Figure 1.
We used a 2×2 factorial design with the factors BEHAVIORAL CONTROL modulation (i.e., ‘control’ vs. ‘no control’ over the noxious stimulation) and SAFETY SIGNALING modulation (i.e., ‘expect low pain’ vs. ‘expect high pain’ noxious stimulation). During two of the four runs, the noxious stimulation could be stopped by the participant (‘control’). In the other two runs, participants had no control over the stimulation (‘no control’). Prior to each trial, participants were informed by a visual cue whether the upcoming stimulation would only consist of moderately painful stimuli (‘expect low pain’) or could also include high-intensity pain stimuli (‘expect high pain’; arrows pointing downwards). Trials during which high-intensity stimulation was delivered were excluded from the analysis (shown in white). Each trial was followed by ratings of pain intensity, helplessness and threat and a baseline period (not shown). Note that the same number of stimuli was applied in the ‘control’ and the ‘no control’ runs. Arrows pointing upwards indicate button presses that stopped the noxious stimulation in the ‘control’ runs and were performed after the stimulation had stopped in the ‘no control’ runs.
Figure 2.
(A) On average, only the safety signaling modulation (i.e., ‘expect low pain’ relative to ‘expect high pain’ stimulation) showed an analgesic effect. No significant difference in pain ratings was found between ‘control’ and ‘no control’ runs. (B) Individual pain intensity ratings, which show a spread such that half of the participants rated the pain as less intense when they could control it and the other half reported the opposite effect, i.e., increased pain during the two ‘control’ runs. In contrast to the pain intensity ratings, participants felt more helpless (C) and threatened (D) during the ‘expect high pain’ stimulation than they did during the ‘expect low pain’ stimulation and during ‘no control’ runs than they did during the ‘control’ runs. The interaction between both factors only reached statistical significance for the helplessness ratings. Error bars indicate the standard error of the mean. *p<0.05, ***p<0.001.
Figure 3.
Right VLPFC correlated with the analgesic effect of the behavioral control modulation.
The signal level in the rVLPFC during ‘control’ trials as compared to ‘no control’ trials was positively correlated with the difference in pain ratings between those conditions (for display purposes thresholded at p<0.005 uncorrected, minimum cluster extent: 5 voxels; shown on a glass brain on the left and overlaid on a standard structural image (in MNI space) on the right).
Figure 4.
Bilateral rACC and left DMPFC/DLPFC correlated with the analgesic effect of the safety signaling modulation.
(A) Brain regions showing a positive correlation with lower pain during ‘expect low pain’ trials than during ‘expect high pain’ trials, displayed on a glass brain (left: sagittal, mid: coronal, right: axial plane; p<0.001 uncorrected, minimum cluster extent: 5 voxels). The bilateral rACC is highlighted in orange. The left DMPFC/DLPFC is highlighted in red. (B) Bilateral rACC and (C) left DMPFC/DLPFC activation overlaid on a standard structural image (in MNI space).
Figure 5.
Left DLPFC correlated with the anxiolytic effect of the behavioral control modulation.
Activation in the left DLPFC was related to relatively lower ratings of anxiety (i.e., the composite score of perceived threat and helplessness) during the control modulation (p<0.001 uncorrected, minimum cluster extent: 5 voxels; shown on a glass brain on the left and overlaid on a standard structural image (in MNI space) on the right).
Figure 6.
PAG and VLPFC showing increased functional connectivity with the rACC during safety signaling modulation.
The right rACC that was correlated with the analgesic effect of the ’expect low pain’ modulation exhibited increased functional connectivity with the left PAG (x,y,z = −6,−33,−18), shown on a glass brain (A), and overlaid on a standard structural image in MNI space (B) (for display purposes thresholded at p<0.005 uncorrected, minimum cluster extent: 5 voxels). The left rACC showed increased connectivity with the right VLPFC (x,y,z = 57,24,0), displayed on a glass brain (C) and overlaid on a standard structural image in MNI space (D) (p<0.001; minimum cluster extent: 5 contiguous voxels).
Figure 7.
Schematic representation of the neural mechanisms underlying the analgesic and anxiolytic effects of both modulatory approaches.
Pain reduction following the ‘control’ modulation was associated with increased activation in the VLPFC. The safety signaling modulation, in contrast, led to pain reduction through the engagement of the descending pain inhibitory system including the rACC, which showed increased functional connectivity with the PAG and the VLPFC. Decreased anxiety during behavioral control was related to increased activity in the DLPFC. No distinct activation could be found for the effect of safety signaling on anxiety.