AI joins the fight against Alzheimer’s and Parkinson’s: NeuroArtP3 Project achievements and the future of “TreC Parkinson”
Il progetto di ricerca nazionale NeuroArtP3, nato
Launched in 2020, the NeuroArtP3 national research project aims to transform the management of central nervous system disorders by leveraging clinical data. The project has now reached two major milestones with the publication of two significant scientific studies demonstrating Trentino’s leadership in applying artificial intelligence (AI) to neurodegenerative diseases . The achievements are the result of a collaboration among ASUIT (the Trentino Integrated University Healthcare System, formerly APSS), Fondazione Bruno Kessler (FBK), and the TrentinoSalute4.0 Competence Center.
Funded by the Italian Ministry of Health and participating regional governments with a total budget of approximately €2.4 million, NeuroArtP3 brings together leading Italian healthcare and research institutions. The Trento branch of the project is led by the ASUIT Neurology Unit under the scientific coordination of Professor Bruno Giometto.
Alzheimer’s Disease: Predicting Cognitive Decline
The first study, published in the international journal Frontiers in Aging Neuroscience, focuses on Alzheimer’s disease. Researchers followed 126 patients in the early stages of the disease—including individuals with mild cognitive impairment and mild dementia—for three years. Twenty-five patients were enrolled at ASUIT Trento and 101 at IRCCS Policlinico San Martino in Genoa.
Scientists from the Data Science for Health unit at FBK’s Center for Digital Health and Wellbeing, working in collaboration with clinicians through TrentinoSalute4.0, developed machine learningin order to build advanced predictive models. The aim is to understand disease progression and identify the clinical and neuropsychological factors most strongly associated with cognitive decline. Key factors that emerged include visuospatial memory (a set of cognitive abilities that enable individuals to perceive, process, and use visual information related to space and the location of objects), long-term verbal memory, the ability to inhibit cognitive interference, the presence of hypertension, and difficulties with instrumental activities of daily living.
This is the first Italian study to use real-world clinical data, collected during routine medical practice rather than relying exclusively on large public datasets. The findings provide new insights into disease progression and support the development of personalized treatment strategies.
Parkinson’s Disease: Predicting Falls and Motor Fluctuations
The second study focuses on Parkinson’s disease, the world’s second most common neurodegenerative disorder. As populations age, the number of people living with Parkinson’s is expected to double by 2030. A multidisciplinary team coordinated by ASUIT, together with FBK, IRCCS Policlinico San Martino, and the University of Genoa, developed an AI-based algorithm capable of predicting patients’ risk of falls and motor fluctuations (the so-called “on-off” episodes).
The project began with the digitization and standardization of clinical data and consisted of two phases:
- Retrospective phase: historical patient records were standardized to characterize patients’ clinical phenotypes and correlate them with subsequent disease progression.
- Prospective phase: the same clinical variables were collected and validated in newly diagnosed patients enrolled at participating clinical centers.
“The relationship between patients’ clinical phenotype and symptom progression—and the early identification of specific symptoms—is critical for risk prediction, personalized treatment, and the planning of effective preventive strategies,” said Lorenzo Gios, Project Manager at TrentinoSalute4.0. Maria Chiara Malaguti, Director of the Neurology Unit in Rovereto, Medical Director at Santa Chiara Hospital in Trento, and Coordinator of ASUIT’s Parkinson’s Clinical Network, added: “Once fully developed and validated, these models will significantly advance our understanding of the disease. For clinicians, they represent an opportunity to manage chronic conditions more effectively by helping us better understand our patients and tailor treatments to their individual needs.”
From Research to Clinical Practice: TreC Parkinson function
The findings extend beyond scientific publications and are already being translated into digital tools with the potential to improve people’s daily lives.
At the end of 2025, the project entered a pilot testing phase. Healthcare professionals began evaluating a clinician dashboard, while patients recruited through the Parkinson’s Association started testing a companion mobile app.
This work is being carried out through a collaboration among ASUIT, Dedagroup, the Parkinson Trento Association, the Artigianelli Institute of Trento, and FBK.Consistent with the vision of TreC+, Trentino’s digital health platform, the team is developing advanced features and educational tools that will be integrated into telemonitoring and telemedicine services for patients and caregivers.
These developments represent an important step in translating AI-based predictive models from research into everyday clinical practice, supporting more personalized care and continuous patient monitoring.
Note: Clinical phenotype refers to the observable clinical characteristics of an individual, resulting from the interaction between genetic factors and environmental influences.
Related articles
- (January 2025): Parkinson’s disease: an algorithm detects falls and motor alterations
- (December 2025) Primed-PD 2025 project to combat Parkinson’s disease approved
- (November 2020) Big data and diseases of the nervous system: APSS and FBK in the NEUROARTP3 project
Share on Facebook Share on Twitter Share on Pinterest
