• Select the area you would like to search.
  • ACTIVE INVESTIGATIONS Search for current projects using the investigator's name, institution, or keywords.
  • EXPERTS KNOWLEDGE BASE Enter keywords to search a list of questions and answers received and processed by the ADNI team.
  • ADNI PDFS Search any ADNI publication pdf by author, keyword, or PMID. Use an asterisk only to view all pdfs.


Question Posted 05/31/20:
My colleagues and I are attempting to summarize information for subjects within the ADNI database and we have encountered some issues when trying to merge information from different .csv files for each participant.

Elaborating on this issue: The .csv file containing basic demographic information of the MR scans (as downloaded from Advanced search results section of the website) contains the participants SubjectID, ImageID and + some cognitive information.
With reference to the website\'s help section, the last 4 digits of the SubjectID is the RID. (E.g. for SubjectID 006_S_4485, RID is 4485).

With this information, when we attempt to combine information from the MMSE.csv containing MMSE scores and other info. (as provided in Download study data > Assessments > Neuropsychological> Mini-Mental State Examination (MMSE) [ADNI1,GO,2,3]) for each participant, It is unclear as to which MMSE score corresponds to the relevant demographic information in the demographics file; (a) the MMSE scores in the demographics file and MMSE file do not match even for the same subject, or (b) in some cases, the number of MMSE scores provided for each subject in the two files are different.

Since the aforementioned attempt to merge information was problematic, we used the ADNIMERGE.csv file available on your website (download study data > ALL > Key ADNI tables merged into one table) instead of using the MMSE.csv file as it contains a summary of most data we need. In this case, when we try to combine information from ADNIMERGE.csv with the demographics file, the dates for the scans and/or tests in the two files do not tally (even if we consider only baseline data). Therefore, if the files are merged based on dates that are close, we still cannot know for sure if the merged information is accurate.

Given these issues arising when merging disparate files, we would highly appreciate if you could guide us on how to obtain all the information available in the ADNIMERGE and/or MMSE files corresponding to subjects in a demographics file that is generated/downloaded through an advanced search in the ADNI website.
Thank you for your time and we look forward to hearing from you soon.
Response posted 05/31/20 by Danielle:
Part of the confusion is that the images use text descriptors of the visit (such as “ADNI2 Initial Visit-Cont Pt” or “ADNI2 Year 1 Visit”, while the .csv files on the Study Data page use VISCODEs (both the tracking VISCODE and VISCODE2 which translates VISCODE into an interpretable month visit (such as “m12” or “m96”). There is a file on the Study Data page (VISITS.csv found under the Enrollment section), that helps to some degree, although the VISNAME variable there is not always an exact match with the text descriptors in the imaging files.

There are explanations for differences in dates: 1) a visit takes place over several (often non-consecutive) days, so the image may be acquired on a different day than the neuropsych testing. The date in the image file would be the date the scan was acquired. 2) the initial visit for a person is split across the screening visit and the baseline visit. Some neuropsych scores are obtained at the screening visit, while others are at the baseline visit. The MRI is at the “screen visit”, so sometimes there are differences just based on which initial visit collected the information; 3) to simplify things, ADNIMERGE does not split the screen and baseline visit, so the measures collected at screen and the baseline visit all fall in the same row of the ADNIMERGE file with a VISCODE of “bl” – I haven’t looked closely as to which date is used there, but it would be either the screen or baseline date.

As for differences in MMSE values, it is difficult to troubleshoot this without knowing more specifics of the search parameters you used to generate your image collection, I did my own search, and although I didn't check all of the MMSE values in the .csv from the image collection compared to MMSE.csv, the ones I checked did match. I'm happy to troubleshoot more if you send me a direct email (djharvey@ucdavis.edu).

LONI is working on a new data query system for ADNI. Once launched, it should help with the merging issue.
Go back to list of topics >