The role of catastrophizing in anxiety and depression symptom scores among knee osteoarthritis patients: a multimodal assessment with EEG and machine learning.

To examine dimensional associations between anxiety- and depression-related symptom severity, pain catastrophizing, and resting-state EEG features in patients with knee osteoarthritis (KOA). Resting-state EEG spectral power (delta, theta, alpha, beta) was analysed in 62 KOA patients from the DEFINE cohort. Emotional symptoms were assessed with the Hospital Anxiety and Depression Scale (HADS), along with the Pain Catastrophizing Scale (PCS) and clinical-demographic variables. Multivariate regression analyses identified significant predictors, while linear and tree-based machine learning models were used post hoc to explore whether multivariate and non-linear approaches converged with the regression findings. Catastrophizing was independently and significantly associated with both anxiety and depression symptom scores across regression and exploratory machine learning models. For depression, a multifactorial pattern was additionally observed: higher bilateral parietal delta power, greater catastrophizing, lower education, and greater body weight showed independent associations with more severe symptom scores. Machine learning analyses indicated that EEG features were weak standalone correlates but showed modest complementary associations when combined with clinical variables. Pain catastrophizing was consistently associated with both anxiety and depression symptom scores, and resting-state EEG features showed limited but complementary associations with depressive symptom scores in KOA. Importantly, these associations were observed within a sample presenting predominantly subclinical HADS scores, and findings should be interpreted as reflecting dimensional associations within a rehabilitation cohort rather than clinical anxiety or depressive disorder.