E-I Dynamics: Analysis Pipeline

Author: Aditya Asopa
Bhalla Lab, NCBS

Installation

Required: Python 3.7+

1. Install Python 3.7.x (compatibility issues for later versions with allenSDK and x-to-nwb)

2. make virtual environment

   python -m venv venv-lab

3. Activate virtual envirionment

   \venv-lab\scripts\activate

4. Install libraries

   pip install -r requirements_detailed.txt

5. Register venv-lab kernel in jupyter-lab/ipython

   python -m ipykernel install --user --name=venv-lab

---

Pre-requisites for running the pipeline on a cell

• Every recorded cell should have its own folder.
• Inside the folder, every experiment (exptID) should have one exptID_rec.abf file.
• Every .abf file is accompanied by an exptID_experiment_parameters.py file that contains the experiment parameters for that experiment.
• Coordinates of optical stimulation are contained in a folder called polygon Protocols. Each file in this folder is a separate polygon protocol with rows as frames, and columns as frame number, grid size x, grid size y, spot coordinates for that frame. For ex:

  3 24 24 205 251 297 389 447
  Means, 3rd frame is a grid of 24x24 pixels, and 5 squares at index 205,251...447 are illuminated.

• An all_cells.py file contains a variable called all_cells that is a list of all the folders (i.e. cells).
• The program is run by executing batchAnalysis.py.

  python batchAnalysis.py -a all_cells.py

---

Structure of the Pipeline

  flowchart-elk LR;
    batch_analysis  --> parse_cell
    rd((recordingData)) --> dt{data type}
    dt  --  if abf file --> parse_recording
    dt  --  if directory--> parse_cell

        parse_cell  --> parse_recording
        parse_cell  --> make_dataframe
        parse_cell  --> data_quality_checks
        parse_cell  --> add_cell_to_xl_db
        parse_cell  --> save_training_set
        parse_cell  --> save_cell

        parse_recording --> load_cell
        parse_recording --> initNeuronObj[init Neuron]
            initNeuronObj   --> cell_param_parser
        parse_recording --> addExperiment
            addExperiment   --> initExptObj[init Experiment]
                initExptObj --> exptParamParser
                initExptObj --> abf_to_data
            addExperiment   --> analyse_experiment
                analyse_experiment  --> if{if protocol}
                    if  -->  FreqSweep  
                    if  -->  SpikeTrain 
                    if  -->  Convergence
                    if  -->  LTMRand  
                    if  -->  Surprise  
            addExperiment   --> update_experiment  
        parse_recording --> saveCell

Loading

---

Examples

1. Analyse all recordings on one cell

• Have a folder named as cellID. for ex. 'CA1PC001'
• Make sure all the recording files that are going to be part of the cell are named with the suffix xxxxx_rec.abf, where xxxx is the experiment ID.
• Make sure all the recording files also have a corresponding xxxx_experiment_parameter.py.
• The experiment parameter file should follow this template

python analysis.py /path/to/cell/folder

2. To analyse one recording

python analysis.py /path/to/xxxx_rec.abf