Given the amygdala's role in survival mechanisms, and its pivotal contributions to psychological processes, it is no surprise that it is one of the most well-studied brain regions. One of the common methods for understanding the functional role of the amygdala is the use of functional magnetic resonance imaging (fMRI). However, fMRI tends to be acquired using resolutions that are not optimal for smaller brain structures. Furthermore, standard processing includes spatial smoothing and motion correction which further degrade the resolution of the data. Inferentially, this may be detrimental when determining if the amygdalae are active during a task. Indeed, studies using the same task may show differential amygdala(e) activation. Here, we examine the effects of well-accepted preprocessing steps on whole-brain submillimeter fMRI data to determine the impact on activation patterns associated with a robust task known to activate the amygdala(e). We analyzed 7T fMRI data from 30 healthy individuals collected at sub-millimeter in-plane resolution and used a field standard preprocessing pipeline with different combinations of smoothing kernels and motion correction options. Resultant amygdalae activation patterns were altered depending on which combination of smoothing and motion correction were performed, indicating that whole-brain preprocessing steps have a significant impact on the inferences that can be drawn about smaller, subcortical structures like the amygdala.
Keywords: 7T; Amygdala; Preprocessing; Reproducibility; Submillimeter resolution.
Published by Elsevier B.V.