Behavioural Force Data (BeForData)

Data structures for handling behavioural force data

This package provides core classes and utilities for loading, processing, and analysing behavioural force data, such as those collected in experimental psychology or neuroscience. It offers a structured approach to manage epochs and records of force measurements, enabling efficient data manipulation and analysis.

BeForData is based on two structured classes of force data: one for the representation of the raw time-based force measurements in the shape of a dataframe (BeForRecord) and one for epoch-based representations as matrices (BeForEpochs).

Features

• Flexible loading and saving of force data in common formats (e.g., CSV, XDF).
• Efficient slicing and indexing of epochs and records for batch analysis.
• Metadata management for experimental context, including event markers and annotations.
• Utilities for preprocessing, such as filtering and baseline correction.
• Integration with scientific Python libraries (NumPy, pandas) for advanced analysis.

Source code: https://github.com/lindemann09/befordata

Documentation: https://lindemann09.github.io/befordata/

(c) Oliver Lindemann

Data Structures

• BeForRecord

  Represents a single continuous recording of force data, including metadata such as sampling rate, channel information, and experimental annotations. BeForRecord supports data cleaning, resampling, and extraction of epochs, and provides convenient access to raw and processed force signals.

  The data structure has the following attributes:

  • dat: DataFrame containing force measurements and optionally a time column.
  • sampling_rate: Sampling rate of the force measurements (Hz).
  • sessions: List of sample indices where new recording sessions start.
  • time_column (optional): Name of the column containing time stamps (if any).
  • meta (optional): Arbitrary metadata associated with the record.

• BeForEpochs

  A container class for managing multiple epochs of force data. Each epoch represents a segment of continuous force measurements, typically corresponding to a trial or experimental condition. BeForEpochs provides methods for slicing, indexing, and batch-processing epochs, as well as for loading and saving epoch data from various formats.

  The data structure has the following attributes:

  • dat: 2D numpy array containing the force data (epochs x samples).
  • sampling_rate: Sampling rate of the force measurements (Hz).
  • design: DataFrame containing design/metadata for each epoch.
  • zero_sample: Sample index representing the sample of the time zero within each epoch (default: 0).
  • baseline (optional): 1D numpy array containing baseline values for each epoch at zero_sample.
  • meta (optional): Arbitrary metadata associated with the epochs.

Typical Workflow

1. Load raw force data into a BeForRecord object.
2. Preprocess and annotate the data as needed.
3. Segment the data into epochs using event markers, creating a BeForEpochs object.

Install via pip

pip install befordata

Julia

A Julia implementation of BeForData is available as a beta release.