Dear CNI Users,

In August we wrote to you regarding a problem we found within the CNI Flywheel instance (cni.flywheel.io). In some cases the physiological data (respiratory and cardiac waveforms) from the scanner were not matched to the correct acquisition in Flywheel, or were not uploaded at all. We have determined that such failures occurred most often when a scan ended prior to the prescribed acquisition duration, The matching logic within the reaper used the duration information.  In some cases the mis-match led the reaper to associate the wrong physiological data with an acquisition; in other cases no match was found and thus no physiological data were uploaded.

We maintain backups of the physiological data, and this is enabling us to correct the association of the physiological data with its proper acquisitions. The restoration has been ongoing in the background for some time, and we expect to complete the process within the next couple of weeks.

Incorrectly associated physiological data in an acquisition will remain within the acquisition and be renamed using the following convention:

 <exam>_<series>_<acquisition>_original.gephysio.zip

Projects in which we found one or more acquisitions containing incorrect or missing physiological data will have a CSV file attached to the project. The files will be named using the following convention:

physio_fix_<group_name>-<project_name>.csv

Those ‘fix’ files will contain detailed information describing each acquisition that had physiological data associated with it. The contents of the file are described at the end of this message.

Please note that physiological data is only automatically associated for EPI-type sequences. If you store physiological data for other sequence (e.g. T2 spin-echo, T1-weighted fast gradient-echo, etc.) then you will need to manually upload the physiological data to Flywheel if you’d like it kept with your acquisition.  Please contact us directly if you have questions about this process.

We appreciate your patience as we work through this issue. If you have any questions regarding this process, or concerns about the data that have been restored to your sessions, please reach out and let us know either via email or Slack.

Contents of the ‘fix’ file

The ‘fix’ file is a text description (CSV) about the corrected physiological data.  It includes:

• Session ID – A string representing the Flywheel Session ID
• Acquisition ID – A string representing the Flywheel Acquisition ID
• Flywheel Path - A human readable string showing the path to the acquisition in Flywheel
• FIX Status – A string indicating whether the new data were ‘found’ or ‘replaced’
• Message – describing the action performed, or an errors during execution.
• Exam – Exam Number
• Series – Series Number
• TR – A number defining the TR of the acquired data in milliseconds
• NTP – An integer defining the  number of temporal samples in the physiological data file
• RAW NTP Number of temporal positions (from the DICOM or PFile header),
• Scan Duration – Calculated scan duration
• PPG_Samples - the number of Physio Samples
• PPG_Lines_Expected – the number of lines we expected in the PPG file
• PPG_Lines – the number in samples found in the PPG file found in Flywheel (if applicable)
• PPG_Lines_Fix – Number of sample found in the Matched/Fixed data,
• DELTA_FIX_EX – the difference between the fixed data and what was expected
• DELTA_FW_EX – the difference between what was in Flywheel and what we expected
• Date and Time – of the acquisition
• FIX Time Delta Difference between expected timestamp and FIX file timestamp in Sec
• FW Time Delta Difference between expected timestamp and FW file timestamp in Sec
• FW File Base Base file name of data found in Flywheel
• Regx – The regular expression used to find the matched data

Thanks and stay safe!

The CNI Team