A posterior insula to lateral amygdala pathway transmits US-offset information with a limited role in fear learning

This repository stores the raw data that gave rise to the study by Palchaudhuri et al. (2025) published in Cell Reports 44(2):115320 (doi: 10.1016/j.celrep.2025.115320). Please refer to the original publication regarding details of experimental design and methodology of data acquisition and analysis.

The data is indexed and described in the provided metadata (.csv) files which refer to individual .ZIP archives containing the raw and accessory data files.

Please note that the repository expands to a second and third repository accessible with separate dois: 10.5281/zenodo.14647369, and 10.5281/zenodo.14647435. The extension repositories contain no metadata files but .ZIP archives of the raw data beyond the space quota of the master repository. 

General structure of the repository:

·         The data is organized in 24 subsets. Each subset contains the dataset giving rise to the results in a given panel / group of panels shown in the main, or Supplementary Figures of the paper. Each subset is documented by its own metadata (.csv) file.

·         The .csv metadata files, one per data subset, describe the content of the individual .ZIP data archives. Field separator: comma.

·         Individual .csv metadata files are indexed by the main metadata file "metadata_master.csv". It also provides the location of the individual .ZIP archives (doi: of either the main of extension repositories). Field separator: comma.

Data formats of the raw data:

·         Widefield fluorescent images of coronal mouse brain sections were uploaded either as original unmodified datasets in a proprietary .vsi format of Olympus slide scanning microscope (for anatomical characterization of neural connectivity or mouse line characterization), or were registered to the digital Allen Brain Atlas, binned, and uploaded as composite image stacks in TIFF format (for the post-hoc histological analysis of the virus expression, fiber/lens placement, etc.). Confocal stacks were provided as composite image stacks in TIFF format. Both “.vsi” and TIFF formats are readable by Bioformats plugin under open-source free software packages FIJI (https://fiji.sc/) and/or QuPath (https://qupath.github.io/). Imaging metadata such as microscope settings and/or pixel resolution is embedded inside the individual .vsi or TIFF files. Correspondence between the image color channels and fluorescent signals, as well as the anterior-posterior spacing between the coronal brain sections, are provided in the corresponding .csv metadata files.

·         Due to excessive data volume (~1.5 Tb), the raw data (videos) for in-vivo miniature microscope Ca2+ imaging are not added to the repository. The data will be made available from the corresponding author upon request. Here we provide output .csv files generated by the CaImAn package from analysis of the miniature microscope Ca2+ imaging data. These include the fluorescence traces for each accepted and rejected cell, deconvolved Ca2+ events, and outlines of the ROIs for each detected neuron. We also provided the .csv files of GPIO signals exported from the Inscopix Data Analysis Software that contain digital timestamps of events (CS, US, LED light) needed for synchronization of the Ca2+ imaging data with the behavioral videos.  

·         Video recordings of animal behavior during the fear conditioning protocol were provided as unmodified “.wmv” files created by the VideoFreeze acquisition software (Med Associates Inc). Video stream parameters: wmv3 codec, color space yuv420p, 320x240 pixels, 30 fps, bitrate 300 kb/s. We also provide separate .csv files with the timestamps of events (CS, US) linked to the video timing, as well as the movement index traces resulting from the analysis of video recordings using ezTrack software package.

·         Data for patch clamp recordings were provided as original .dat files generated by the acquisition software PatchMaster (HEKA Elektronik, Germany). These files can be read using an Igor Pro extension “bpc_ReadHeka.xop” (for 32-bit Igor Pro versions 5.xx - 6.37; by Dr. Holger Taschenberger, https://www.wavemetrics.com/project/bpc_ReadHeka), or using a Python script by Luke Campagnola (https://github.com/campagnola/heka_reader), or using a Matlab script “HEKA PatchMaster Importer” by Christian Keine (https://github.com/ChristianKeine/HEKA_Patchmaster_Importer). In some cases, the numbers of recorded series within the PatchMaster data files corresponding to particular recording conditions were reported in the accompanying .csv files.

·         Extracellular optrode recording data were converted into the open HDF5 .h5 format from the original .mcd proprietary format of the MC_Rack software using the Multi Channel DataManager software (Multichannel Systems). The raw unfiltered electrode data is stored as a 32-bit integer matrix 16xN (16 - number of electrodes, N - number of sampling points @ 40 kHz) in the container Data->Recording_0->AnalogStream->Stream_1->ChannelData of the HDF5 files. The raw values have to be divided by the conversion factor 1.25e+6 to obtain the units of volts. Correspondence of the rows of the data matrix to the electrodes “E1”-“E16” is indexed by the string array Data->Recording_0->AnalogStream->Stream_1->I_Label. The electrodes were physically arranged into four tetrodes in the following groups: E1-E4, E5-E8, E9-E12, E13-E16. Digital TTL signals corresponding to the CS & US, or the laser light applied during the behavioral session are encoded, respectively, by the first and the second low bits of the integer array Data->Recording_0->AnalogStream->Stream_0->ChannelData.