Multisite Comparison of MRI Defacing Software Across Multiple Cohorts

February 2021
Journal: Frontiers in Psychiatry
Lead Author: Athena E. Theyers

All Authors: Athena E. Theyers; Mojdeh Zamyadi; Mark O’Reilly; Robert Bartha; Sean Symons; Glenda M. MacQueen; Stefanie Hassel; Jason P. Lerch; Evdokia Anagnostou; Raymond W. Lam; Benicio N. Frey; Roumen Milev; Daniel J. Müller; Sidney H. Kennedy; Christopher J. M. Scott; Stephen C. Strother; on behalf of The ONDRI Investigators and Stephen R. Arnott

With improvements to both scan quality and facial recognition software, there is an increased risk of participants being identified by a 3D render of their structural neuroimaging scans, even when all other personal information has been removed. To prevent this, facial features should be removed before data are shared or openly released. While there are several publicly available software algorithms to do this, there has been no comprehensive review of their accuracy within the general population. This paper addresses this need, through cross-collaboration with Ontario Brain Institute’s CAN-BIND and POND programs.