George Mason University, USA

Title: From Neuron Classification to Spiking Neural Network Simulations: A Neuroinformatics Approach to Data-Driven Computational Models

Abstract: Neuroscience textbooks describe the brain as a massive network of spiking neurons with complex dynamics. Neural circuits, the story goes, are largely defined by the connectivity between axons, reliably propagating all-or-none action potentials up to very distant targets, and dendrites, integrating synaptic inputs into local arbors. In such conceptual model, short- and long-term activity-dependent plasticity continuously alters synaptic strength, but a fixed neurotransmitter identity determines signal type (e.g., excitatory vs. inhibitory). Are these ingredients sufficient to explain the emergence of cognition? If additional details are required, which ones exactly? This keynote aims to illustrate a strategy to tackle such fundamental questions for the hippocampal formation, which plays a core role in episodic memory and spatial navigation. I will demonstrate that a systematic recognition of neuronal diversity is the key to connect the dots of disparate experimental data into a coherent quantitative framework to explore the relationship between synapses and behavior.

University of Oxford, UK and Aarhus University, Denmark

Title: Whole-brain Modelling: Cartography of Eudaimonia and Flourishing In the Human Brain

Abstract: In order to survive, the brain must constantly extract, predict and recognise the essential spacetime features of complex environments. This distributed computation of information relies on having a hierarchy of optimal information transfer across the whole brain at the lowest possible metabolic cost. Suboptimal brain orchestration has been linked to mental illness, yet the fundamental principles of brain orchestration over fast and slow timescales are still not well understood. I will show how significant progress has been made using whole-brain modelling of neuroimaging data using new frameworks based on stochastic thermodynamics and turbulence. A series of studies have already furthered our understanding of human flourishing using data from experiments including music, food, social interactions, meditation and psychedelics. Overall, this new evidence has given rise to a deeper understanding of experiences that can give rise to both flourishing and suffering, providing meaning and purpose to life, and may eventually help to find novel ways to rebalance the brain in neuropsychiatric disorders.

The University of Bari Aldo Moro, Italy

Title: The Human Brain Wasn’t Built in a Day: The Building Blocks of Psychiatric Risk Trajectories

Abstract: Psychiatric treatments continue to focus on symptom categories instead of the biological mechanisms that produce them. Progress toward more effective interventions depends on clarifying how mental disorders develop and how genetic risk interacts with biological and environmental influences. The current research program investigates genetic risks associated with highly heritable disorders such as schizophrenia and how these risks unfold across the lifespan. We examine how genes, environmental influences, and developmental timing (G×E×T) interact to shape brain function and behavior. The central hypothesis is that genetic risk affects specific biological pathways, which determine how a disorder manifests, progresses, and responds to treatment. We test this hypothesis by integrating large-scale, multimodal data from over 100,000 individuals. Using genome-wide association studies, polygenic scores, and pathway-specific genetic indices, we link genetic variation to biological mechanisms through analyses that used brain imaging, post-mortem transcriptomics, and induced pluripotent stem cells. Our findings indicate that biological markers of psychiatric risk emerge long before clinical symptoms. Signs of genetic risk for schizophrenia can already be seen in early childhood, influencing cognition even before measurable brain changes. In adolescence, these genetic influences are associated with brain connectivity during a period especially sensitive to environmental effects. Genetic vulnerability is also linked to personality traits that increase the likelihood of being bullied, contributing to early psychotic-like experiences. At the neural level, this pathway involves dopamine-related genetic variation associated with altered striatal function. These findings highlight how genetic risk interacts with development and experience, offering new targets for interventions that foster resilience.

Brain And SIgnal Research and Analysis (BASIRA) Laboratory, UK

Title: Brains and AI: A Two-Way Journey

Abstract: Brains and artificial intelligence (AI) are converging through a two-way exchange: network neuroscience informs new learning paradigms, while AI models increasingly help decode and generate brain connectivity. Graph-based learning provides a shared formalism for this convergence, capturing both neural and computational architectures within a unified framework. This keynote will highlight recent conceptual leaps in graph neural networks (GNNs)—progressing from unification to reasoning and ultimately to cognification, where cognitive principles are directly embedded into GNN learning. Together, these developments outline a path toward interpretable, robust, and brain-aligned AI that simultaneously advances network neuroscience. By integrating principled graph learning, cognition-aware generative modeling, and unified training across sensory and cognitive domains, we move closer to AI systems that mirror the organization and function of the human brain.

University College Cork, Ireland

Title: Explainable Artificial Intelligence for EEG Research and Scientific Discovery

Abstract: Electroencephalography (EEG) is a core method for studying and understanding brain activity and processes. Traditional methods for EEG data analysis have been complemented with methods from Artificial Intelligence (AI). These especially include machine learning and deep learning methods that have been exploited for various tasks, including supervised disease classification, EEG-neural conditions mapping, unsupervised mental states clustering or fully automated artefact detection and mitigation. They have demonstrated satisfactory capability to model the intrinsic non-linearities and non-stationarities of EEG signals, within and across subjects. However, while AI-based approaches are now widely recognised and used to analyse and interpret EEG signals, unfortunately, they lead to models that often lack transparency and interpretability, thus limiting scientific discovery. This keynote will shed light on the application of eXplainable Artificial Intelligence (XAI) and its novel methods for constructing solutions to the above problem. Specifically, it will delve, for example, into approaches for interpreting the inner inferential mechanisms of the complex models learnt via deep learning or the identification of higher-level features that are responsible for certain outputs. Similarly, it will showcase how scientific discovery could be performed with EEG data and how recurrent abstract representations of brain activity within and across subjects can be extracted.