Infrequent strong connections constrain connectomic predictions of neuronal function (3/3)

Currier, Timothy A.1 ; Clandinin, Thomas R.1

Published Jun 10, 2025 on Dryad. https://doi.org/10.5061/dryad.kh18932k1

Data files

Jun 10, 2025 version files 217.56 GB

ImagingData1B.zip
217.56 GB
README.md
6.83 KB

Abstract

How does circuit wiring constrain neural computation? Recent work has leveraged connectomic datasets to predict the functions of cells and circuits in the brains of multiple species. However, many of these hypotheses have not been compared with physiological measurements, obscuring the limits of connectome-based functional predictions. To explore these limits, we characterized the visual responses of 43 cell types in the fruit fly and quantitatively compared them to connectomic predictions. We show that these predictions are accurate for some response properties, such as orientation tuning, but are surprisingly poor for other properties, such as receptive field size. Importantly, strong synaptic inputs are more functionally homogeneous than expected by chance and exert a disproportionately large influence on postsynaptic responses. Finally, we quantitatively define the subset of connections that best describe the functional differences between cell types. Our results establish a powerful set of constraints for improving the accuracy of connectomic predictions.

This is a README file for the data deposition associated with the manuscript “Infrequent strong connections constrain connectomic predictions of neuronal function” by Timothy A. Currier and Thomas R. Clandinin (Cell, 2025). Any questions or comments can be directed to Tim (currier@stanford.edu)

DOIs

The raw imaging data have been split over three Dryad depositions:
- All files except ImagingData1 https://doi.org/10.5061/dryad.pg4f4qs1j
- ImagingData1A https://doi.org/10.5061/dryad.bnzs7h4ns
- (THIS DATASET) ImagingData1B https://doi.org/10.5061/dryad.kh18932k1

LOOKING FOR STRFS?

If you are simply looking for STRFs for each recorded cell/cell type:
- Data S1 and Data S2 in the manuscript
- Download ALL_RESPONSES.npy here

The deposition includes the following:
- Response summaries for all recorded neurons (.npy)
- Connectome data (various formats)
- Brain and Cell logs (.xls)
- Stimulus & ROI metadata files (.hdf5)
- High-resolution anatomical volumes (.nii)
- ROI annotations (.key)
- Functional scan volumes (.nii)

WARNING!

Python packages required to fully interact with this VERY RAW data:
- Visanalysis
- Original code for this manuscript

DETAILED INFORMATION

The following files (except half of the .nii files in this dataset) can be found at https://doi.org/10.5061/dryad.pg4f4qs1j

ALL_RESPONSES.npy

A compendium of all cells recorded for the study. The last dimension of each dict value is 571, the number of ROIs. The dict contains the following keys:

'cell_IDs', the unique cell identifiers (ROIs with the same cell_ID were recorded from the same cell)
'cell_types', the cell type assignment for each ROI (empty elements indicate that no identification was made)
'blue_STRFs', the STRF for each ROI responding to blue noise (dims: x, y, t, ROI#)
'uv_STRFs', the STRF for each ROI responding to UV noise (dims: x, y, t, ROI#)
'STRF_time', the time vector for STRF data
'flicker_responses', the dF/F response for each ROI responding to full-screen flicker at 0.1, 0.5, 1, or 2 Hz (dims: f, t, ROI#)
'flicker_time', the time vector for flicker data

brain_log.xlsx

A list of the flies imaged for the study, including the recording date, the fly genotype, and an estimate of the number of cells in each fly.
This file is for housekeeping only (redundant with .hdf5 metadata files)

ME0708_full_log_snap.xls

A list of ROIs recorded for the study. My analysis scripts use this log as the primary source of metadata.
Includes information about the stimulus series # for blue, UV, and flicker presentation, as well as a cell type assignment, when an ID was possible.
ROIs from unique cells have a unique cell ID. When multiple ROIs come from the same neuron, the cell ID is repeated for those ROIs.
The weird file name is retained for compatibility with the GitHub repo.

Stimulus & ROI metadata files (.hdf5)

Located in the StimData folder. Each recording date has one file.
Contains stimulus and ROI metadata for each fly recorded on the corresponding date.
To open, launch the Visanalysis DataGUI.py file and load the .hdf5 file. More details are available in the Visanalysis documentation.

Connectome Data (various formats)

These are companion files to run a number of the analysis scripts in the GitHub repo; it is recommended to drop the unzipped folder into your cloned repo directory.
complete_metrics courtesy of the Reiser Lab, summary quantifications of neuronal morphology
Filenames that begin with __EXTRACTED__ are the full type-to-type connectivity data from Nern et al., Nature, 2025 - used for calculating Euclidean connectivity distance
Filenames that begin with Rsr_ are type-to-type weighted Jaccard distance matrices for the Nern et al. data
FAFB_distances courtesy of Sebastian Seung - type-to-type weighted Jaccard distances for flywire data, Matsliah et al., Nature, 2024
HighN_InFracs... scraped from the Reiser Lab’s Cell Type Explorer Website

Anatomical Scans (.nii)

Each file contains a high-resolution scan through the optic lobe of a recorded fly. For dates with multiple files, compare the volume to the snapshots in the ROI annotation files to determine the specific fly for each file.
Typically, lower-numbered anatomical scans on a specific date refer to lower-numbered fly IDs (see brain_log.xls).
When two channels are present for a specific scan, channel_1 is myr::tdTomato and channel_2 is GCaMP8m. When only one channel is present, it is GCaMP8m.
These scans were exclusively used for the anatomical identification of recorded cells.

ROI Annotation Files (.key)

Each recording date has one annotation file, comprised of slides with snapshots of the high-resolution anatomical scan and numbers indicating the ROIs for each brain on that date.
Snapshots are ordered from most superficial (i.e., most posterior part of the brain) to most deep (most anterior).
This is meant to be a handy visual guide to direct users to more careful inspection in the full anatomical scan. The exact ROI position and shape can be seen in DataGUI.

Functional Scans (.nii)

These files are located at https://doi.org/10.5061/dryad.bnzs7h4ns and
https://doi.org/10.5061/dryad.kh18932k1

Located in the ImagingData folders. ImagingData1(A, B) contains the first ~300 cells that were largely collected over a 6-month period in late 2022 - early 2023. ImagingData2 contains an additional ~215 cells that were collected a few months later in Oct-Dec of 2023.
Each “series” (stimulus) has two associated files: the motion-corrected fluorescence data (TSeries-*.nii, xyzt data), and a volume with the fano factor of each pixel (*_thresh_brain.nii, xyz data).
The series number corresponding to blue noise, UV noise, or full-screen flicker presentation for any given recording date can be found in ME0708_full_log_snap.xls.
It is not recommended to open the fluorescence data files directly, as they are quite large. Visanalysis uses an ROI mask to extract raw fluorescence timecourses on which all subsequent analyses are based.
The fano factor volumes were used as a preliminary means of identifying noise-responsive cells for ROI curation.