Intracellular membrane potential noise recordings in cortical pyramidal cells after blocking action potentials and synaptic transmission and simulations of miniature excitatory postsynaptic potentials in cortical pyramidal cells. The data in this repository was used to benchmark 'minis' software:

Martynas Dervinis 23bb145499 Delete 'images/outputStructure.png' 7 months ago
recs 378d7aae9e Added p131c 7 months ago
temps eb4efcf0ad Added p103a 8 months ago
LICENSE fab3eb3c5a Added CC-BY 4.0 license 7 months ago
README.md 952b2291d2 Update 'README.md' 7 months ago
datacite.yml ea4a7dc7b7 Update 'datacite.yml' 7 months ago

README.md

minis-benchmarking-data

Intracellular membrane potential noise recordings in cortical pyramidal cells after blocking action potentials and synaptic transmission (0.5 or 1 μM TTX, 12.5 μM gabazine, 40 μM NBQX, and 50 μM CPP) and simulations of miniature excitatory postsynaptic potentials (mEPSPs) in cortical pyramidal cells. The data in this repository was used to produce the manuscript of the minis software benchmarking study.

Repository content

recs

recs folder holds data for all individual recordings in separate folders. Folders are named according to recording IDs: p103a, p106b, p108a, p108b, p108c, p120b, p122a, p124b, p125a, p127c, p128c, p129a, p131a, p131c. Recording IDs are formed by appending the animal ID with the slice ID letter. Each recording ID folder has the following structure:

Folder name Description
abf This folder contains Axon Binary Format files produced by simulating excitatory postsynaptic potentials (EPSPs) and superimposing them on top of the filtered and smoothed intracellular membrane potential noise recordings. These files are produced by minis software running in the Simulate mode. minis in turn calls simulateDetectEvaluate function which performs the simulation, runs the minis detection algorithm on the noise + simulated EPSP time series data, and evaluates detection performance using signal detection measures. There are 4 simulated files per each test condition as described in the manuscript. File name parts identifying different conditions are listed below the table.
abf_raw This folder contains ABF files that correspond to files in abf folder prior to being smoothed. These files are used to detect simulated EPSPs by Clampfit. Smoothed recordings are not suitable for Clampfit detection.
csv_MA This folder contains Mini Analysis simulated EPSP detection output tables saved in the CSV format. The CSV files correspond to each ABF file in the abf folder.
csv_pC_raw This folder contains Clampfit simulated EPSP detection output tables saved in the CSV format. The CSV files correspond to each ABF file in the abf_raw folder.
mat_MA Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf and csv_MA folders. Mini Analysis part of the manuscript Figures 4 and 6-8 are based on the data stored in this folder. The output is saved in the MAT format and contains variables that are explained below this table. The output files are produced by the evaluateMiniAnalysisPostprocessing function.
mat_MA_select Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf and csv_MA folders. Mini Analysis part of the manuscript Figures 9-11 are based on the data stored in this folder.
mat_MA2 Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf and csv_MA folders. Mini Analysis part of the manuscript Figures 12-15 are based on the data stored in this folder.
mat_minis Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf folder. minis part of the manuscript Figures 4 and 6-8 are based on the data stored in this folder. The output is saved in the MAT format and contains variables that are explained below this table. The output files are produced either by simulateDetectEvaluate or by evaluateMinisPostprocessing functions.
mat_minis_select Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf folder. minis part of the manuscript Figures 9-11 are based on the data stored in this folder.
mat_minis2 Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf folder. minis part of the manuscript Figures 12-15 are based on the data stored in this folder.
mat_pC_raw Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf_raw and csv_pC_raw folders. Clampfit part of the manuscript Figures 4 and 6-8 are based on the data stored in this folder. The output is saved in the MAT format and contains variables that are explained below this table. The output files are produced by the evaluatepClampPostprocessing function.
mat_pC_raw_select Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf_raw and csv_pC_raw folders. Clampfit part of the manuscript Figures 9-11 are based on the data stored in this folder.
mat_pC_raw2 Simulated EPSP detection performance evaluation analysis output files corresponding to individual files in the abf_raw and csv_pC_raw folders. Clampfit part of the manuscript Figures 12-15 are based on the data stored in this folder.
<filename>.abf Recording of intracellular membrane potential noise stored in the ABF file.

Test Condition File Naming

Files stored in folders abf, abf_raw, csv_MA, csv_pC_raw, mat_MA, mat_MA_select, mat_MA2, mat_minis, mat_minis_select, mat_minis2, mat_pC_raw, mat_pC_raw_select, and mat_pC_raw2 correspond to individual testing conditions when part of their name contain the following characters:

  • amp0p3_sAmp0p05_n500_<...>_noiseScaleFactor4p2: 0.3 mV simulated EPSPs at 2.5 events/second with the noise scaling factor of 4.2.
  • amp0p3_sAmp0p05_n500_<...>_noiseScaleFactor2p6: 0.3 mV simulated EPSPs at 2.5 events/second with the noise scaling factor of 2.6.
  • amp0p3_sAmp0p05_n500_<...>_noiseScaleFactor1p8: 0.3 mV simulated EPSPs at 2.5 events/second with the noise scaling factor of 1.8.
  • amp0p3_sAmp0p05_n500_<...>_noiseScaleFactor1p4: 0.3 mV simulated EPSPs at 2.5 events/second with the noise scaling factor of 1.4.
  • amp0p3_sAmp0p05_n500_<...>_noiseScaleFactor1p2: 0.3 mV simulated EPSPs at 2.5 events/second with the noise scaling factor of 1.2.
  • amp0p3_sAmp0p05_n500_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 2.5 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n1000_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 5 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n2000_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 10 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n4000_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 20 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n8000_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 40 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n1600_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 80 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n3200_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 160 events/second with the noise scaling factor of.
  • amp0p3_sAmp0p05_n6400_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 320 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp0p05_n12800_<...>_noiseScaleFactor1: 0.3 mV simulated EPSPs at 640 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp1e_09_n500_<...>_noiseScaleFactor1: Varying amplitude simulated EPSPs at 13 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp1e_09_n625_<...>_noiseScaleFactor1 or amp0p3_sAmp1e_09_n1250_<...>_noiseScaleFactor1: Varying amplitude simulated EPSPs at 33 events/second with the noise scaling factor of 1.
  • amp0p3_sAmp1e_09_n1000_<...>_noiseScaleFactor1 or amp0p3_sAmp1e_09_n2000_<...>_noiseScaleFactor1: Varying amplitude simulated EPSPs at 53 events/second with the noise scaling factor of 1.

Detection Performance Evaluation Analysis Output

Detection performance evaluation analysis output is produced separately for each testing condition and is saved in Matlab data files (MAT format) containing the following variables:

Variable name Description
dPrime d' sensitivity index for each simulation file.
FPR False positive rate for each simulation file.
sensitivity Sensitivity (or true positive rate) for each simulation file.
specificity Specificity (or true negative rate) for each simulation file.
performance Performance indicators for each simulation file. There are 7 indicators (rows) with the same sampling rate as the original data (columns). Row 1: Simulated EPSP positions. Row 2: Positions of hits (where detected rather than simulated) and misses. Row 3: Positions of hits and false alarms. Row 4: Hit positions. Row 5: Missed event positions. Row 6: False alarm positions. Row 7: Positions of correctly rejected noise events.
falseI Indices of noise events. Also could be reprhrased as sample points corresponding to the peaks of designated noise events.
falseT Times of noise events (peaks of designated noise events).
filename Noise recording file name.

temps

Clampfit EPSP templates (Axon Text File format) used to detect simulated EPSPs and described in the benchmarking manuscript.

datacite.yml
Title Intracellular membrane potential noise recordings and mEPSP simulations in cortical pyramidal cells used to benchmark 'minis' software
Authors Dervinis,Martynas;University of Copenhagen;ORCID:0000-0002-4676-7456
Description Intracellular membrane potential noise recordings in cortical pyramidal cells after blocking action potentials and synaptic transmission (0.5 or 1 μM TTX, 12.5 μM gabazine, 40 μM NBQX, and 50 μM CPP) and simulations of miniature excitatory postsynaptic potentials (mEPSPs) in cortical pyramidal cells. The data in this repository was used to produce the manuscript of the minis software benchmarking study.
License Creative Commons Attribution 4.0 (https://creativecommons.org/licenses/by/4.0/legalcode)
References Martynas Dervinis, Guy Major (2022) Novel method for reliably measuring spontaneous postsynaptic potentials/currents in whole- cell patch clamp recordings in the central nervous system. bioRxiv 2022.03.20.485046; doi: https://doi.org/10.1101/2022.03.20.485046 [https://doi.org/10.1101/2022.03.20.485046] (IsSupplementTo)
Martynas Dervinis. (2023). dervinism/minis-benchmarking: minis-benchmarking (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.8336130 [https://doi.org/10.5281/zenodo.8336130] (IsReferencedBy)
Funding
Keywords Neuroscience
PSP
EPSP
Electrophysiology
Cortex
Ephys
Synapse
PSC
EPSC
Patch clamp
pClamp
Clampfit
minis
EPSC
Icephys
Quantal analysis
MiniAnalysis
Benchmarking
Resource Type Dataset