Advanced Brain Imaging-TKA (fMRI-TKA)

Using 50 pre-operative brain networks derived from advanced brain imaging will be assessed as predictors of postoperative patient satisfaction.

Patient satisfaction measured by PROMIS questionnaire Preoperative resting-state fMRI brain connectivity will be assessed as a predictor of postoperative patient satisfaction measured using PROMIS questionnaires, which assess pain and function.

Preoperative resting-state fMRI brain connectivity will be assessed as a predictor of postoperative patient satisfaction measured using the Oxford Knee Score, which assesses knee function and pain.

50 brain networks derived from advanced brain imaging will be assessed longitudinally to characterize changes in brain networks from baseline to 12 months. Brain regions will be defined using the Schaefer functional atlas and the Allen Brain Atlas. Longitudinal changes in connectivity will be assessed using linear mixed-effects models with individual satisfaction change scores as continuous dependent variables. These models will include fixed effects for time and satisfaction, as well as random effects for subject ID to account for repeated measures. Changes will be computed by subtracting baseline from 12-month connectivity values, and these difference matrices will serve as the input features for modeling. The machine learning models for longitudinal change will follow the same 2:1 training/test split, k-fold cross-validation strategy, and performance evaluation.

Brain network nodes will be identified from rsfMRI-derived brain networks that are associated with postoperative dissatisfaction. Network nodes will be analyzed to identify potential neuroimaging biomarkers relevant to postoperative outcomes and future therapeutic targeting. Nodes will be ranked based on their contribution to predictive models and graph-theoretical importance metrics.