Back

Retinotopic mapping data permit accurate matching of participants across different datasets

Schwarzkopf, D. S.; Morgan, C. A.; Roberts, R. P.; Dakin, S. C.

2024-12-21 neuroscience
10.1101/2024.12.19.629552 bioRxiv
Show abstract

Public sharing of neuroimaging data is becoming increasingly common for the advancement and validation of scientific research. However, this sharing poses challenges regarding privacy and data safety, and associated questions about the ownership of research data. Here we show that a simple pattern correlation algorithm can match nominally deidentified participants across two separate functional magnetic resonance imaging (fMRI) experiments. This re-identification procedure is effective despite functional maps being spatially warped to a common template brain. This work highlights the need for appropriate safeguards against possible misuse of shared neuroimaging data.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.