Specification Overview¶
The signalJourney specification provides a standardized JSON format for describing biosignal processing pipelines and their outputs. Its primary goal is to enhance reproducibility and data sharing by capturing detailed provenance information.
Core Concepts¶
- Pipeline Description: Defines the overall goal, software environment, and execution context of the processing workflow.
- Processing Steps: Details each individual operation performed on the data, including the specific software, function calls, parameters used, inputs, and outputs.
- Data Provenance: Explicitly links processing steps, defining dependencies and tracking data flow.
- Quality Metrics: Allows for embedding quantitative or qualitative metrics about the data quality or processing outcomes at various stages.
- Extensibility: Uses a namespace system to allow for domain-specific extensions (e.g., for EEG, MEG) while maintaining a core standard.
File Structure¶
A signalJourney file is a JSON object with several key top-level fields:
sj_version
: The version of the signalJourney specification the file adheres to.schema_version
: The version of the JSON schema file itself.description
: A brief, human-readable description of the pipeline documented in the file.pipelineInfo
: An object containing metadata about the overall pipeline (name, version, type, execution date, etc.).processingSteps
: An array of objects, each detailing a single step in the pipeline.summaryMetrics
(optional): An object containing summary quality metrics for the entire pipeline output.extensions
(optional): An object containing namespaced, domain-specific extensions.versionHistory
(optional): An array documenting changes to the signalJourney file itself.
Purpose¶
By standardizing how processing pipelines are documented, signalJourney aims to:
- Improve reproducibility by capturing exact parameters and software versions.
- Facilitate data sharing and meta-analysis by providing rich, machine-readable provenance.
- Enable automated analysis of processing pipelines across different studies and labs.
- Provide a clear audit trail for complex data transformations.
See the Fields section for a detailed description of each component.