Meta-analytic activation maps can help identify affective processes captured by contrast-based task fMRI: the case of threat-related facial expressions.

Abstract

Meta-analysis of functional magnetic resonance imaging (fMRI) data is an effective method for capturing the distributed patterns of brain activity supporting discrete cognitive and affective processes. One opportunity presented by the resulting meta-analysis maps (MAMs) is as a reference for better understanding the nature of individual contrast maps (ICMs) derived from specific task fMRI data. Here, we compared MAMs from 148 neuroimaging studies representing emotion categories of fear, anger, disgust, happiness and sadness with ICMs from fearful > neutral and angry > neutral faces from an independent dataset of task fMRI (n = 1263). Analyses revealed that both fear and anger ICMs exhibited the greatest pattern similarity to fear MAMs. As the number of voxels included for the computation of pattern similarity became more selective, the specificity of MAM–ICM correspondence decreased. Notably, amygdala activity long considered critical for processing threat-related facial expressions was neither sufficient nor necessary for detecting MAM–ICM pattern similarity effects. Our analyses suggest that both fearful and angry facial expressions are best captured by distributed patterns of brain activity, a putative neural correlate of threat. More generally, our analyses demonstrate how MAMs can be leveraged to better understand affective processes captured by ICMs in task fMRI data.