Pipeline Configuration File

The configuration file for PyLossess is a YAML file that specifies various parameters and settings for the processing steps. The file can be read and parsed when instantiating the the class:~pyLossless.config.Config class, e.g. pylossless.config.Config("path/to/my/config_file.yaml"). If you don’t already have a configuration file, you can instantiate an empty configuration class (e.g. pylossless.config.Config()) and then use the meth:~pylossless.config.Config.load_default method to initialize the pipeline with default settings.

If you load the default configuration, you will most likely want to modify some of the settings before running the pipeline. You can do this by either editing the configuration file directly or by modifying the attributes of the class:~pyLossless.config.Config instance in your Python code. To save the configuration instance back to a YAML file (for parameter editing or for later use), you can use the meth:~pylossless.config.Config.save method, e.g. config_instance.save("path/to/my/modified_config_file.yaml").

The configuration file includes settings for various steps in the PyLossless pipeline, but also includes settings for describing metadata about the dataset, and settings for passing arguments to third-party dependencies (e.g. MNE-Python).

To see examples of the Configuration files, see below:

• Adult Example

• Infant Example

Configuring Pipeline Steps

Each processing step in the PyLossless pipeline has its own section in the configuration file. Each section includes parameters that control the behavior of that step. For example, the preprocessing section includes parameters for filtering, resampling, and artifact rejection. The source_reconstruction section includes parameters for selecting the source space, forward model, and inverse method.

Pipeline steps

Step Name	Description
filtering	Parameters for band-pass, high-pass, and low-pass filtering of the data.
find_breaks	Parameters for detecting and marking breaks between experimental tasks.
noisy_channels	Parameters for identifying and marking noisy channels in the data.
noisy_epochs	Parameters for identifying and marking noisy time periods in the data.
uncorrelated_channels	Parameters for identifying and marking uncorrelated channels in the data.
uncorrelated_epochs	Parameters for identifying and marking uncorrelated time periods in the data.
bridged_channels	Parameters for identifying and marking bridged channels in the data.
ica	Parameters for performing ICA to identify and remove artifacts from the data.

Pipeline helper configurations

Helper Name	Description
epoching	Parameters for specifying how to epoch the continuous data into segments.
nearest_neighbors	Parameters for specifying sensor neighbors.
montage_info	Parameters for specifying sensor montage information.