Electron-Microscopy-Based Synaptic Connectivity in Early Postnatal Mouse Cerebellum

This dataset contains information about synaptic connectivity in mouse cerebellum (vermal lobule VIII) at postnatal days 3 and 7. Information was obtained by measuring annotations of two serial-section electron microscopy image volumes. These image volumes and our annotations can be visualized at http://bossdb.org/project/wilson2019.

For each dataset, a spreadsheet describes the synaptic inputs to a fully reconstructed Purkinje cell. It also identifies all the synaptic targets of those axonal inputs that were classified as either mossy fibers or climbing fibers, based on morphology and connectivity. Various properties of these axons and their synapses were also characterized and are documented.

This work was used to describe changes in synaptic connectivity that occur between climbing fiber axons and Purkinje cells during the first postnatal week, when there is evidence that one of the multiple climbing fiber inputs to a Purkinje cell may be becoming stronger than the others (see Kano et al. 2018 but also Scelfo and Strata 2005). We find that climbing fibers increase their synaptic strength during this time by selectivity adding synapses onto a few of their Purkinje cell targets, in a rich-get-richer fashion, while remaining connected with non-preferred Purkinje cell targets. We also find that this synapse addition does not require other climbing fiber inputs to lose synapses. We find that branches of the same climbing fiber have shared preferences for Purkinje cells in a local region. Finally, we find that the number of climbing fiber axons that innervates a local region of cerebellar cortex at this time roughly matches the number of Purkinje cells there. Thus, developmental circuit rewiring, or synapse rearrangement, between climbing fibers and Purkinje cells in the cerebellum appears to be economical, meaning that just enough axons innervate a local region in early postnatal development so that all can participate in adult connectivity there, and that no axon branched to that region in vain.

A preprint manuscript detailing our results can be found on bioRxiv: https://www.biorxiv.org/content/10.1101/627299v1

Code that makes use of this information is available on the Baby Andross repository: https://github.com/amwilson149/baby-andross.