Skip to content

Specification Fields

This section details the primary fields defined in the signalJourney.schema.json.

Top-Level Fields

These fields are required at the root level of every signalJourney JSON file.

  • sj_version (string, required)
    • Description: The version of the signalJourney specification that this file conforms to. Must follow semantic versioning (e.g., "0.1.0").
    • Example: "sj_version": "0.1.0"
  • schema_version (string, required)
    • Description: The version of the specific signalJourney.schema.json file used for validation. Must follow semantic versioning.
    • Example: "schema_version": "0.1.0"
  • description (string, required)
    • Description: A brief, human-readable summary of the processing pipeline described in this file.
    • Example: "description": "EEG preprocessing pipeline including filtering and ICA."
  • pipelineInfo (object, required)
  • processingSteps (array, required)
    • Description: An ordered array detailing each individual step performed in the pipeline.
    • See: Processing Step Object

Optional Top-Level Fields

  • summaryMetrics (object, optional)
    • Description: Contains summary quality metrics that summarizes multiple qualtiyMetrics fileds across pipline or dataset. For example, sumamrizing quality metrics such as average powerline noise across subjects in a dataset.
    • Structure: Key-value pairs where keys are metric names and values are the metric results. Can be nested objects.
    • See: Quality Metrics Object
    • Example: "summaryMetrics": { "finalSNR": 35.2, "percentDataRejected": 5.1 }
  • extensions (object, optional)
    • Description: Container for domain-specific or custom extensions to the core specification, organized by namespace.
    • See: Namespaces for details on structure and governance.
    • Example: "extensions": { "eeg": { "channelCount": 64, "referenceType": "average" } }
  • versionHistory (array, optional)
    • Description: Records changes made to this specific signalJourney file over time. Each item is an object.
    • See: Version History Object

Pipeline Info Object

Describes the overall pipeline.

  • name (string, required)
    • Description: A descriptive name for the pipeline.
    • Example: "name": "Standard EEG Preprocessing"
  • description (string, required)
    • Description: A more detailed description of the pipeline's purpose and methods.
  • version (string, required)
    • Description: The version of this specific pipeline script/implementation (distinct from sj_version).
    • Example: "version": "1.2.0"
  • pipelineType (string, optional)
    • Description: A category describing the pipeline's main function (e.g., "preprocessing", "ica", "source-localization", "connectivity", "statistics").
    • Example: "pipelineType": "preprocessing"
  • executionDate (string, optional, format: date-time)
    • Description: The date and time when the pipeline was executed (ISO 8601 format).
    • Example: "executionDate": "2024-05-02T10:00:00Z"
  • institution (string, optional)
    • Description: The institution where the pipeline was run.
  • references (array, optional)
    • Description: List of relevant publications or resources related to the pipeline methods.
    • Items: Object with doi (string, required) and citation (string, optional) fields.
    • Example: "references": [ { "doi": "10.1016/j.neuroimage.2019.116046" } ]

Processing Step Object

A single step within the processing pipeline.

Required Fields: * stepId (string): A unique identifier for this step within the pipeline (e.g., "01_filter", "ica_run"). * name (string): A human-readable name for the step (e.g., "High-pass Filter", "ICA Decomposition"). * description (string): A brief description of what the step does. * software (array): An array of softwareDetails objects detailing the software used. * inputSources (array): An array of inputSource objects defining the inputs. Must contain at least one item.

Optional Fields: * parameters (object): Key-value pairs representing parameters specific to this step. * outputTargets (array): An array of outputTarget objects defining the outputs. * qualityMetrics (object): A qualityMetricsObject containing metrics relevant to this step's output. * extensions (object): Container for extensions.

inputSource Object

Defines a source of input data for a processing step.

  • description (string, required): Describes the input data (e.g., "Raw EEG data", "ICA component activations").
  • location (string, required): Path to the input file, relative to the dataset root or the signalJourney.json file itself (convention TBD, likely relative to dataset root for BIDS).
  • format (string, optional): The file format (e.g., "SET", "FIF", "EDF").

outputTarget Object

Defines the target location or method for storing the output of a processing step.

  • description (string, required): Describes the output data (e.g., "Filtered EEG data", "Cleaned IC weights", "Rejected channels list").
  • targetType (string, required): Specifies the storage method. Must be one of:
    • file: The output is stored in an external file (maps to location).
    • in-memory: Output exists in memory (e.g., a specific data structure in a script) but isn't explicitly saved to a file or variable name in this record. Use format for data type description.
    • variable: Output is stored as a named variable within the execution environment (maps to name). Use format for data type description.
    • report: Output is a report (e.g., HTML, PDF). May optionally use location. Use format for report type.
    • userDefined: A custom output type not covered by others (maps to details).
    • inlineData: The output data is embedded directly within the JSON (maps to data, encoding, formatDescription).
  • location (string, required if targetType is file, optional for report): Path to the output file, relative path convention same as inputSource.location.
  • format (string, optional, recommended for file, in-memory, variable, report): The file format, data type, or report type (e.g., "SET", "FIF", "TSV", "NumPy array", "HTML").
  • entityLabels (object, optional, applicable if targetType is file): Key-value pairs representing BIDS-like entities for the output file. Example: {"desc": "filtered"}.
  • name (string, required if targetType is variable): Name of the variable stored.
  • details (string, required if targetType is userDefined): Description of the user-defined output.
  • data (any, required if targetType is inlineData): The embedded data itself (e.g., a list of bad channel names, an ICA weight matrix as a nested array, QC metrics as an object).
  • encoding (string, optional if targetType is inlineData, default: "utf-8"): Specifies how the data is encoded if it's binary or needs special handling (e.g., "base64", "gzip+base64"). Defaults to standard JSON types if omitted.
  • formatDescription (string, required if targetType is inlineData): A human-readable description of the data field's format (e.g., "List of bad channel labels", "NumPy array serialized to list", "JSON object with QC scores").

Example inlineData Usage:

{
  "description": "List of channels rejected during artifact removal",
  "targetType": "inlineData",
  "formatDescription": "List of bad channel labels",
  "data": ["Fp1", "Oz", "T7"]
}

{
  "description": "ICA weight matrix",
  "targetType": "inlineData",
  "formatDescription": "ICA weights matrix (channels x components) as nested list",
  "encoding": null, // Or omit if standard JSON types
  "data": [
    [0.1, 0.5, -0.2],
    [-0.3, 0.8, 0.1],
    // ... more rows ...
  ]
}

Software Details Object

Describes the software used in a processing step. Corresponds to items in the processingSteps[].software array.

Required Fields: * name (string): The name of the software package or library (e.g., "MNE-Python", "EEGLAB", "FieldTrip"). * version (string): The specific version of the software used.

Optional Fields: * functionCall (string): The specific function or command executed (potentially abbreviated or representative). Example: "functionCall": "raw.filter(l_freq=1.0, h_freq=40.0)" * operatingSystem (string): The operating system the software was run on. * repository (string, format: uri): URL to the software's source code repository (e.g., GitHub link). * commitHash (string): Specific Git commit hash of the software version used, if applicable.

Quality Metrics Object

Container for quality metrics, used in summaryMetrics and processingSteps[].qualityMetrics.

  • Structure: Free-form key-value pairs. Keys should be descriptive metric names. Values can be numbers, strings, booleans, arrays, or nested objects.
  • Recommendation: Use consistent naming conventions (e.g., camelCase, snake_case) and include units where applicable (e.g., snrDb, latencyMs). Use namespaces for custom or tool-specific metrics within the extensions object if necessary.
  • Examples: 
    "qualityMetrics": {
  "channelsInterpolated": ["EEG 053"],
  "numComponentsRemoved": 3,
  "percentVarianceExplainedICA": 85.2,
  "eogCorrelationThreshold": 0.8
}

Extensions Object

Container for domain-specific or tool-specific information not covered by the core schema.

  • Structure: Key-value pairs where keys are namespace prefixes (e.g., "eeg", "mne", "iclabel") and values are objects containing the extension data.
  • See: Namespaces for details.
  • Location: Can appear at the top level (extensions) or within a processingStep (extensions).
  • Example (within a step): 
    "extensions": {
  "iclabel": {
    "componentClassifications": [
      {"icIndex": 0, "label": "Brain", "probability": 0.95},
      {"icIndex": 1, "label": "Eye", "probability": 0.88}
    ]
  }
}

Version History Object

Describes a single entry in the top-level versionHistory array.

Required Fields: * version (string): The version identifier for this entry (e.g., "1.1.0"). Should correspond to schema_version at the time of change. * date (string, format: date): The date this version was created (ISO 8601 format YYYY-MM-DD). * changes (string): A description of the changes made in this version of the signalJourney file.

Optional Fields: * author (string): The person or entity who made the changes.