Data Release notes for the Cubic Millimeter (MM^3) Dataset
The MICrONS Cubic Millimeter dataset is undergoing continued proofreading and analysis. Public data releases include changes to the proofreading tables, additional annotation tables related to upcoming publications, manual work provided for the scientific community as part of VORTEX, and data products derived from the structural reconstructions and functional recordings.
Data releases:
Click on each version to view the detailed release notes page
Version: 117 — June 11, 2021
The original data release. Highlights include: automatic segmentation with manual proofreading of X cells, synapse detection, nucleus segmentation, and manual cell typing.
Version: 343 — February 24, 2022
Highlights include: the ‘flat segmentation’ of the cellular segmentation meshes, enabling multi-resolution viewing and download; cell-type classifiers trained on the segmentation as of v117.
Version: 661 — April 6, 2023
Highlights include: an additional X manually proofread cells; and coregistration between the two-photon functional data and the electron microscopy cell positions.
Version: 795 — August 23, 2023
Highlights include: updated cell type classifiers run on the segmentation as of v661.
Version: 943 — January 22, 2024
Highlights include: derived visual tuning properties of the functional recordings, associated with the coregistered cells.
Version: 1078 — June 6, 2024
Highlights include: Additional axonal extension for 109 cells, proofreading of 6 astrocytes; manual labels for myelin, nodes of Raniver, and spine-synapse compartments targeted by basket cells. Additional manual coregistration of functional cells, and automatic nearest-match. Automatic labeling of synapse compartment types.
Version: 1181 — September 16, 2024
Highlights include: Additional axonal extension for 273 cells, proofreading of 6 astrocytes; manual labels for spine-synapse compartments as part of VORTEX.
Consider also:
For information on the manual proofreading efforts, see Proofreading Strategies
View the datastack metadata via the Info Service (requires google login).