Emotion-motion interactions in conversion disorder: an FMRI study

Abstract

Objectives: To evaluate the neural correlates of implicit processing of negative emotions in motor conversion disorder (CD) patients.

Methods: An event related fMRI task was completed by 12 motor CD patients and 14 matched healthy controls using standardised stimuli of faces with fearful and sad emotional expressions in comparison to faces with neutral expressions. Temporal changes in the sensitivity to stimuli were also modelled and tested in the two groups.

Results: We found increased amygdala activation to negative emotions in CD compared to healthy controls in region of interest analyses, which persisted over time consistent with previous findings using emotional paradigms. Furthermore during whole brain analyses we found significantly increased activation in CD patients in areas involved in the 'freeze response' to fear (periaqueductal grey matter), and areas involved in self-awareness and motor control (cingulate gyrus and supplementary motor area).

Conclusions: In contrast to healthy controls, CD patients exhibited increased response amplitude to fearful stimuli over time, suggesting abnormal emotional regulation (failure of habituation / sensitization). Patients with CD also activated midbrain and frontal structures that could reflect an abnormal behavioral-motor response to negative including threatening stimuli. This suggests a mechanism linking emotions to motor dysfunction in CD.

Fig 1. ROI analysis: Group effect in the Emotion Model.

Upper panel: Activation in the amygdalae in response to emotional faces showing a group effect (CD patents > healthy controls). Statistical maps have been thresholded at p<0.005 uncorrected for display purposes. A significant increased activity was found in the left amygdala (FWEp = 0.027, z = 3.3, [–22–8–14]). Of note, the right amygdala activation displayed here did not reach conventional levels of significance after correction for multiple comparisons (FWE p = 0.078 at MNI [22–4–14]). Lower panel: Graph showing contrast estimates for left amygdala activation in response to each negative emotion (sad, fear) separately for the 2 groups: CD patients (in orange) and healthy controls (in green).

Fig 2. ROI analysis: Interaction effect in the Linear-change Model.

Graph showing group-level beta estimates for the linear change in the responsiveness of the left amygdala to the presentation of emotional stimuli. Here a positive beta indicates a linear increase in the magnitude of the response elicited by stimuli following repeated exposure. Negative betas indicate the response amplitude reduced over time (see S1 Fig for a graphical representation of the predicted BOLD time series scaled by these group beta coefficients).

Fig 3. Whole-brain analysis: Group effect in the Emotion Model.

Statistical maps have been thresholded at p<0.001 uncorrected. Upper left panel: Activation in a frontal cluster showing a significant group effect (CD patients > healthy controls). Upper right panel: Plots of beta estimates at MNI [–16 22 60] (left supplementary motor area).Lower left panel: Activation in a midbain cluster (including the periaqueducal grey area) showing a significant group effect (CD patients > healthy controls). Lower right panel: Plots of beta estimates at MNI [–4–28–20] (midbrain).