Second Computational Psychiatry Day in Parma

On November 28–29, 2024, the city of Parma hosted the Second Italian Day of Computational Psychiatry, confirming the growth of the Italian Network for Computational and Precision Psychiatry. Building on the success of the inaugural meeting in Ferrara, this second edition consolidated the community and expanded the international dimension, with contributions from leading Italian and European scholars.

The symposium opened on Thursday afternoon with a welcome by Paolo Ossola (University of Parma), who highlighted the importance of computational psychiatry in clarifying treatment mechanisms. Rodolfo Rossi (University of Rome Tor Vergata) followed with an overview of Bayesian belief updating in psychosis, demonstrating how probabilistic modeling can reveal altered inference processes. After the first break, Chiara Montemitro (Linköping University and NIH) presented the potential of reinforcement learning paradigms to guide treatment in substance use disorders. Erdem Pulcu (University of Oxford) then discussed treatment-resistant depression, focusing on the role of learning, memory, and underlying brain mechanisms. The session concluded with Paolo Ossola presenting Drift Diffusion Models and their relation to prefrontal cortex function in affective disorders.

Friday’s program was dedicated to data-driven approaches. Martino Belvederi Murri (University of Ferrara) opened with a presentation on Item Response Theory and how computational methods can improve the use of clinical questionnaires. Linda Antonucci and Alessandra Raio (University of Bari Aldo Moro) discussed applications of machine learning in psychotic disorders. Alessandro Pigoni (IRCCS Ca’ Granda Policlinico, Milan) compared supervised and unsupervised approaches in psychotic and affective disorders, while Filippo Corponi (University of Edinburgh) explored the role of neural networks and self-supervised learning in bipolar disorder. The final session, dedicated to how data-driven approaches can reshape psychiatric treatments, prepared the ground for a collective discussion and concluding panel.

The Parma meeting highlighted the dual trajectory of computational psychiatry: theory-driven models, such as Bayesian inference and reinforcement learning, that offer mechanistic explanations of symptoms, and data-driven approaches, such as machine learning and network modeling, that extract clinically relevant patterns from complex datasets. Together, these perspectives illustrated the potential of computational psychiatry to refine diagnosis, predict outcomes, and tailor interventions.

Importantly, the event also emphasized the humanistic and clinical grounding of this field. By integrating advanced quantitative methods with clinical insight, computational psychiatry in Italy is shaping itself as a discipline that remains attentive to patient experience and the challenges of everyday practice.

The Second Italian Day of Computational Psychiatry thus consolidated a vibrant community of clinicians, neuroscientists, and data scientists. It strengthened international collaborations, provided visibility to young investigators, and reaffirmed the commitment to building a national and international network. The Parma symposium demonstrated that computational psychiatry is no longer a peripheral interest but an essential tool for the future of psychiatric research and care.