SpikeSorting: Object Space Task Pipeline

Automatic and manual spike sorting for tetrode recordings based on SpikeInterface.

Dependencies

Same as SpikeInterface

Important: This pipeline uses the SpikeInterface version from 2020. Compatibility has not been tested with newer versions.

Please install each sorter as described here:
SpikeInterface Sorter Installation Guide

If the method above causes issues, try these detailed installation instructions (PDF).

When using Matlab-based sorters, save them in:

~/Documents/SpikeSorting/

To be used on Linux . Add this folder to your path temporarily:

bash export PYTHONPATH=$PYTHONPATH:/path/to/this/folder

To make it permanent, paste it into your bashrc:

bash gedit ~/.bashrc

---

Spike Sorting Pipeline

1. Activate your conda environment

bash conda activate <Name of environment>

---

2. Check for corrupt or uneven length files

bash python -m check_length /full/path/to/folder_with_Study_day_subfolders

Ensure trial folders dont have duplicate channel files like 100_CH1.continuous and 100_CH1_2.continuous.

---

3. Select trials or merge trials (Trial1 + PostTrial1 or full-day)

Use merge_channels_revised_T1_PT1.m or merge_channels_revised.m in MATLAB.
Ensure correct sampling rate (fs):
- Rats 19 used 20kHz, not 30kHz.

Check folders variable in the MATLAB script and make sure it lists the correct trial folders in the right order.

Refer to the file:
RAT_OS_EPHYS_Channel_Normalization_Across_Animals.xlsx (Dropbox)

Run merging twice:
1. Trial1 + PostTrial1
2. Presleep + T1 + PT1 + + T5 + PT5 (+ Novelty trial if applicable)

Refer to this Google Planning Sheet

Generate hpc.xlsx and cortex.xlsx with tetrode IDs and channel info.
Skip tetrodes with only 1 channel or reference tetrodes.

---

4. Fix file names if needed

bash python -m fix_channel_name /path/to/continuous_files

Copy the following files from the same study day (any trial or post-trial folder):

• all_channels.events
• Continuous_Data.openephys
• messages.events
• settings.xml
• tetrode.prb (same for all rats)
• hpc.xlsx and cortex.xlsx (omit hpc.xlsx if not sorting HPC)

Work locally! Copy the merged folder to your computer's data hard drive.

---

5. Rearrange data by tetrode

bash python -m rearrange_folders /path/to/merged_data

If only sorting cortical tetrodes, ensure rearrange_folders.py uses:

python folders = ['cortex']

Instead of:

python folders = ['hpc', 'cortex']

---

6. Run automatic spike sorting

bash python -m run_tetrodes /path/to/brain_region_folder_with_tetrodes

Prefer using Spyder (prevents crashes):

bash conda activate spiketutorial spyder

In Spyder: - Load run_tetrodes.py - Set the correct path in the folder variable - Press Ctrl+A then F9 to run the script

---

7. Create JSON for consensus

bash python -m create_json /path/to/brain_region_folder

Copy JSON files to the respective brain area folders of the full-day merged folder.

To run manual curation:

bash python -m run_tetrodes_manual /path/to/brain_region_folder_with_tetrodes

---

8. Manual curation with Phy

During manual curation: - Alt+N = Noise
- Alt+M = MUA
- Alt+G = Pyramidal unit
- Leave unlabeled = Interneuron

Always save first, then close. If you close first and then save, your curation won't be saved.

---

9. Activation matrix & consensus post-curation

Both scripts (run_tetrodes.py or run_tetrodes_manual.py) create a phy folder with:

• spike_times.npy
• spike_clusters.npy

You can generate activation matrices (25 ms bins) as actmat_auto_tetrode#.

To concatenate and detect assemblies:

bash python -m phy2assembly

To view previous detections:

bash os.system('phy template-gui phy_AGR/params.py')

Post-curation consensus:

bash python -m consensus_post_curation /path/to/tetrode_folder

---

10. Quality Metrics

bash python -m quality_metrics

Requires:

bash conda env create -f environment_qm.yml conda activate spiketutorial_qm

---

Running in Spyder

1. Open run_tetrodes.py
   
2. Press Ctrl+A, then F9

---

Running in Linux Terminal (Adrian's PC only)

bash ./loop

This calls run_tetrodes_brain_areas.py in a loop.
Download loop.sh rename to loop.

Add loop to your PATH in ~/.bashrc.

Modify sorter.py: add exit() to the end of ms4() and auto() functions.

---

Reading corrupted Open Ephys files

Navigate to:

/home/Username/anaconda3/envs/spiketutorial/lib/python3.6/site-packages/pyopenephys

Modify:

• openephys_tools.py
• core.py

---

Newest scripts (yet to be described)

• run_tetrodes_brain_areas.py
• run_tetrodes_json.py
• run_tetrodes_loop.py
• truncate2.py