The effect of perceiving control on glutamatergic function and tolerating stress

Control over stressful experience results in subsequent inhibition of fear responses. We indexed glutamatergic levels in the ventromedial prefrontal cortex (vmPFC) before, and following, exposing humans to electric shocks, half of whom believed that they could control the shocks and subsequently presented with aversive images. Our findings show for the first time that perceived lack of control results in an increase in glutamatergic response in the vmPFC and heightened avoidance of a subsequent stressor. Our perceived capacity to control adverse events is critical for psychological functioning. Believing in the fact that one controls the outcomes, even when one does not actually have control, helps to reduce subsequent fear reactivity.1,2,3 Similarly, rats exposed to escapable (versus inescapable) shock are 'immunized' against subsequent stressors, which is reflected in reduced fear responses.4 Although there is evidence of the role of glutamate alterations following control manipulation in rats,5,6 there is no parallel evidence in humans. We investigated (a) the role of glutamate in perceived control over aversive experience in humans and (b) the relationship between glutamate change and subsequent stress response. We focused on the vmPFC because glutamate is activated in the vmPFC during stress, and there are projections from the vmPFC to emotional and decision-making functional regions.7 We hypothesized that lacking perceived control would result in greater glutamate because this would aggravate the experience of stress. We assessed glutamate levels in 29 healthy participants; the voxel of interest (15x15x15mm3) was placed adjacent to the anterior border of the genu of corpus callosum and centered on the interhemispheric fissure in order to include both left and right anterior cingulate cortex primarily (Figure 1). Participants were then subsequently exposed to standardized electric shocks. Half of the participants were (erroneously) informed that they could terminate the shock by pressing a button. Glutamate levels were then re-assessed. Fifteen minutes later, participants were then presented with aversive and neutral images, and were instructed that they could terminate the image by pressing a button; this task indexed the tolerance of distress (Supplementary Methods). Change in glutamate level before and after the shock differed between control conditions (F1,27=4.31, P=0.05; (Figure 1). Participants who lacked perceived control increased their glutamate level, following shock (t13=3.32, P=0.006), whereas those who perceived control did not (t14=0.28, P=0.79; Supplementary Results). Participants who lacked perceived control subsequently terminated aversive images sooner than those who perceived control (t24=3.13, P=0.005). There was no difference in viewing time for neutral images (t24=0.41, P=0.68). In the Control condition, glutamate decrease was associated with perceived prediction of the shocks (R2=0.57, P=0.025), but this was not apparent in the No-Control condition (R2=0.03, P=0.91) (see Supplementary Figure S1 and Supplementary Results). The observation that lacking perceived control led to increased glutamate in the vmPFC is consistent with evidence that acute stress leads to activation of glutamatergic transmission in the vmPFC,8 with subsequent activation of dopaminergic stress responses.7 The association of glutamate decrease and perceived prediction of shocks in those who perceived control suggests that perceived mastery over adverse stimuli (including capacity to predict occurrence) may be associated with diminished glutamate activation in the vmPFC. The 15-minute duration between manipulated control perception and measurement of subsequent behavior may have impacted the relationship, and it remains to be determined how glutamatergic activity after perceived control impacts subsequent behavior over different time courses. There are significant differences that are relative to animal research, where (a) only actual control over aversive stimuli can be manipulated and (b) control reduces gamma-aminobutyric acid (GABA)ergic interneurons in the vmPFC, which results in activation of glutamatergic output neurons to the dorsal raphe nucleus-the role of glutamate in the animal vmPFC is unknown9 (Supplementary Discussion). Future research needs to study the compatibility between neurotransmitter response to control over stressors in animals and humans, and also to determine neural responses to actual versus perceived control over aversive events. In summary, this study provides a novel direction for future research to investigate a potential neural modulator of how perceived control may impact behavior. This direction could provide important insights into how humans manage stressful events and their aftermath.