Research interests

With Antonio Celani (ICTP, Italy), Daniel M. Busiello (MPIPKS, Germany) and many other collaborators, I am currently working on understanding how biological and artificial agents process information and make decisions in complex and stochastic environments. We are seeking universal principles that can be applied across different scales, from biochemical to neural networks, exploiting ideas from Information Theory and tools from machine learning.

Our work so far has dealt with foundational questions. We have studied how stochastic systems can optimize information harvesting from hidden signals, showing the existence of energy-efficient transduction mechanisms and how our results can be applied to understand the links between flickering and mechanical stress in red blood cells.
We have also investigated how the existence of a spectrum of dynamical timescales shape how information is propagated between the components of multiscale systems, and how it can be used to define functional relationships between different layers of a network.
Finally, we have introduced the idea of information interference in systems coupled with unpredictable and changing environments. We have studied the conditions upon which the information between internal degrees of freedom can be disentangled from the one due to the external environment, and how this can be used to define the concept of effective interactions in complex systems. In general, however, we have shown how information interference emerges from environmental changes, making the system intrinsically context-dependent.

With Amos Maritan (University of Padova), Andrea Rinaldo (EPFL) and Prajwal Padmanabha (University of Lausanne) I am studying spatial metapopulation models through the lenses of Statistical Physics. By studying ecological systems in arbitrary network topologies, we seek to understand how dispersal affects survival, coexistence, and dynamics of ecological metacommunities.

Our work focuses on how spatial features can drastically change the fate of ecological communities. We have shown that landscape heterogeneity can support the coexistence of competing species by fostering the spontaneous formation of ecological niches. Similarly, the connectivity of dispersal network can affect the stability of the system, either helping them survive or pushing them to the brink of extinction. We use methods from the Statistical Physics of disordered systems, allowing us to study the emergence of macroscopic ecological patterns starting from simple microscopic, individual-based rules.

With Samir Suweis (University of Padova) and several experimental collaborators, I have been studying since the beginning of my PhD the emergence of critical signatures in neural activity and how tools from Statistical Physics can be used to understand the dynamics of the brain.

Our work has focused on understanding the emergence of spatiotemporal patterns in brain dynamics, such as the formation of scale-free neural avalanches and long-range temporal and spatial correlations. We have highlighted how slow drivers of neural activity are necessary for the onset of neural avalanches, and how scale-free neural correlations emerge from LFP recordings in the brain of rodents. We have recovered similar results in whole-brain models on the human connectome, showing how brain-wide oscillations are tightly linked to phase transitions. Finally, we have shown that strong temporal correlations in EEG recordings of newborns are linked to the prenatal experience with language, suggesting that the brain is already shaped by the maternal language before birth.