Title: Platform and digital tests for remote screening of children with learning disability

Abstract
A televisit platform based on Web applications will be presented, which allows the clinician to connect with the child in two different places and allows them to interact. The platform allows audio-video communication between clinician and child mediated by SW Jitsi. It also allows the clinician to view the current test and perform real-time scoring and the child to carry out the test by providing the associated visual material. At the end, a digital report is provided, which determines the score in the test for the child's class. Some tests for the screening of dyslexia, dysgraphia and dyscalculia will be presented.televi

Title: Playing Together: Tabletop Interaction with NAO Robots in Autism Therapy

Abstract
This work presents a humanoid robot platform designed to support tabletop play interaction with children with Autism Spectrum Condition (ASC). Using a NAO robot, the system enables turn-taking and object-mediated play through reaching, grasping, and manipulation of physical blocks during therapeutic activities. The presentation will discuss the co-design process with clinicians and psychologists, the development of robot behaviors for dyadic interaction, and the integration of the platform within real therapeutic settings, including a pilot study conducted with children with ASC in rehabilitation centers. The contribution reflects on the opportunities and challenges of deploying humanoid robots in therapeutic contexts.

Title: Personality Is Not a Prompt: Toward Psychologically Grounded LLM Simulations of Personality

Abstract
Large Language Models (LLM) can convincingly imitate conversational styles, roles, and short persona descriptions. However, sustaining a coherent personality over time remains a more complex challenge: simulated agents may contradict their own autobiographical facts, shift preferences, flatten emotionally, or drift under conversational pressure. This talk frames personality simulation as an interdisciplinary problem at the intersection of psychology and artificial intelligence. Rather than treating personality as a static prompt, it proposes a modular view in which personality is represented as a persistent and inspectable state composed of traits, values, goals, affective dynamics, interpersonal style, autobiographical memory, and narrative identity. The presentation will outline the core design principles of this approach, discuss key evaluation challenges, and raise open questions concerning memory, emotional continuity, role integrity, robustness, and ethical interpretation. The goal is to stimulate discussion on how psychologically grounded models could make LLM-based personality simulations more coherent, testable, and useful for research.

Title: Neurotechnologies facing mental health - merging technical, ethical, and human rights-based perspectives

Abstract
Neurotechnologies comprise a wide range of devices. Since 2018, direct-to-consumer (DTC) companies have overtaken the medical ones, giving to consumer grade users a plethora of solutions in the grey zone of wellness, wellbeing, and mental wellbeing.
In December 2025, the U.S. Food and Drug Administration (FDA) approved the world’s first at-home brain stimulation treatment for depression. While brain stimulation therapies have demonstrated their efficacy especially for medication resistant mental disorders, the mass distribution of such technologies give rise to numerous technical, ethical, and legal concerns.
In this talk, we will mainly dive in the concepts of neurosecurity and neurorights, giving some concrete examples on how these disruptive technologies can be exploited by malicious actors, undermining vulnerable users’ autonomy, mental privacy, and informational self-determination.

Title: Fractal geometry and machine learning applications to MRI in psychiatry

Abstract
Fractal geometry offers a powerful framework for quantifying the structural complexity of the brain, capturing self-similar patterns in grey matter morphology that traditional linear measures may overlook. The complexity of brain structure, in particular of grey matter, can be evaluated as a marker of disease in patients with schizophrenia and bipolar disorder, or to stratify individuals according to symptoms’ severity. This approach, combined with machine learning methods, opens promising avenues for automated classification and biomarker identification in psychiatry. Machine learning methods are indeed gaining increasing popularity in the medical field, representing a considerable step towards personalized psychiatry. Fractal analysis has shown high sensitivity in discriminating patients from healthy controls across diverse physiological and behavioral measures, though the heterogeneity of variables examined calls for further studies in this promising line of research. Together, these approaches suggest that the integration of fractal-based neuroimaging features and machine learning algorithms may represent a valuable tool for advancing precision psychiatry.

Title: Learning from Brain Signals: AI Approaches to EEG Analysis in Neurodegenerative Diseases

Abstract
Understanding the neural signatures of neurodegenerative diseases such as Alzheimer's from EEG data requires more than accurate classifiers, requiring models that generalize across clinical settings and whose predictions can be meaningfully interpreted. This talk presents an introduction to AI-based spatio-temporal EEG analysis, with a focus on clinical pipelines developed to handle real-world data heterogeneity. Building on this, explainability and neurosymbolic reasoning are introduced as a more principled route to identifying robust EEG biomarkers across heterogeneous settings. Together, these contributions outline a way forward for AI systems that are not only high-performing, but also clinically trustworthy and suitable for longitudinal monitoring scenarios.