A transdiagnostic approach to non-motor symptoms in Parkinson’s disease: exploring internalizing, bipolar spectrum, and impulsive compulsive behavior profiles via cluster analysis
Parkinson’s disease (PD) is associated with diverse non-motor symptoms (NMS), including depression, anxiety, anhedonia, bipolar-spectrum features, and impulsive-compulsive behaviors (ICBs), which substantially impair quality of life. Although these domains may share dopaminergic and frontostriatal mechanisms, they are often studied separately. This study adopted a transdiagnostic, person-centered approach to identify clinically meaningful NMS profiles in PD. In this cross-sectional study, 126 patients with PD and no dementia diagnosis from two university centers completed validated assessments of ICBs, impulsivity, anhedonia, depression, anxiety, and bipolar-spectrum symptoms. K-means clustering was applied to standardized variables, and cluster validity was evaluated using multiple stability metrics. The four-cluster solution demonstrated the best validity and identified Low-Symptom (44.3%), Moderate Internalizing (29.4%), Bipolar–Impulsive (16.7%), and Severe Internalizing–Impulsive (9.5%) profiles. Clusters differed significantly across depression, anxiety, anhedonia, bipolar-spectrum symptoms, ICBs, and impulsivity. The Bipolar–Impulsive group was significantly younger, whereas the Moderate Internalizing profile was older and predominantly female. The Severe Internalizing–Impulsive cluster showed the highest psychiatric burden, motor severity, disease stage, dopaminergic treatment load, and amantadine use. No significant between-cluster differences emerged in global cognitive performance. These findings support a transdiagnostic model in PD in which affective dysregulation and impulsivity interact with disease-related factors to form distinct neuropsychiatric phenotypes. Recognizing these profiles may facilitate earlier identification of vulnerable patients and more individualized neuropsychiatric management strategies beyond a solely dopaminergic framework.