produced ~50 terabytes of data. This dataset (called FAFB) became the foundation for the FlyWire wiring diagram project.

Intermediate: Zheng et al. produced the FAFB (Full Adult Fly Brain) dataset: the complete Drosophila brain imaged at ~4x4x40 nm resolution using automated tape-collecting ultramicrotomy (ATUM) and multi-microscope SEM. Key technical achievements: ~7,000 serial sections with only 36 unusable, multi-microscope parallelization for throughput, and computational alignment of ~21 million tile images. The paper demonstrates tracing of individual neurons across the full brain, validating that the image quality is sufficient for connectomics.

Advanced: FAFB is the imaging foundation of the FlyWire connectome and illustrates the engineering scale of modern connectomics. The ATUM workflow preserves sections (unlike SBEM), enabling re-imaging and troubleshooting. The 10x z-anisotropy (4 nm lateral, 40 nm axial) creates directional biases in segmentation accuracy — thin processes running in-plane are better resolved than those running axially. The paper’s reporting of section loss rate, stitching quality, and alignment accuracy provides benchmarks for future projects. The multi-microscope parallelization strategy (partitioning the tape across instruments) was essential for practical timeline and was subsequently adopted by larger projects.

Key figures: Fig. 1 (whole-brain overview), Fig. 2 (ATUM workflow), Fig. 3 (section quality), Fig. 4 (neuron tracing examples)

Discussion prompts:

Related content: FlyWire whole-brain, EM principles, Acquisition QA

7. Cook et al. (2019) — Whole-Animal Connectomes of Both C. elegans Sexes

Citation: Cook SJ, Jarrell TA, Brittin CA, Wang Y, Bloniarz AE, Yakovlev MA, et al. Whole-animal connectomes of both Caenorhabditis elegans sexes. Nature. 2019;571(7763):63-71. DOI: 10.1038/s41586-019-1352-7

Tags: case-studies:C-elegans case-studies:whole-brain case-studies:dense-reconstruction connectomics:connectome-comparison connectomics:graph-theory methodology:reproducibility

Summaries

Beginner: The original C. elegans wiring diagram from 1986 mapped the hermaphrodite (the more common sex). This paper completes the picture by mapping the male nervous system too, and also corrects errors in the original reconstruction. The male has about 385 neurons (83 more than the hermaphrodite), with the extra neurons forming circuits for mating behavior. Comparing the two sexes shows which wiring is shared and which differs.

Intermediate: Cook et al. present updated connectomes for both C. elegans sexes, substantially revising the original White et al. (1986) hermaphrodite connectome and providing the first complete male connectome. They identified ~1,500 previously missed synapses in the hermaphrodite and mapped the complete male nervous system including 91 male-specific neurons. The comparative analysis reveals: shared circuits are quantitatively conserved between sexes, while male-specific neurons form modular additions that modify shared circuit function for reproductive behavior.

Advanced: This paper addresses a fundamental limitation of single-animal connectomics by incorporating data from multiple individuals and both sexes. The finding that ~1,500 synapses were missed in the original White et al. reconstruction (a ~25% increase in edge count) has sobering implications for the completeness claims of any connectome. The modular organization of sex-specific circuits — added as appendages to a conserved shared backbone rather than deeply integrated — suggests an evolutionary mechanism for circuit diversification. The quantitative conservation of shared connectivity between sexes (high correlation of connection strengths) establishes a baseline expectation for individual-to-individual variability in this species.

Key figures: Fig. 1 (sex comparison overview), Fig. 2 (updated hermaphrodite connectivity), Fig. 3 (male-specific circuits), Fig. 5 (shared vs. dimorphic connectivity)

Discussion prompts:

Related content: C. elegans revisited, Connectome history

8. Takemura et al. (2013) — A Visual Motion Detection Circuit Suggested by Drosophila Connectomics

Citation: Takemura S, Bharioke A, Lu Z, Nern A, Vitaladevuni S, Rivlin PK, et al. A visual motion detection circuit suggested by Drosophila connectomics. Nature. 2013;500(7461):175-181. DOI: 10.1038/nature12450

Tags: case-studies:Drosophila case-studies:retina connectomics:graph-theory neuroanatomy:synapse neuroai:structure-function

Summaries

Beginner: This paper is a classic early connectomics study that mapped the wiring of neurons in the fly’s visual system to figure out how the brain detects motion. Scientists used electron microscopy to trace the connections between specific cell types in the fly’s medulla (a visual processing region). By identifying which neurons connect to which, they proposed a circuit diagram that explains how the fly computes the direction of moving objects — showing that the physical wiring of the brain can predict the computations it performs.

Intermediate: Takemura et al. used serial section TEM to densely reconstruct columns of the Drosophila medulla, identifying the cell types and synaptic connections involved in motion detection. The reconstructed connectivity revealed a circuit architecture consistent with variants of the Hassenstein-Reichardt correlator model for elementary motion detection. Key findings include: identification of specific columnar cell types (Mi1, Tm3, Mi4, Mi9, T4, T5) and their directionally asymmetric wiring; quantitative synapse counts revealing connection strength differences between pathway branches; and a wiring diagram that predicts the computational algorithm (correlation of luminance signals across space and time). This was among the first demonstrations that connectomics data could directly constrain computational models.

Advanced: This paper exemplifies the power of targeted connectomics — reconstructing a specific circuit to answer a specific computational question, rather than pursuing whole-brain completeness. The motion detection circuit had been studied electrophysiologically and computationally for decades, but the cellular implementation remained disputed. The connectomic data resolved this by identifying the actual cell types and connections implementing the computation. Methodological strengths include reconstruction of multiple columns for statistical power and explicit mapping from anatomy to algorithm. Limitations: the reconstruction covered a small number of columns, synapse identification relied on morphological criteria without functional validation, and the mapping from connectivity to computation required assumptions about synaptic sign (excitatory vs. inhibitory) that were later tested genetically. Subsequent functional studies (e.g., Maisak et al., 2013) confirmed many of the circuit predictions made from the anatomy.

Key figures: Fig. 1 (medulla column reconstruction), Fig. 2 (cell type identification and connectivity), Fig. 3 (motion detection circuit diagram), Fig. 4 (computational model comparison)

Discussion prompts:

Related content: FlyWire whole-brain, NeuroAI bridge

9. Bock et al. (2011) — Network Anatomy and In Vivo Physiology of Visual Cortical Neurons

Citation: Bock DD, Lee WCA, Kerlin AM, Andermann ML, Hood G, Wetzel AW, et al. Network anatomy and in vivo physiology of visual cortical neurons. Nature. 2011;471(7337):177-182. DOI: 10.1038/nature09802

Tags: case-studies:mouse case-studies:visual-cortex imaging:serial-section imaging:ATUM neuroai:structure-function methodology:correlative-imaging

Summaries

Beginner: This paper was one of the first to combine two powerful techniques: watching neurons fire in a living mouse brain, and then mapping their connections with an electron microscope. Scientists first showed mice visual patterns and recorded which neurons responded to which orientations (using two-photon calcium imaging). Then they sliced and imaged the same piece of brain tissue to trace the wires between those neurons. They discovered that inhibitory neurons receive input from neurons with many different orientation preferences — meaning inhibition is broadly tuned, not selective.

Intermediate: Bock et al. performed correlative functional-structural imaging in mouse primary visual cortex. In vivo two-photon calcium imaging characterized orientation selectivity of neurons in a local patch of V1, after which the same tissue was processed for serial section EM and reconstructed. By matching functionally characterized neurons to their EM profiles, they traced synaptic inputs onto specific inhibitory interneurons. Key finding: inhibitory neurons receive synaptic input from excitatory neurons spanning the full range of orientation preferences, rather than from a functionally homogeneous subset. This broad connectivity provides a structural basis for non-selective inhibition in cortical circuits.

Advanced: This paper pioneered the correlative functional-structural approach that later scaled up dramatically in the MICrONS project. The technical achievement of registering in vivo two-photon data to ex vivo EM was non-trivial, requiring fiducial landmarks and careful tissue processing. The biological conclusion — that inhibitory neurons sample broadly from the local excitatory population — addressed a longstanding debate about the role of inhibition in sharpening tuning curves. Limitations include: the small number of reconstructed inhibitory neurons (constraining statistical power), the inability to identify all presynaptic partners (many axons could not be traced to their somata), and the use of calcium imaging rather than electrophysiology for functional characterization (limiting temporal resolution and sensitivity). Despite these limitations, the paper established a paradigm — function first, then structure — that became central to modern connectomics.

Key figures: Fig. 1 (correlative imaging pipeline), Fig. 2 (functional characterization of neurons), Fig. 3 (EM reconstruction of inhibitory neurons), Fig. 4 (input selectivity analysis)

Discussion prompts:

Related content: MICrONS visual cortex, NeuroAI bridge

10. Witvliet et al. (2021) — Connectomes Across Development Reveal Principles of Brain Maturation

Citation: Witvliet D, Mulcahy B, Mitchell JK, Meiber Y, Chisholm R, Wang Y, et al. Connectomes across development reveal principles of brain maturation. Nature. 2021;596(7871):257-261. DOI: 10.1038/s41586-021-03778-8

Tags: case-studies:C-elegans case-studies:dense-reconstruction connectomics:connectome-comparison neuroanatomy:synapse methodology:reproducibility

Summaries

Beginner: This paper asked a simple but profound question: how does a brain’s wiring change as an animal grows up? Scientists mapped the complete wiring diagram of the roundworm C. elegans at eight different time points during development, from a newly hatched larva to an adult. They found that which neurons connect to which stays mostly the same throughout life, but the strength of those connections (how many synapses link two neurons) changes dramatically. This is the first time anyone has tracked a complete brain’s wiring across an entire developmental timeline.

Intermediate: Witvliet et al. reconstructed eight complete C. elegans connectomes spanning development from the first larval stage (L1) to adulthood. By comparing these connectomes, they identified two key principles of brain maturation: (1) connection identity is largely stable — most neuron-to-neuron connections present in the adult are already established in the L1 larva; (2) connection strength is highly dynamic — the number of synapses per connection changes substantially during development, with some connections strengthening and others weakening. Additionally, they identified a small number of connections that are added or pruned during development, suggesting both activity-dependent and genetically programmed wiring refinement.

Advanced: This is the first developmental connectomics study with multiple complete time points, enabling statistical analysis of wiring changes rather than pairwise comparison. The finding that connection identity is stable while connection strength is dynamic has important implications: it suggests that the “connectome” as a binary graph is largely genetically specified, while synaptic weight tuning may reflect experience-dependent plasticity or developmental programs. Methodological considerations include: inter-individual variability confounds developmental trends (each time point is a different animal), the definition of “same connection” across developmental stages requires careful cell identity tracking, and synapse count is a proxy for connection strength that may not capture other forms of plasticity (receptor composition, release probability). The dataset provides a unique resource for testing models of neural development and establishes baseline expectations for how much connectome variability is developmental versus individual.

Key figures: Fig. 1 (developmental series overview), Fig. 2 (connection stability analysis), Fig. 3 (synaptic strength changes), Fig. 4 (added and pruned connections)

Discussion prompts:

Related content: C. elegans revisited, Connectome history