Complex relationships between depression and chronic pain have been reported in previous studies. However, only a few neuroimaging studies have investigated similarities and differences in neural systems underlying them. We examined the brain functions in the resting state of 43 patients with depression, 41 patients with chronic pain (somatoform pain disorder) and 41 healthy controls, by using regional homogeneity (ReHo) and functional connectivity analysis. Depressive symptoms were assessed by using the Beck Depression Inventory-Second Edition (BDI-II). ReHo values for the dorsolateral prefrontal cortex (DLPFC) significantly decreased for chronic pain patients, and functional connectivity between the DLPFC and thalamus decreased only for these patients. These findings are indicative of distinct brain functions related to depression and chronic pain. Understanding these differences would further elucidate the pathophysiology of these conditions.

A number of epidemiological studies have shown that both depression and chronic pain lead to decreased productivity, social disability, increased suicide rates and higher health care cost1,2,3,4. The association between depression and chronic pain has been supported by previous studies, including biological studies on neuroplastic, neurochemical, electrophysiological and hormonal variables, and psychological studies on pessimism and low self-esteem5, 6. Furthermore, different randomized controlled trials have reported that antidepressants have beneficial effects on both depressive symptoms and pain perception7, 8. Thus, it appears that depression and chronic pain might have certain commonalities.

Subtle differences between chronic pain and depressive patients have been reported. For instance, certain experimental studies on pain perception using thermal, or electrical stimuli have shown that chronic pain patients exhibit higher pain sensitivity than healthy controls9,10,11. However, other studies of depressive patients have indicated that they were less likely to perceive pain stimuli compared to controls12, 13. To our knowledge, there is only one study that has directly compared pain perception between depressive and chronic pain patients. Normand et al. examined potential differences in experimentally induced pain perception and diffuse noxious inhibitory control efficacy (e.g. “pain inhibits pain” phenomenon) between depressive and chronic pain patients by using a tonic thermal test and a cold pressor test14. They reported that chronic pain could be distinguished from depression by pain ratings during the cold pressor test, which might be related to the pain inhibition system. Another study has suggested that efficacy of antidepressants against depressive symptoms and pain, including the dose and onset of efficacy, might be based on independent mechanisms15. Moreover, path analysis has suggested that analgesic effects of antidepressants in chronic pain patients might be separately caused by a direct analgesic effect and an antidepressant effect16. They have indicated that 80% and over of the change in pain intensity in chronic pain patients could be ascribed to a direct analgesic effect of antidepressants, with the remaining effect affected by an antidepressant effect.

Various brain regions involved in the processing of sensation, emotion, cognition, and nociception are associated with the neuropathology of both depression and chronic pain17,18,19. The amygdala, hippocampus, insula, anterior cingulate cortex (ACC) and the prefrontal cortex (PFC) show alterations in depression19. On the other hand, the somatosensory cortex, thalamus, amygdala, hippocampus, insula, ACC and the PFC are involved in chronic pain17. Various studies of the linkage between depression and chronic pain have reported the ACC to be a critical region20, 21, and many neuroimaging studies have suggested that the ACC is significantly involved in depression or chronic pain17,18,19, 22, 23. On the other hand, based on the above-mentioned studies, distinctive brain regions involved in chronic pain would appear to be the somatosensory cortex and thalamus. Moreover, pain-related catastrophizing, that is mainly associated with etiology of chronic pain, and depression differed in the ways in which they impacted on pain experiences such as pain prediction24. Pain-related catastrophizing is closely related to activity in the dorsolateral prefrontal cortex (DLPFC)25, and of the PFC structures, the DLPFC seems to show the most differentiation between depression and chronic pain, playing different roles in each of these phenomena. Based on these points, we hypothesized that the activities of the DLPFC, thalamus and somatosensory cortex would be different between depression and chronic pain, and that the ACC would be a common, activated region in depressive and chronic pain patients, but not in healthy participants.

Recent evidence indicates that resting-state functional magnetic resonance imaging (R-fMRI) might be useful for investigating human cognitions, behaviors, emotions and somatic sensations26, 27. It is known that regional homogeneity (ReHo) has been shown to be sufficient to measure the local temporal synchronization of the time series of nearest neighbors during the resting state28, 29, and the functional connectivity method can manifest longer inter-regional changes19, 28. Previous ReHo or functional connectivity studies have suggested that these techniques were useful for increasing our understanding of neuropathology related to mental disorders or chronic pain27, 30,31,32. Furthermore, coordinating the ReHo method in the seed-based functional connectivity analysis may facilitate the sensitivity of each analysis and reduce the uncertainty of seed extraction33, and more sensitive analyses could contribute to further elucidation of the complex neurocircuitry underlying depression and chronic pain.

In order to test the present hypotheses, we examined underlying resting state neural abnormalities of 43 depressive patients, 41 chronic pain patients, and 41 healthy controls using ReHo (local activity) and functional connectivity methods (long-distance connectivity). We conducted a one-way analysis of covariance (ANCOVA) with Beck Depression Inventory-Second Edition (BDI-II) scores as a continuous factor, BDI-II scores * group as the interaction and age and gender as covariates of no interest.

Detailed demographic and clinical characteristics of participants are presented in Table 1. There were no significant differences in age, or gender among the groups (p = 0.46 for age and p = 0.38 for gender). A one-way ANOVA revealed a significant main effect of BDI-II scores, F (2, 122) = 113.8, p < 0.001, with BDI-II scores of depressive patients being significantly higher than those of chronic pain patients and the controls (Bonferroni p < 0.001). Moreover, BDI-II scores of chronic pain patients were significantly higher than those of the controls (Bonferroni p < 0.001).

Regional brain functions of each group are shown in Fig. 1. It can be seen that regions such as the medial temporal lobe, posterior cingulate cortex, postcentral gyrus and precuneus exhibited significantly higher ReHo values during the resting state.

Figure 1Full size image
ReHo values for one sample t-test (A) for patients with depression, (B) for patients with chronic pain, and (C) for healthy controls.

To examine differences in resting-state regional brain functions among depressive, chronic pain and control groups, we performed an ANCOVA with group as a categorical factor, BDI-II scores as a continuous factor, BDI-II scores * group as the interaction term and age and gender as covariates of no interest, because it is important to assess the implication of depressive state for each group. Significant BDI-II score * group interactions were observed for the DLPFC (Fig. 2A and Table 2: FWE corrected p < 0.05).

Figure 2
Figure 2

(A) Different ReHo values in the DLPFC are shown (p (FWE corrected) < 0.05). Scatter-plots illustrate these correlations. *p < 0.01, Bonferroni post hoc test. DLPFC; dorsolateral prefrontal cortex, D; depression, C; chronic pain and H; healthy controls. Each error bar represents one standard error. (B) Hypoconnectivity in the DLPFC and thalamus is noted in the chronic pain patients (p (FWE corrected) < 0.05). Scatter-plots illustrate these correlations. *p < 0.01, Bonferroni post hoc test. Each error bar represents one standard error.

Full size image

After controlling for…