Mesocorticolimbic Interactions Mediate fMRI-Guided Regulation of Self-Generated Affective States

Abstract

Increasing evidence shows that the generation and regulation of affective responses is associated with activity of large brain networks that also include phylogenetically older regions in the brainstem. Mesencephalic regions not only control autonomic responses but also participate in the modulation of autonomic, emotional, and motivational responses. The specific contribution of the midbrain to emotion regulation in humans remains elusive. Neuroimaging studies grounding on appraisal models of emotion emphasize a major role of prefrontal cortex in modulating emotion-related cortical and subcortical regions but usually neglect the contribution of the midbrain and other brainstem regions. Here, the role of mesolimbic and mesocortical networks in core affect generation and regulation was explored during emotion regulation guided by real-time fMRI feedback of the anterior insula activity. The fMRI and functional connectivity analysis revealed that the upper midbrain significantly contributes to emotion regulation in humans. Moreover, differential functional interactions between the dopaminergic mesocorticolimbic system and frontoparietal networks mediate up and down emotion regulatory processes. Finally, these findings further indicate the potential of real-time fMRI feedback approach in guiding core affect regulation.

Keywords: anterior insula; emotion regulation; midbrain; periaqueductal gray; real-time fMRI; reinforcement learning.

Figure 1

Violin plots representing estimated beta values for the contrast up > down emotion regulation in the midbrain (5, −26, −10), prefrontal cortex (PFC, −21, 30, −10), medial parietal cortex (MPC, 2, −30, 50), and left and right AI (−37, 20, 0 and 42, 13, 5). White circles show the medians; box limits indicate the 25th and 75th percentiles as determined by the R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles; polygons represent density estimates of beta values and extend to extreme values. Violin plots were created with BoxPLotR (http://shiny.chemgrid.org/boxplotr/) [65].

Figure 2

SPM maps showing the whole brain and specific midbrain activity during up emotion regulation with respect to down emotion regulation. Activation maps are superimposed on a Ch2 template using the MRIcron software (version 2016) [66].

Figure 3

Group-independent component analysis (ICA) showing highly functionally-connected networks involving mesencephalic areas during up emotion regulation (a,b) and during down emotion regulation (c).

Figure 4

Patterns of functional connectivity measured as a positive and negative correlation of the midbrain with the whole brain (seed-to-voxel connectivity maps) during up and down emotion regulation, separately (first and second row); positive and negative correlation of the AI with the whole brain during up and down emotion regulation, separately (third and fourth row).