A Randomized Controlled Neuroimaging Trial of Cognitive Behavioral Therapy for Fibromyalgia Pain

Abstract

Objective: Fibromyalgia (FM) is characterized by pervasive pain-related symptomatology and high levels of negative affect. Mind-body treatments such as cognitive behavioral therapy (CBT) appear to foster improvement in FM via reductions in pain-related catastrophizing, a set of negative, pain-amplifying cognitive and emotional processes. However, the neural underpinnings of CBT's catastrophizing-reducing effects remain uncertain. This randomized controlled mechanistic trial was designed to assess CBT's effects on pain catastrophizing and its underlying brain circuitry.

Methods: Of 114 enrolled participants, 98 underwent a baseline neuroimaging assessment and were randomized to 8 weeks of individual CBT or a matched FM education control (EDU) condition.

Results: Compared with EDU, CBT produced larger decreases in pain catastrophizing post treatment (P < 0.05) and larger reductions in pain interference and symptom impact. Decreases in pain catastrophizing played a significant role in mediating those functional improvements in the CBT group. At baseline, brain functional connectivity between the ventral posterior cingulate cortex (vPCC), a key node of the default mode network (DMN), and somatomotor and salience network regions was increased during catastrophizing thoughts. Following CBT, vPCC connectivity to somatomotor and salience network areas was reduced.

Conclusion: Our results suggest clinically important and CBT-specific associations between somatosensory/motor- and salience-processing brain regions and the DMN in chronic pain. These patterns of connectivity may contribute to individual differences (and treatment-related changes) in somatic self-awareness. CBT appears to provide clinical benefits at least partially by reducing pain-related catastrophizing and producing adaptive alterations in DMN functional connectivity.

Figure 1.

CONSORT flow diagram of the study.

Figure 2.

Treatment-induced changes. A, Changes in clinical outcome measures (BPI interference, severity, FIQR, and PCS). B, Inter-correlation between changes in outcome measures (BPI interference, FIQR, and PCS). CBT: Cognitive Behavioral Therapy group, EDU: Education control group, BPI: Brief Pain Inventory, PCS: Pain Catastrophizing Scale, FIQR: Fibromyalgia Impact Questionnaire-Revised, PRE: Baseline/before-treatment, POST: Post-treatment. * P < 0.05, ** P < 0.01.

Figure 3.

Mediation analysis evaluating changes in PCS scores as mediators for group differences of changes in clinical outcome measures. Mediation models were controlled for demographic covariates as well as general measures of negative affect (i.e., anxiety and depression, which are generally moderately associated with pain catastrophizing). CBT: Cognitive Behavioral Therapy group, EDU: Education control group, PCS: Pain Catastrophizing Scale, FIQR: Fibromyalgia Impact Questionnaire-Revised. * P < 0.05, ** P < 0.01.

Figure 4.

vPCC functional connectivity during pain catastrophizing. A, At baseline (N = 98), during engagement with pain catastrophizing statements, whole-brain functional connectivity from vPCC was reduced to other default mode network areas (e.g., PC and mPFC), and was increased to salience network areas (i.e., aINS, aMCC). B, Following CBT, vPCC connectivity to somatomotor (S1, M1) and salience (aMCC) network areas was reduced. The M1 cluster showed a significant interaction between Group (CBT and EDU) and Time (PRE- and POST-treatment) (N: CBT = 55, EDU = 26). PC: precuneus, mPFC: medial prefrontal cortex, SMA: supplementary motor area, aINS: anterior insula cortex, aMCC: anterior midcingulate cortex. S1: primary somatosensory cortex, M1: primary motor cortex, CBT: Cognitive Behavioral Therapy group, EDU: Education control group, PRE: baseline/before-treatment, POST: Post-treatment, R: right hemisphere, L: left hemisphere.