We propose an extension to the SNIRF specification to support Speckle Contrast Optical Spectroscopy (SCOS) data.
Several groups are now working on SCOS systems, and these are poised to be the next generation of blood flow monitoring devices. The SCOS modality is not included in the current specification, and we propose to include support for it through inclusion of new dataTypes to support SCOS data, and additional fields in the probe description to include exposure times.
The core idea is to include support for SCOS data adhering to SNIRF guiding principles enabling accurate data interpretation and open data sharing. This extension would allow SCOS systems to store the principal temporal outputs of a given acquisition in one standardized file while preserving their shared probe geometry, channel information, exposure times, and timing.
It would also allow SCOS and DCS data to coexist in the same SNIRF file.
Since much of SCOS literature is still evolving, we propose including the raw data in addition to blood flow index data streams, to facilitate evolution of data processing pipelines and cross validation studies in future which will further firm up the foundations of SCOS as a robust modality for blood flow monitoring. Readers that do not recognize the proposed SCOS data types would still be able to access the time-series arrays, while SCOS-aware readers could interpret the data types and exposure-time indexing.
The proposal adds:
- An exposureTimes field under /nirs(i)/probe.
- SCOS data types in the 500+ range.
- A defined relationship between exposureTimes and the existing dataTypeIndex field.
The existing SNIRF data structures are otherwise sufficient to store the SCOS time series.
The addition of exposure times supports including multi-exposure SCOS data, by tying each channel with an exposure time specification. This further aids in verification or cross validation of the data with theoretical frameworks that require the specification of exposure time.
We propose including three raw data types: Intensity mean, Intensity variance, Raw speckle contrast(Kraw2), Corrected speckle contrast (Kcorr2) and the processed data type : Blood flow index.
Allowing the flexibility to include Intensity mean and variance at the raw and/or processed stage of the speckle contrast data can provide an opportunity to evaluate the impact of noise-correction steps on the processed data. We believe at the current stage of technology development in SCOS, this is an important aspect to enable faithful data interpretation.
The blood flow index is included as a processed data type since it can be either generated from the speckle contrast directly, or using a model-based framework. We recommend including this information in the metadata.
At minimum, the BFi data should identify whether it was derived using:
inverseKcorr2, for example BFi = 1 / Kcorr2; or
modelDerived, where BFi was obtained from a SCOS forward or inverse model, calibration, or fit.
Additionally, we also recommend, when available, to provide a link to a repository that contains the source image files generated in the SCOS acquisition, and a description of the processing methods.
We believe, the existing SNIRF metadata fields allow a complete description of the hardware utilised, and do not need additional modifications at this stage.
The proposed changes make minimal modifications to the existing SNIRF format and support including the SCOS data into the format.
We propose an extension to the SNIRF specification to support Speckle Contrast Optical Spectroscopy (SCOS) data.
Several groups are now working on SCOS systems, and these are poised to be the next generation of blood flow monitoring devices. The SCOS modality is not included in the current specification, and we propose to include support for it through inclusion of new dataTypes to support SCOS data, and additional fields in the probe description to include exposure times.
The core idea is to include support for SCOS data adhering to SNIRF guiding principles enabling accurate data interpretation and open data sharing. This extension would allow SCOS systems to store the principal temporal outputs of a given acquisition in one standardized file while preserving their shared probe geometry, channel information, exposure times, and timing.
It would also allow SCOS and DCS data to coexist in the same SNIRF file.
Since much of SCOS literature is still evolving, we propose including the raw data in addition to blood flow index data streams, to facilitate evolution of data processing pipelines and cross validation studies in future which will further firm up the foundations of SCOS as a robust modality for blood flow monitoring. Readers that do not recognize the proposed SCOS data types would still be able to access the time-series arrays, while SCOS-aware readers could interpret the data types and exposure-time indexing.
The proposal adds:
The existing SNIRF data structures are otherwise sufficient to store the SCOS time series.
The addition of exposure times supports including multi-exposure SCOS data, by tying each channel with an exposure time specification. This further aids in verification or cross validation of the data with theoretical frameworks that require the specification of exposure time.
We propose including three raw data types: Intensity mean, Intensity variance, Raw speckle contrast(Kraw2), Corrected speckle contrast (Kcorr2) and the processed data type : Blood flow index.
Allowing the flexibility to include Intensity mean and variance at the raw and/or processed stage of the speckle contrast data can provide an opportunity to evaluate the impact of noise-correction steps on the processed data. We believe at the current stage of technology development in SCOS, this is an important aspect to enable faithful data interpretation.
The blood flow index is included as a processed data type since it can be either generated from the speckle contrast directly, or using a model-based framework. We recommend including this information in the metadata.
At minimum, the BFi data should identify whether it was derived using:
inverseKcorr2, for example BFi = 1 / Kcorr2; or
modelDerived, where BFi was obtained from a SCOS forward or inverse model, calibration, or fit.
Additionally, we also recommend, when available, to provide a link to a repository that contains the source image files generated in the SCOS acquisition, and a description of the processing methods.
We believe, the existing SNIRF metadata fields allow a complete description of the hardware utilised, and do not need additional modifications at this stage.
The proposed changes make minimal modifications to the existing SNIRF format and support including the SCOS data into the format.