Parkinson disease (PD) is the second most common neurodegenerative disease, with diverse etiological risk factors, including advancing age, male, genetic factors, family history, head trauma, exposure to chemicals and medications, being demonstrated to be associated with PD development and disease progression (Belvisi et al., 2020). While it is probable that multiple factors and their mutual interactions contribute to the pathogenesis of PD, advancing age is a primary risk factor for developing PD (Collier et al., 2011).

A multitude of studies have explored the complex relationships between aging and PD. Studies have found that PD prevalence increases steadily with age, regardless of sex or geographic variations (Brakedal et al., 2022, Coleman and Martin, 2022, Kang et al., 2023, Okunoye et al., 2022, Parkinson's Foundation, P.G., 2022) Furthermore, the age at onset of PD affects not only the predominant symptoms but also disease progression (Jankovic and Kapadia, 2001). Based on the stochastic acceleration hypothesis, the cellular mechanisms for dopamine neuron degeneration in aging and PD are similar, whereby aging can be considered a pre-PD state (Collier et al., 2011). Thus, individuals who eventually present as PD have experienced an accelerated aging effect, potentially influenced by a factor or combination of factors to the point where the dopaminergic neuron loss has exceeded the threshold for PD. In addition, extensive albeit inhomogeneous brain structural changes in PD have been well demonstrated by studies in neuropathology (Tritsch and Sabatini, 2012), neuroimaging (Pan et al., 2012), and animal models (Pickrell et al., 2011). These findings, indicating the extensive and heterogeneous changes in the PD brain across different stages of disease progression, suggest that the brain degeneration associated with PD may be region-specific and not uniformly presented in the brain.

Biological age is an index integrating a combination of multiple biomarkers which may be superior than chronological age for characterizing the age-related biological changes observed in a given individual (Levine, 2013). The aging change of the structural brain is region-specific and occurs in non-linear patterns. This brain biological age can be assessed non-invasively using MRI (Franke and Gaser, 2019). Brain biological age assessment has been developed by incorporating the neuroimaging data of healthy participants within a machine-learning framework to predict individual biological age. Individual brain health may be assessed in terms of the aging trajectory by utilizing the difference between anticipated brain age and chronological age as a single quantitative metric, known as the brain age gap (BAG) (Cole and Franke, 2017, Franke and Gaser, 2019). The BAG is used to assess the relationships between clinical assessments and outcomes in both healthy and disease populations, as well as how neurological and neuropsychiatric disorders affect the natural aging process of the brain (Cole et al., 2017, Ning et al., 2020). The majority of recent brain age studies have been conducted using a single measure of global brain age while not considering the importance of focal regional brain age, which would provide more comprehensive and accurate information regarding the pathophysiology of diseases (Kaufmann et al., 2019).

The construction of multiscale brain age prediction models that include both global and region information for evaluating individual brain health in terms of aging trajectory could offer novel insights into the pathophysiology of PD. To estimate individual brain age in an unbiased manner in the present study, the multiscale brain age prediction model within a machine-learning framework was developed based on brain MRI dataset from a large-size community-based aging cohort consisting of 1240 healthy participants. Using proposed multi-level brain age prediction model, we further evaluated whether patients with PD demonstrated both global and regional-specific advanced brain aging profiles. We further explored the relationships between advanced brain aging and clinical profiles, including disease severity and cognitive functions to determine whether advanced brain aging is involved in the pathophysiology of PD.