Objective The neural mechanisms of pain in knee osteoarthritis ( OA ) are not fully understood, and some patients have neuropathic‐like pain associated with central sensitization. To address this, we undertook the present study in order to identify central sensitization using neuroimaging and Pain DETECT and to relate it to postarthroplasty outcome. Methods Patients awaiting arthroplasty underwent quantitative sensory testing, psychological assessment, and functional magnetic resonance imaging ( fMRI ). Neuroimaging ( fMRI ) was conducted during punctate stimulation (n = 24) and cold stimulation (n = 20) to the affected knee. The postoperative outcome was measured using the Oxford Knee Score, patient‐reported moderate‐to‐severe long‐term pain postarthroplasty, and a range of pain‐related questionnaires. Results Patients with neuropathic‐like pain presurgery (identified using Pain DETECT ; n = 14) reported significantly higher pain in response to punctate stimuli and cold stimuli near the affected joint ( P < 0.05). Neural activity in these patients, compared to those without neuropathic‐like pain, was significantly lower in the rostral anterior cingulate cortex ( P < 0.05) and higher in the rostral ventromedial medulla ( RVM ) during punctate stimulation ( P < 0.05), with significant functional connectivity between these two areas (r = 0.49, P = 0.018). Preoperative neuropathic‐like pain and higher neural activity in the RVM were associated with moderate‐to‐severe long‐term pain after arthroplasty ( P = 0.0356). Conclusion The psychophysical and neuroimaging data suggest that a subset of OA patients have centrally mediated pain sensitization. This was likely due to supraspinally mediated reductions in inhibition and increases in facilitation of nociceptive signaling, and was associated with a worse outcome following arthroplasty. The neurobiologic confirmation of central sensitization in patients with features of neuropathic pain, identified using PainDETECT, provides further support for the investigation of such bedside measures for patient stratification, to better predict postsurgical outcomes.