News‎ > ‎

Centering matters for PPI analysis

posted Jan 21, 2017, 10:56 AM by Xin Di
We found that the default setting of PPI calculation in SPM5 and SPM8 may generate spurious PPI effects, because of an interaction between mean centering and deconvolution process. We suggest that the psychological variable should be centered before calculating the PPI term, or we should add a deconvolve-reconvolve reversion of the physiological variable into the PPI model to account for reconvolution artifacts. For details on this issue, please read our new paper, which was just published in Human Brain Mapping. 

Di X, Reynolds RC, Biswal BB (2017). Imperfect (de)convolution may introduce spurious psychophysiological interactions and how to avoid it. Hum Brain Mapp doi: 10.1002/hbm.23413.