Dénes Szűcs

The paper estimates how much fMRI connectivity can be reliably detected given within-subject time course reliability and finds that less than a quarter of observed connectivity is reliably detectable.

The paper evaluates sample size trends in neuroimaging studies from 1990-2012 and 2017-2018, finding increasing sample sizes and low use of pre-study power calculations.

The paper critiques null hypothesis significance testing (NHST) and advocates for alternative statistical methods in biomedical and psychological research.