Imaging Pain

Abstract

The challenges and understanding of acute and chronic pain have been illuminated through the advancement of central neuroimaging. Through neuroimaging research, new technology and findings have allowed us to identify and understand the neural mechanisms contributing to chronic pain. Several regions of the brain are known to be of particular importance for the maintenance and amplification of chronic pain, and this knowledge provides novel targets for future research and treatment. This article reviews neuroimaging for the study of chronic pain, and in particular, the rapidly advancing and popular research tools of structural and functional MRI.

Keywords: Brain-based therapies; Chronic pain; MRI; MVPA; Neuroimaging; Resting-state networks.

Fig. 1

Schematic of data sources and clinical applications related to identification of pain signature patterns imaging data can provide large sources of detailed and objective biomarkers for pain (top, gray box). Various sources of imaging biomarkers include (1) structural abnormalities measured with MRI (eg, DTI of white matter tractography; gray matter volumetry), (2) functional differences measured with fMRI (eg, resting-state fMRI networks and functional connectivity between brain regions; brain activity in response to evoked stimulation or during a task), and (3) functional differences measured with non-MRI modalities, such as EEG. Nonimaging data sources can provide additional objective biomarkers (middle, green box). These include, for example, genotype information, biometrics from wearable technology (eg, actigraphy), actively reported biometrics (eg, via handheld devices for recording patients’ symptoms throughout the day), psychometrics including reaction time tests and voice analysis (eg, to measure emotional states, such as depression or anxiety) and actively reported psychometrics (ie, demographic, psychological, and clinical questionnaires). The identified biomarkers from both imaging and nonimaging sources can be combined as input for an MVPA. The MVPA uses clustering, regression, and/or classification technology and is capable of identifying the most meaningful biomarkers to provide a multivariable signature pattern of pain (right, red box). The MVPA-derived pain signature can then be used in a variety of clinical applications (bottom, red box). Starting as a complex prognostic measure, the pain signature could provide the basis for stratification of an individual to a specific treatment program (“Prognosis & Stratification”). Surrogate endpoints, follow-up measurements after treatment, could be used to evaluate the treatment as effective or noneffective for the individual patient. Ultimately, the final outcome measures would indicate the pain signature’s prognostic sensitivity and specificity. Alternatively, as a more traditional clinical course of action (“Prognostic [no treatment]”), the pain signature could be used simply as a diagnostic (for clinical or legal purposes) or to predict a patient’s prognosis over time regardless of treatment. (Courtesy of Ming-Chih Kao, PhD, MD, Palo Alto, CA.)