Restricted access data in the neurosciences: are the restrictions always justified?

Personal healthcare information has historically been protected in law through physician–patient privilege, and recent legislation in the USA and the EU has sought to impose firm restrictions on sharing of such information. However, data access is a fundamental aspect of scientific research, and there is increasing conflict between data security/anonymity and the concepts of falsifiability and open science. This conflict is discussed from the perspective of the neurosciences (including cognitive science and experimental psychology) where researcher access to relevant data, even if anonymous, is increasingly constrained. In this age of ‘data protection', restricted access to personal data has popular appeal. But is it always justified? Epidemiological data are a special case because maximizing their utility constrains anonymization, and restrictions on access may be necessary. However, do we include anonymous genomic/transcriptomic sequence information, brain imaging data, brainwave recordings, eye-tracking data, body-posture recordings, Rorschach tests, or even microbiome studies? There is no evidence to date that, if properly anonymized, the identity of any individual can be deduced from such data (unless relevant data are already on the internet - ‘chicken and egg’). With the exception of epidemiological data, it is argued that objective evaluation is needed, and that restrictions on sharing anonymized datasets of other types should either be empirically based or set aside.