Editorial: Advancing personalized diagnosis and treatment in Parkinson's disease: integrating biomarkers, neuroimaging, and artificial intelligence

Editorial on the Research Topic
Advancing personalized diagnosis and treatment in Parkinson's disease: integrating biomarkers, neuroimaging, and artificial intelligence

Introduction

Once viewed primarily as a late-onset movement disorder, Parkinson's disease (PD) is increasingly recognized as a multisystem syndrome with motor and non-motor symptoms arising from diverse genetic, molecular, neurological, and environmental factors.

With the number of individuals affected more than doubled over the past three decades (Global Burden of Disease 2021) and the dramatic growing prevalence of early-onset forms (Li et al.), researchers and clinicians are exploring how to integrate epidemiological insights, lifestyle determinants, and emerging biomarkers to identify PD pathophysiology and build personalized approaches to early diagnosis and treatment.

This Research Topic, Advancing personalized diagnosis and treatment in Parkinson's disease: integrating biomarkers, neuroimaging, and artificial intelligence, brings together original studies and reviews that showcase innovative methodologies and clinical insights pointing the way toward integrated patient-specific care for PD (Table 1).

Table 1

Table 1. Summary of the study design, hypothesis and results of the studies included in the Research Topic.

Neuroimaging and electrophysiological insights into Parkinson's disease

One of the biggest challenges in moving toward personalized care in PD lies in its clinical heterogeneity. Patients differ not only in the motor symptoms they display but also in the occurrence of non-motor symptoms, such as cognitive and mood disturbances. Neuroimaging approaches have been providing a deeper understanding of the neural basis of PD and several contributions to this Research Topic used such techniques to explore PD heterogeneity.

Using MRI, Song et al. conducted the largest meta-analysis to date of regional homogeneity (ReHo) alterations, combining 72 datasets from over 2,000 patients with PD and 1,400 controls. This study revealed a distributed dysfunctional network, involving the visual, somatomotor, dorsal and ventral attention networks, confirming that PD pathology extends beyond dopaminergic circuits to large-scale networks.

Another study (Xu T. et al.) investigated hemispheric lateralization, a hallmark of asymmetric dopaminergic degeneration associated with distinct symptom profiles (Voruz et al., 2025). Compared to right-onset patients, left onset patients with PD (LPD) showed reduced cortical area in the right supramarginal gyrus, right precuneus, left inferior parietal lobule, and left lingual gyrus. In LPD, the right precuneus area positively correlated with MMSE cognitive scores, consistent with previous reports (Syrimi et al., 2017; Lee et al., 2015). At the network level, LPD patients exhibited altered topological organization, with increased path length and reduced small-world index, indicative of reduced efficiency.

Building on this network-level approach, another study examined different stages of PD using resting-state fMRI and graph theory analysis (Wei et al.). Both early- and middle-to-late-stage patients with PD showed reduced clustering coefficients, indicating decreased local networks' specialization. More advanced patients showed an overall decline in network efficiency, both locally and across the whole brain. Importantly, centrality within the left middle frontal gyrus and right middle temporal pole correlated with clinical measures of motor severity and disease stage.

Going beyond PD staging and subtyping, another critical challenge lies in distinguishing idiopathic PD from secondary conditions, such as drug-induced Parkinsonism (DIP). Zhou et al. found greater reduction in bilateral whole hippocampal volume and subfields atrophy in DIP patients compared to patients with PD and controls, which correlated with cognitive deficits, depressive and motor symptoms. However, DIP patients' lower MoCA cognitive scores may partially explain these volumetric differences, emphasizing the need to control for baseline cognitive functioning in future studies.

Non-invasive electrophysiological (EEG) studies added further insights. Zhao Y. et al. showed that cognitively impaired patients with PD displayed lower EEG alpha (8–13 Hz) power in parieto-occipital and posterior temporal regions compared to cognitively intact patients, which correlated positively with MoCA score and best differentiated PD patients with and without cognitive impairment.

These findings align with longitudinal studies showing that alpha slowing predicts progression to PD-dementia (Klassen et al., 2011; Olde Dubbelink et al., 2013), reinforcing its value as a biomarker. Importantly, advances in wearable EEG technology and artificial intelligence now make it feasible to translate these markers into home-based monitoring systems, enabling continuous and personalized assessment of the patients' brain activity and cognition outside the clinic (Sigcha et al., 2023).

Overall, these contributions highlight how neuroimaging techniques can capture several dimensions of PD heterogeneity, from volumetric changes to network dysfunction, lateralization, and EEG slowing. By linking structural and functional alterations to specific symptoms, these methods move beyond simple PD subtyping to identify patients at risk of rapid progression or cognitive decline, guide treatments, provide tools for monitoring disease trajectories, and enrich clinical trials with biologically defined subgroups, thereby accelerating the translation of precision medicine approaches into practice.

Genetic, molecular and cellular biomarkers

Complementing neuroimaging and electrophysiological evidence, other contributions in this Research Topic highlight how genetic, fluid, and metabolomic biomarkers can refine PD diagnosis, staging, and phenotyping, and how they can be combined or mapped onto brain networks for improved personalized care.

A single-center longitudinal study (Xu H.-L. et al.) found that, compared to GG homozygotes, BST1 rs4698412 A-allele carriers had faster motor, but not cognitive, decline, mainly driven by rigidity and bradykinesia. This highlights a common PD-risk variant as a potential progression biomarker, useful for refining prognosis and optimizing trials for accelerated motor decline.

In the WPBLC cohort, He and colleagues found that elevated plasma Parkin in patients with PD had moderate diagnostic accuracy. Combining Parkin with homocysteine, total protein, and urea improved discrimination. Parkin correlated with blood α-synuclein oligomers and phosphorylated α-synuclein, but not with motor severity. Mediation analyses suggested partial effects via albumin and carcinoembryonic antigen, and transcriptomics pointed to PINK1-PRKN mitophagy and related metabolic pathways. Together, these data support a minimally invasive multi-analyte blood panel for PD (He et al.).

Another study integrated plasma neurofilament light (NfL) with cortical morphometry and connectivity and observed that patients with PD with excessive daytime sleepiness (EDS) show higher NfL, focal parietal thinning (left supramarginal gyrus; right postcentral gyrus), and reduced parietal-frontal coupling. NfL partially mediated the link between supramarginal thickness and EDS severity. These results support combined neuroimaging and plasma NfL biomarkers to clarify EDS mechanisms and track non-motor progression (Chen J. et al.).

Finally, blood metabolomics identified seven molecules distinguishing PD from controls and a distinct three-metabolite pattern specific to PD with REM sleep behavior disorder (PD-RBD). Pathway enrichment indicated disruption of central carbon metabolism in PD and inactivation of PPAR signaling in PD-RBD, supporting these metabolites as candidate (Chen H. et al.).

Epidemiology, digital, and clinical tools

Beyond neuroimaging and molecular markers, population-based studies have also identified systemic biomarkers of PD. In a cross-sectional analysis of NHANES 2005–2020 data, Zhao Y. et al. found that higher systemic immune-inflammation index (SII) values, derived from routine blood counts, were linked to greater PD prevalence, particularly at higher SII levels. These findings highlight PD's multisystem nature and support immune-inflammatory markers as useful indicators.

Similarly, Zhou and Cheng reported that cardiovascular health, reflected by higher Life's Essential 8 (LE8), correlated with lower PD risk, suggesting that maintaining cardiometabolic health may help reducing disease risk.

Digital and AI-based tools are also advancing PD diagnostics. In a cross-sectional study, a digital clock drawing test (dCDT) was implemented to differentiate PD patients with mild cognitive impairment (MCI) from Alzheimer's disease (AD) (Wang et al.). Combined metrics separated AD-MCI from PD-MCI with high accuracy and the overall drawing score correlated with the MoCA visuospatial/executive subtest score, supporting dCDT's value in cognitive assessment.

Eye-tracking, increasingly integrated with machine learning (ML) and virtual reality (VR), is further refining PD phenotyping. In their systematic review, Culicetto et al. reported that oculomotor metrics, such as saccade velocity, fixation duration, and pupil size, correlated with disease severity, and that integrating these metrics with ML/VR pipelines improve diagnostic accuracy and scalability. Together, these advances position eye-tracking as a promising biomarker platform for motor and cognitive dysfunction in PD, though standardization across devices and protocols remains essential before clinical adoption.

Finally, Twala reviewed AI applications in PD diagnosis and treatment, reporting accuracies spanning 78–96% across modalities, with neuroimaging leading on mean accuracy and multimodal systems offering the best generalizability. Building on this, the study illustrates a novel multimodal AI framework that achieves 94.2% overall accuracy and strong early-stage PD detection and outperforms traditional clinical assessment methods (Twala). So far, validation of this model used a simulated PD cohort, hence real-world, multi-site studies are needed before clinical use.

Interventions

Deeper insights into PD risk factors and progression paired with biomarker-guided patient stratification are paving the way for effective personalized therapies. This Research Topic illustrates also this bench-to-bedside trajectory through two complementary studies: a flavonoid-based neuroprotective approach in a mouse model (Huang et al.) and a dual-target deep brain stimulation in a patient with asymmetric motor features (Deng et al.).

Huang et al. demonstrated that galangin mitigated MPTP-induced dopaminergic neurodegeneration in mice, restoring striatal dopamine levels, reducing neuroinflammatory cytokines and α-synuclein accumulation, increasing antioxidant enzymes and improving motor performance. Mechanistically, galangin activated PI3K/AKT/CREB-BDNF signaling, inhibited Beclin-1-dependent autophagy, and preserved TH-positive neurons, indicative of its disease-modifying efficacy in preclinical models (Huang et al.).

Taking another innovative approach, Deng et al. reported the first documented case of combining posterior subthalamic area (PSA) and globus pallidus internus (GPi) deep brain stimulation (DBS) in a single PD patient with pronounced left-right asymmetry. The PSA is a tremor-responsive target, whereas the GPi is favored for treating rigidity and dyskinesia. A left PSA lead suppressed right-sided tremor, and a right GPi lead treated left-sided rigidity and dyskinesia. After six months of dual-target DBS, motor status improved and tremor frequency decreased, while cognition remained intact with no significant adverse effects. As a single case with short follow-up, these findings support feasibility of symptom specific dual-target neuromodulation for individualized therapy, although larger controlled studies are needed to establish effectiveness (Deng et al.).

Conclusions

As PD emerges as a multifaceted disorder extending beyond dopaminergic degeneration, progress in care will depend on closer integration of technology and neurobiologically informed clinical practice. Collectively, the studies in this Research Topic show that combining biomarkers, advanced neuroimaging, and AI-driven analytics enables earlier diagnosis, personalized treatment, and improved monitoring of disease progression, indicating an ongoing shift toward precision care.

Author contributions

AG: Writing – review & editing, Writing – original draft. SS: Writing – review & editing, Writing – original draft. ET: Writing – original draft, Writing – review & editing.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. Dr. Elisa Tatti acknowledges Dr. Lice Ghilardi and the Department of Defense (grant W81XWH-19-1-0810) and National Institute of Health (grant U54MD017979) for supporting her research.

Acknowledgments

We thank Dr. Alberto Cacciola for his valuable contribution to this Research Topic. The authors are also thankful to the contributors to this Research Topic, the reviewers, and the Editorial support of the Journal.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.

Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Klassen, B. T., Hentz, J. G., Shill, H. A., Driver-Dunckley, E., Evidente, V. G. H., Sabbagh, M. N., et al. (2011). Quantitative EEG as a predictive biomarker for Parkinson disease dementia. Neurology 77, 118–124. doi: 10.1212/WNL.0b013e318224af8d

PubMed Abstract | Crossref Full Text | Google Scholar

Lee, E. Y., Sen, S., Eslinger, P. J., Wagner, D., Kong, L., Lewis, M. M., et al. (2015). Side of motor onset is associated with hemisphere-specific memory decline and lateralized gray matter loss in Parkinson's disease. Parkinsonism Relat. Disord. 21, 465–470. doi: 10.1016/j.parkreldis.2015.02.008

PubMed Abstract | Crossref Full Text | Google Scholar

Olde Dubbelink, K. T. E., Stoffers, D., Deijen, J. B., Twisk, J. W. R., Stam, C. J., and Berendse, H. W. (2013). Cognitive decline in Parkinson's disease is associated with slowing of resting-state brain activity: a longitudinal study. Neurobiol. Aging 34, 408–418. doi: 10.1016/j.neurobiolaging.2012.02.029

PubMed Abstract | Crossref Full Text | Google Scholar

Sigcha, L., Borzì, L., Amato, F., Rechichi, I., Ramos-Romero, C., Cárdenas, A., et al. (2023). Deep learning and wearable sensors for the diagnosis and monitoring of Parkinson's disease: a systematic review. Expert Syst. Appl. 229:120541. doi: 10.1016/j.eswa.2023.120541

Crossref Full Text | Google Scholar

Syrimi, Z. J., Vojtisek, L., Eliasova, I., Viskova, J., Svatkova, A., Vanicek, J., et al. (2017). Arterial spin labelling detects posterior cortical hypoperfusion in non-demented patients with Parkinson's disease. J. Neural Transm. 124, 551–557. doi: 10.1007/s00702-017-1703-1

PubMed Abstract | Crossref Full Text | Google Scholar

Voruz, P., Guérin, D., and Péron, J. A. (2025). Impact of motor symptom asymmetry on non-motor outcomes in Parkinson's disease: a systematic review. Npj Park Dis. 11:188. doi: 10.1038/s41531-025-01046-4

PubMed Abstract | Crossref Full Text | Google Scholar