Dr. Paul Thompson, Director at ENIGMA Center for Worldwide Medicine, Imaging & Genomics

Professor of Ophthalmology, Pediatrics, Neurology, Psychiatry and the Behavioral Sciences, Radiology, Biomedical Engineering, Electrical Engineering and Quantitative and Computational Biology, University of Southern California

This is in response to a request for a brief comment on the implementation of Nipoppy and Neurobagel tools for the ENIGMA-PD datasets. As PI for ENIGMA Consortium as a whole, I and my colleagues are immensely grateful for the work done to create an informatics infrastructure for ENIGMA-PD. As detailed in the paper, the data organization, streamlining, and standardization of workflows has many benefits, including the ability to understand what meta-data are available, run reproducible analyses, and manage data and workflows as the project scales. As multisite analyses grow in scale (more sites, more data modalities) and as multiple projects are running concurrently, there are immense requirements for informatics systems that can handle the complexity of the data and how it is managed. I am really impressed by the extremely well thought out efforts to create tools that allow scalable, reproducible analyses and there have been significant successes already.

Looking beyond this project, there are some challenges for other working groups/institutes/consortia trying to adopt similar standards and data practices. The Nipoppy team has handled these well. We all know that data management and informatics systems require resources (funding), skill, and dedicated interaction between technical teams and less technically skilled team members to help adopt these systems. In the long run, time is saved, and with better management, benefits accrue. In this case there were carefully structured in person meetings with the Nipoppy team traveling internationally to help address user questions. Having had their team visit us in person also, their commitment and skill is exceptional. There is a need for funding and dedicated personnel for these efforts to succeed, it cannot just be assumed that people will adopt them. Although in principle, in person hands-on interaction could be done remotely, in my experience this is challenging for all but the simplest tools, unless there is a large immediate pay-off. I have seen many good ideas that people do not adopt, due to limited attention span, overwhelm with other tasks, lack of time, and difficulty coordinating interactions on site. The in person aspect seems to have helped a lot here, as various simplifications (Zoom, conference workshops) do not always gain as much traction as an in person visit.