Virtual Observatory of the Cortex

A BRAIN initiative funded U24 project to bring the IARPA MICrONS dataset to the scientific research community.

The MICrONS dataset is public and open access, ready for analysis. But manual edits to the segmentation (“proofreading”) continue to improve data quality.

The automatic segmentation from EM imagery to 3D reconstruction was effective, and the only feasible way to process data at this scale (The MICrONS Consortium et al. 2025). However, due to imaging defects and the nature of thin, branching axons, the automated methods make mistakes that have large impacts on the biological accuracy of the reconstructions.

This is our ongoing effort to provide manual correction of segmentation and additional biological annotation.

The VORTEX program provides expert proofreading as a service to and directed by the scientific community. We are funded to facilitate your scientific questions. Read on to find how!

How it works

Is there an observation you could make from this dataset that would help you answer a scientific question, but you aren’t sure how to make it? Or maybe you know how to make it, but don’t have the time or resources to repair or annotate enough examples to answer your question.

The Virtual Observatory of the Cortex is a grant funded mechanism designed to assist you in precisely this situation. We provide the labor and experience to make corrections and annotations in the data. The results of these efforts will be integrated with an evolving version of the dataset, and shared with you and the larger scientific community.

We think of it as a metaphor for a large scale astronomy survey: we are providing the infrastructure and focusing mechanisms, the community is steering the observations to the spots of maximum interest, and everyone can see what comes out!

 
  • Members of the scientific community request and justify proofreading effort in the dataset. For example:

    • proofreading axons and dendrites for a specific group of neurons

    • annotating the locations of dense core vesicles in these particular cells

    • annotating locations where microglia have engulfed synapses

  • A Scientific Steering Group will evaluate all the plans and prioritize the manual work based upon the resources available to the virtual observatory.

    The steering group meets 2-3 times annually to review proposals.

  • Plans will then be carried out by staff of the virtual observatory, with no costs to the requestor.

    Plans will be completed within the assigned project period of 6 months.

  • The data produced from the efforts of the manual proofreaders is then shared with the entire scientific community for everyone to benefit from.

    Data releases will occur at least quarterly.

If you are uncertain what proofreading you want, or even if you need it, book a time with our scientists to discuss the best options for your scientific question:

A scientist at the Allen Institute will work with you on developing a plan that helps address the scientific question you are interested in, including any manual proofreading and annotation work that needs to be done.

Tip: if you have a general question about the accessing the data, or bugs to report, contact the MICrONS team at info@microns-explorer.org or review the MICrONS Tutorial.


Scientific Request Form

We are accepting scientific proposals for our Proofreading and Annotation program: Virtual Observatory of the Cortex (VORTEX).

This is for the first project period of our grant’s Year 3, and will consider all new proposals, including from previous requesters.

Application Deadline: August 1, 2025

Scientific Steering Group: August 2025

Project Period: September 2025 - March 2026

Upcoming Data Releases: July 2025, October 2025, January 2026, April 2026

Peruse our completed projects below and on the tutorial page: Vortex Outcomes.

 

Data Releases

Data Sharing Policy

Data created by the Vortex project, such a proofreading and annotations, will be integrated into broader publicly available dataset for everyone to use. Proofreading updates the segmentation, which will be released along with any other changes that are part of the centralized dataset, no matter their source. Annotations will be made on the dataset and released as tables via the CAVEclient, and periodically as static file dumps available via google cloud. See the specific dataset pages for details on data access.

  • Our quarterly release includes the following new tables:

    • vortex_thalamic_proofreading_status : selected proofreading of thalamic axons, as part of the VORTEX effort

    • multi_input_spine_predictions_ssa : a new prediction of which synapses share a spine head, based on analysis from synapse_target_predictions_ssa

    The following tables have updated status or entries, from VORTEX proofreading and annotations

    • proofreading_status_and_strategy (newly proofread cells and upgraded cell status)

    • vortex_astrocyte_proofreading_status (newly proofread cells and upgraded cell status)

    See Release Manifests v1412 for more details

  • Our quarterly release includes the following new tables:

    • synapse_target_predictions_ssa : a new synapse-target prediction classifier run on 200 million synapses

    • baylor_gnn_cell_type_fine_model_v2 : a new cell type classifier with neuron excitatory/inhibitory subclasses

    • coregistration_auto_phase3_fwd_apl_vess_combined_v2 : an update to the automatch table generated by combining coregistration_auto_phase3_fwd_v2 and apl_functional_coreg_vess_fwd

    • coregistration_auto_phase3_fwd_v2 : an update to the automatic coregistration model

    • cg_cell_type_calls, gamlin_2023_mcs, gamlin_2023_mcs_met_types : a table collection associated with a forthcoming publication from Gamlin et al.

    The following tables have updated status or entries, from VORTEX proofreading and annotations

    • proofreading_status_and_strategy (newly proofread cells and upgraded cell status)

    • vortex_compartment_targets (new manual annotations)

    • vortex_manual_myelination_v0 (new manual annotations)

    See Release Manifests v1300 for more details

  • Our quarterly data release includes:

    • Axon extension on 273 wholly new pyramidal cells, and upgrades to 400 previously proofread but truncated cells

    • Manual labeling of 20,000 synapse targets.

    See: v1181 Release Manifest for more

  • We have concluded our first Project Period, with:

    • Axon extension on 109 functionally co-registered pyramidal cells (6 previously proofread with incomplete extension, 103 wholly new).

    • Manual labeling of 73,000 synapse targets from 20 basket cells (putative parvalbumin cells).

    • Manual annotation of myelination and nodes of Ranvier on 53 mm of axon.

    • Selection of 12 astrocytes for segmentation quality and completeness, and ‘cleaning’ to remove incorrectly merged neurites from 3 astrocytes.

    See: v1078 Release Manifest for more

  • Highlights include: derived visual tuning properties of the functional recordings, associated with the coregistered cells. And updated flat segmentation.

    See Release Manifests for v943

  • Release of public datastack v795. Adds annotation tables containing the updated cell typing and models based on multiple methods, and further proofreading of various cell types including L5ET and L5IT excitatory cells.

    Public annotation tables can be viewed with the Dash App and the current segmentation explored with neuroglancer.

    Follow instructions in Visualization for data access

  • This released focused on releasing and updated version of the segmentation with more proofreading, including 1,051 cells with at least cleaned axons. It also marked the first release where almost 100% of the neurons have been made single soma. It also contained more co-registration, including a manual table with 13,925 entries, and a larger automated registration with >20,000 entries that had high confidence. Finally, the release contained updated cell type calls and models based on multiple methods.

    This release was accompanied by an update to the preprint describing the dataset (https://www.biorxiv.org/content/10.1101/2021.07.28.454025v3)

  • This version has an updated segmentation, which included more proofread cells including cells with at least clean 410 axons. It also contained an update to functional coregistration which includes now 9518 coregistered ROIs. Finally, it now includes a broader array of cell type calls in the dataset, including both manual and automated tables.

  • This was the initial release of the dataset, that included all the image data, the synapse detection, nucleus detection, and the functional dataset. It included 249 cells with at least clean dendrites, and 200 functional coregistered cells. It had some basic cell type calls including a model for what was likely neurons in the dataset, and excitatory vs inhibitory calls, both manual and automated.

 

Current Efforts

Topic Task Status
Proximity metrics for pyramidal cell dendrites Analysis support Ongoing
Subcellular distribution of large secretory vesicles Annotation of dense core vesicles Ongoing
Neuromodulation at the blood-brain barrier Axon extenstion of putative neuromodulatory axons Ongoing
Neuron structure modeling, using skeletonized neurons Axon extension: pyramidal cells; Axon extension: inhibitory cells Ongoing
Microglia interaction at synapses Annotation of microglia lysosomes Ongoing
Uncovering local and long-range circuit organization Axon extension: pyramidal cells; Axon extension: inhibitory cells Ongoing
Connectomics of response heterogeneity in mouse visual cortex Axon extension of functionally recorded cells Ongoing
Design principles underlying astrocyte morphology Proofread 'cleaning' of astrocyte Ongoing
Connection between SST subtypes and pyramidal neurons Axon extension: inhibitory cells Ongoing
Dopaminergic projections to cortex Axon extenstion of putative neuromodulatory axons Ongoing



Previous Efforts

Topic Task Status
How does the orientation tuning of visual cortical neurons arise from the functional organization of inputs onto their dendritic tree? Axon extension of functionally recorded cells Complete: v1078 and v1300
How does neural computation arise from the combination of inputs along the dendritic tree? Axon extension of functionally recorded cells Complete: v1078 and v1300
To what extent do interareal cortical connections exhibit precise recurrent interactions, and are these connection probabilities governed by shared retinotopy? Axon extension of functionally recorded cells Complete: v1078 and v1300
What are the detailed connectivity properties of cortical PV basket cells? Annotation of parvalbumin interneurons Complete: v1078 and v1300
How does the connectivity pattern of SST/Chodl cells give rise to their physiological feature: synchronization of cortical states? Axon extension of putative Chodl cells Incomplete: pending putative Chodl cells
How do astrocytes interact with neurons, at multiple scales of organization and operation? Proofread 'cleaning' of astrocytes Complete: v1078
How does the presence of vasculature affect the distribution of electric charge due to an applied electric field (brain stimulation)? Analysis support Complete: v1078
What is the mechanism of electrical neuro-vasculature coupling, at the microscopic level? Analysis support Complete: v1078
Census of spinehead apparatus Annotation of pyramidal spine inputs Complete: v1300
OPC spacing and cilia orientation OPC proofreading and annotation Incomplete: pending resources
Myelin density in cortex Annotation of myelin Complete: v1300
Perivascular fibroblast Analysis support Complete
Local environment of spine synapses Annotation of pyramidal spine inputs Complete: v1300
Inhibitory-Excitatory connectivity motifs Axon extenstion of functionally coregistered cells Complete: v1300

VORTEX Contributions

Community Manager: Bethanny Danskin

Proofreading Management: Szi-chieh Yu, Celia David

Proofreading: Chi Zhang, Erika Neace, Rachael Swanstrom, Agnes Bodor, Mai Dumale, Elanur Acmad, Al-Mishrie Sahijuan, Joselito Astrologo, Arvin Caballes, Rey Cabasag, Jovie Cano, Jessi Deocampo, Emely Deresas, Mark Encallado, Mary Krestine Guzarem, Hannie Grace Hordillo, Logic Domme Ibanez, Angeline Libre, Geronimo Mon, Geomar Pagalan, Joren Paulines, Joshua Angel Remullo, Rezi Selgas,

Principal Investigators: Forrest Collman, Nuno Da Costa, R. Clay Reid, Casey Schneider-Mizell H. Sebastian Seung

NIH BRAIN Initiative: U24 NS120053