Simplified Power and Sample-Size Calculations Using Prevalence and Magnitude of Active Peaks

• Activity Number: 506
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Joke Durnez*+ and Beatrijs Moerkerke and Thomas E. Nichols
• Companies: Ghent University and Ghent University and University of Warwick
• Keywords: fMRI ; power ; neuroimaging ; sample sizes ; power calculations ; multiple testing

Abstract

There is increasing concern about statistical power in neuroscience research. Not only does low power decrease the chance of detecting a true effect, but it also reduces the chance that a statistically significant result indicates a true effect. In other words, findings from a low-power study are unlikely to be reproducible, and thus a power analysis is key for any paper. While power is straightforward to compute for a univariate response, determining the power of a fMRI study is a formidable task as the magnitude, spatial extent and location of a hypothesized effect are difficult to specify. In this work we present a simple way to characterize the spatial signal in a fMRI study, and a direct way to estimate power based on an existing pilot study. Using just the volume of the brain activated and the average effect size in activated brain regions, we can calculate power for given sample size, brain volume and smoothness. This procedure allows minimizing the cost of an fMRI experiment, while preserving a predefined statistical power. We found our method accurately estimates the effect size in a given study, which enables power and sample size calculations for new studies.