Login / Signup

A novel tool for time-locking study plans to results.

Matan MazorNoam MazorRoy Mukamel
Published in: The European journal of neuroscience (2018)
Often researchers wish to mark an objective line between study plans that were specified before data acquisition and decisions that were made following data exploration. Contrary to common perception, registering study plans to an online platform prior to data collection does not by itself provide such an objective distinction, even when the registration is time-stamped. Here, we adapt a method from the field of cryptography to allow encoding of study plans and predictions within random aspects of the data acquisition process. Doing so introduces a causal link between the preregistered content and objective attributes of the acquired data, such as the timing and location of brain activations. This guarantees that the preregistered plans and predictions are indeed specified prior to data collection. Our time-locking system does not depend on any external party and can be performed entirely in-lab. We provide code for easy implementation and a detailed example from the field of functional Magnetic Resonance Imaging (fMRI).
Keyphrases
  • magnetic resonance imaging
  • electronic health record
  • big data
  • healthcare
  • health insurance
  • primary care
  • machine learning
  • magnetic resonance
  • deep learning
  • data analysis
  • functional connectivity
  • contrast enhanced