Catalogger

Catalogger is a Python package for loading, cataloging, and analyzing data from high-density microelectrode array (HD-MEA) recordings, with a focus on experiments involving organoids and mouse brain slices. It leverages cataloged metadata to streamline data access and analysis.

Features

• Load and update experiment catalogs from metadata and CSV files.
• Extract and process spike data from HD-MEA recordings.
• Annotate catalogs with computed metrics (e.g., firing rates, burstiness).
• Support for filtering experiments based on quality metrics.
• Utilities for handling large-scale electrophysiology datasets.

Installation

First, install the SpikeData dependency:

bash pip install git+https://github.com/braingeneers/SpikeData

Then install Catalogger:

bash pip install git+https://github.com/hschweiger15/Catalogger.git

Or clone the repository and install locally:

bash git clone https://github.com/hschweiger15/Catalogger.git cd Catalogger pip install .

Requirements

• Python >= 3.8
• See setup.py for a full list of dependencies, including: pandas, numpy, scipy, matplotlib, seaborn, h5py, tqdm, joblib, boto3, pyarrow, tables, nptyping, pyqt5, xarray, and more.

Usage

Loading a Catalog

```python from Catalogger.Loaders import AcqmLoader import pandas as pd

Load a catalog CSV

catalog = pd.readcsv('catalogbaseline.csv')

Initialize loader

loader = AcqmLoader(basepath='/path/to/data', catalog=catalog) ```

Updating Catalog with Spike Data

python updated_catalog = loader.update_catalog(gen_metrics=True, min_units=5)

Accessing Spike Data

python spike_data = loader.get_spike_data('Trace_20240619_12_24_17_20174a_day46_ventral_c57')

Data Structure

• catalog_baseline.csv: Main experiment catalog, with columns for UUID, experiment name, date, sample type, species, cell line, media, drug, chip number, etc.
• metadata.json: Detailed metadata for each experiment, including hardware, sample info, and data file paths.

Example Catalog Entry

| uuids | experimentname | experimentdate | sampletype | species | cellline | aggdate | platingdate | orgage | media | patterning | drug | chipnumber | |------------------------------|-------------------------------------------------|---------------------|-------------|-----------|-----------|------------|--------------|-----------------|--------|------------|----------|-------------| | 2024-05-04-e-sakura-day22and24 | Trace2024050413261121985day22ventrale14 | 2024-05-04 13:27:00 | organoid | organoid | E14 | 2024-04-11 | 2024-04-29 | 23 days 13:27:00| sakura | ventral | | 21985e |

Drug Loading

Catalogger includes a DrugLoader class for managing and analyzing drug application metadata and effects in your experiments. This class helps you load drug information, associate it with recordings, and perform downstream analyses.

Example Usage

```python from catalogger.Loaders import DrugLoader import pandas as pd

Load the catalog

catalog = pd.readcsv('catalogbaseline.csv')

Initialize the DrugLoader

drug_loader = DrugLoader(catalog=catalog)

Access drug information for a specific experiment

experimentname = catalog['experimentname'].iloc[0] druginfo = drugloader.getdruginfo(experimentname) print(druginfo) ```

• The DrugLoader can be used to filter experiments by drug, analyze drug effects, and integrate drug metadata into your analysis workflows.

```

Example Usage

```python import pandas as pd import matplotlib.pyplot as plt from catalogger.Loaders import AcqmLoader from catalogger.cell_styler import Styler

Load the catalog

catalog = pd.readcsv('catalogbaseline.csv')

Set your basepath

basepath = '/path/to/data' # Update to your actual data path

Instantiate the loader

loader = AcqmLoader(basepath=basepath, catalog=catalog)

Update the catalog with spike data (optional)

loader.updatecatalog(genmetrics=True, min_units=5)

Select a recording name

recordingname = loader.catalog['experimentname'].iloc[0]

Get the SpikeData object for the recording

sd = loader.getspikedata(recording_name)

Create a Styler for standardized plotting

styler = Styler( fontsize=12, ticksize=10, linewidth=1.5, colorpalette='deep', rastermarker='|', rastermarker_size=8 )

Plot the raster with standardized style

fig, (ax, ax2) = plotstyledraster( sd, title=f"Raster for {recordingname}", sortby_fr=True, styler=styler ) plt.show() ```

• Set sort_by_fr=True to sort neurons by firing rate (highest at bottom).
• The function supports custom time windows, population rate units, and saving plots to file.
• For advanced styling, pass a Styler object (see cell_styler.py).

Styler for Standardized Plotting

The Styler class allows you to standardize the appearance of your plots for publication-quality figures. You can customize font sizes, line widths, color palettes, and marker styles. Pass a Styler instance to plot_styled_raster to apply consistent formatting across your figures.

```python from catalogger.cell_styler import Styler

styler = Styler( fontsize=12, ticksize=10, linewidth=1.5, colorpalette='deep', rastermarker='|', rastermarker_size=8 ) ```

Plotting and Visualization

Catalogger provides an example function for visualizing spike data as a styled raster plot with population firing rate overlay.

plotstyledraster

```python def plotstyledraster(spikedata, title=None, timewindow=None, ylim=None, styler=None, widthpt=None, heightpt=None, savepath=None, filename=None, smoothingwindow=100, sortbyfr=False, timeunit='seconds', poprateunit='Hz'): """ Plot a styled raster plot with population firing rate.

Parameters:
    spike_data (SpikeData): SpikeData object containing spike times and neuron data
    title (str, optional): Title for the plot
    time_window (tuple, optional): Time window to display in time units specified by time_unit
    y_lim (tuple, optional): Y-axis limits for population rate (min, max)
    styler (Styler, optional): Styler object for plot formatting
    width_pt (float, optional): Width of the figure in points
    height_pt (float, optional): Height of the figure in points
    save_path (str, optional): Path to save the figure
    smoothing_window (float, optional): Bin size in ms for population rate calculation
    sort_by_fr (bool, optional): Whether to sort neuron IDs by firing rate
    time_unit (str, optional): Unit for x-axis, either 'seconds' or 'minutes'
    pop_rate_unit (str, optional): Unit for population rate, either 'Hz' or 'kHz'
Returns:
    tuple: (fig, (ax1, ax2)) where ax2 is None if no spikes are present
"""
# ...see examples/ex_usage.py for full implementation...

Contributing

Contributions are welcome! Please open issues or pull requests on GitHub.

License

MIT License

Author

Hunter Schweiger (hschweig@ucsc.edu) Ash Robbins