infrastructure: 

hardware

software

• import from <places> into Posda then into NBIA
• There is a hangup between SECTRA to Posda-to-ARIES
• Cancerimagingarchive.net has a landing page (Wordpress, Confluence, NBIA-for-search-and-download)

MongoDB to manage nonDICOM data?

1. What processes <are, should be> built?
2. JSON IS DECIDED as the right standard format to feed MongoDB 
   1. Potentially, in addition, CSV for tabular data (Posda understands DICOM, JSON, CSV)
   2. how to handle embedded formulae in excel multitabular data for example will be a Posda feature
3. Annotation based on this type of file format?
4. Potential descriptions from the PI for each column on a form
5. What does search mean - which data elements are Posda cleaning versus which pieces are Posda Harmonizing
6. Deale neuropsych ontology; Poldrack ontology?

1. Linkage between ARIES ID and Virmani lab ID is established via secure server site
2. Some of the data were not in SECTRA and don't have MRN (stored by name) but there exists a mapping table to ARIESID for all (we think)
3. This lab has loaded data to ~200 ID to SECTRA since ~2-3 years ago but older scans are on CD / hard drive in the lab
   1. OSF = "OutSideFilms" - initial project MR 
      1. some have multiple timepoints (watch for longitudinal changes)
   2. CT may have been used on pt that have Deep Brain Stim equipment installed - let's leave these out for now
   3. Dr. Virmani's lab would have to tell us which strings describing which scans to share
   4. DATSCAN SPECT PET scans - should also be ignored at this time
4. Start loading "things" to ARIES using the SECTRA PACS to pull to Posda & curate there.
   1. Phil may be getting added to this IRB to facilitate, email from Tuhin came recently discussing "routine clinical data includes images"
5. Many dates will be converted into intervals, perhaps age-at-time-of-X
6. There are items in AccessDB that are being converted to REDcap, how do I then put that into ARIES (from REDcap)

1. Passthrough ARIES IDs? Or consistent process through Sohrab's tool for this lab's project IDs? He can set this up for NDL (does not have linkages required for AR-CDR or SECTRA
2. Discussion with Aaron and Bill to set up sharedrive to pull data into Posda
   1. ORIGINAL DATA: DICOM
   2. "Processed Data" - NIfTI then (BIDS which also works with EEG format with JSON sidecar) CIfTI?  (to consider later)
   3. Spreadsheets associated have been deidentified; do NOT yet have ARIES ID in the spreadsheets - interval offsets can be applied programmatically but may not yet have been

1. Sohrab needs to know who from their team needs access to the utility that will create IDs for their study
2. Aaron can pass a bag of disks of images for import to Posda
3. This group is not planning to provide table data that doesn't already exist in the data warehouse (so we should leave that in its native form)
4. Which study were these subjects consented under - 

1. Can pull from SECTRA to Posda with project distinctions via AETitles
2. Quasar got Posda test up and running, Michael is downloading images and Terri can be trained when Bill are back

(data warehouse / EPIC / images)

these data might be fit for HPCcluster - Neurodocker container would need converted to singularity; singularity works better on a shared system than docker does

get neuro pipelines set up (test to get data from PRISM into local HPC quickly)

Brukker etc for small animal imaging

NIfTI

CIfTI

BIDS

Aperio