Osteoarthritis (OA) is a leading cause of chronic pain and disability in older adults, which most commonly affects the joints of the knee, hip, and hand. To date, there are no established disease modifying interventions that can halt or reverse OA progression. Therefore, treatment is focused on alleviating pain and maintaining or improving physical and psychological function. Rehabilitation is widely recommended as first-line treatment for OA as, in many cases, it is safer and more effective than the best-established pharmacological interventions. In this article, we describe the presentation of OA pain and give an overview of its peripheral and central mechanisms. We then provide a state-of-the-art review of rehabilitation for OA pain—including self-management programs, exercise, weight loss, cognitive behavioral therapy, adjunct therapies, and the use of aids and devices. Next, we explore several promising directions for clinical practice, including novel education strategies to target unhelpful illness and treatment beliefs, methods to enhance the efficacy of exercise interventions, and innovative, brain-directed treatments. Finally, we discuss potential future research in areas, such as treatment adherence and personalized rehabilitation for OA pain.

Previous work suggests that sleep disruption can contribute to poor pain modulation. Here, we used experimental sleep disruption to examine the relationship between sleep disruption‐induced pain sensitivity and functional connectivity (FC) of cognitive networks contributing to pain modulation. Nineteen healthy individuals underwent two counterbalanced experimental sleep conditions for one night each: uninterrupted sleep versus sleep disruption. Following each condition, participants completed functional MRI including a simple motor task and a noxious thermal stimulation task. Pain ratings and stimulus temperatures from the latter task were combined to calculate a pain sensitivity change score following sleep disruption. This change score was used as a predictor of simple motor task FC changes using bilateral executive control networks (RECN, LECN) and the default mode network (DMN) masks as seed regions of interest (ROIs). Increased pain sensitivity after sleep disruption was positively associated with increased RECN FC to ROIs within the DMN and LECN (F_(4,14) = 25.28, pFDR = 0.05). However, this pain sensitivity change score did not predict FC changes using LECN and DMN masks as seeds (pFDR > 0.05). Given that only RECN FC was associated with sleep loss‐induced hyperalgesia, findings suggest that cognitive networks only partially contribute to the sleep‐pain dyad.