Exercise Strengthens Central Nervous System Modulation of Pain in Fibromyalgia

Abstract

To begin to elucidate the mechanisms underlying the benefits of exercise for chronic pain, we assessed the influence of exercise on brain responses to pain in fibromyalgia (FM). Complete data were collected for nine female FM patients and nine pain-free controls (CO) who underwent two functional neuroimaging scans, following exercise (EX) and following quiet rest (QR). Brain responses and pain ratings to noxious heat stimuli were compared within and between groups. For pain ratings, there was a significant (p < 0.05) Condition by Run interaction characterized by moderately lower pain ratings post EX compared to QR (d = 0.39-0.41) for FM but similar to ratings in CO (d = 0.10-0.26), thereby demonstrating that exercise decreased pain sensitivity in FM patients to a level that was analogous to pain-free controls. Brain responses demonstrated a significant within-group difference in FM patients, characterized by less brain activity bilaterally in the anterior insula following QR as compared to EX. There was also a significant Group by Condition interaction with FM patients showing less activity in the left dorsolateral prefrontal cortex following QR as compared to post-EX and CO following both conditions. These results suggest that exercise appeared to stimulate brain regions involved in descending pain inhibition in FM patients, decreasing their sensitivity to pain. Thus, exercise may benefit patients with FM via improving the functional capacity of the pain modulatory system.

Keywords: chronic pain; exercise; fibromyalgia; imaging; modulation.

Figures

Figure 1

Pain intensity (Int) and unpleasantness (Unp) ratings during the first run of each condition. FM = fibromyalgia; CO = control; EX = post-exercise; QR = post quiet rest.

Figure 2

Maps of BOLD responses to pain in nine FM patients. (A) BOLD response post-exercise (EX); (B) BOLD response post-quiet rest (QR); (C) within group differences in BOLD responses in the bilateral anterior insula (EX-QR). Color represents the β coefficient (% signal change ranging from −1% to 1%) and opacity represents t-statistic, with full opacity at a voxelwise t corresponding to p < 0.01. The significant cluster at α < 0.05 (corresponding to a cluster size threshold of 17 voxels) is outlined in white. Analyses were performed at 4 mm × 4 mm × 4 mm resolution using a mask of regions determined a priori from the hypotheses, which is highlighted in the background of image C; (D) Violin plots illustrating condition differences in %BOLD signal change in the left and right anterior insulae. Each point represents the average %∆BOLD in the cluster during one run, so that each subject is represented by three connected points. Shadows connecting points between violin plots indicate data from the same individual during each condition (EX & QR). Note that the region from which these points are drawn was chosen because there is a significant difference between EX and QR; the plot is intended to help clarify the within-subjects’ results.

Figure 3

Maps of BOLD responses to pain in 9 fibromyalgia (FM) patients and nine pain-free controls (CO). (A) BOLD response by group and condition (EX = exercise; QR = quiet rest); (B) significant group by condition interaction in BOLD responses in the left dorsolateral prefrontal cortex (DLPFC). Color represents the β coefficient (% signal change ranging from −1% to 1%) and opacity represents t-statistic, with full opacity at a voxelwise t corresponding to p < 0.01. The significant cluster at α<0.05 (corresponding to a cluster size threshold of 17 voxels) is outlined in white. Analyses were performed at 4 mm × 4 mm × 4 mm resolution using a mask of regions determined a priori from the hypotheses, which is highlighted in the background of image B; (C) Violin plot illustrating the simple main effects underlying the interaction in the DLPFC. Each point represents the average %∆BOLD in the indicated cluster during one run, such that each subject is represented by three connected points. Shadows connecting points between violin plots indicate data from the same individual during each condition (EX & QR). Note that the region from which these points are drawn was chosen because there is a significant interaction between group and condition; the plot is intended to further clarify the interaction.