Mario Lavanga

I am currently working as a postdoctoral researcher at the INS focusing on functional connectivity in the contest of healthy ageing. I am part of the theoretical neuroscience group and I work as part of a team dedicated to the resting state brain. I spend most of my time working on Python applications in Jupyter-lab to compute a variety of biomarkers associated to ageing and reproduce with biophysical models. My team and I are investigating functional connectivity dynamics and static functional connectivity in both the context of pathological and healthy ageing. In parallel, we are deploying a variety of computational strategies to tune our models and reproduce the empirical patterns by means of supercomputer facilities.

Thanks to my doctoral studies, I can bring my perspective in the empirical data analysis, especially related to ageing. On top of that, I provide specific views to enhance the comprehension of our team’s results to the clinical side of the project. I enjoy learning Python and Jupyter as tools for scientific computing, as well as the schedule manager such as Slurm or parallel computing tools such as SnakeMake.

During my years as a SB PhD fellow, I developed a model to quantify stress and pain in premature infants and to understand the physiological processes behind the infants’ development. The first part of my research focused on the assessment of development inside the neonatal intesive care unit by means of physiological signal processing. I investigated the scalp-EEG connectivity, the nonlinear properties of the neonatal EEG signal and the fractality of the heart-rate varaibility of the premature infant in the context of NICU development. The quantification of maturation trends by data processing is usually known as brain-age models. The codes associated to each of these studies were developed in Matlab.

The second part of my PhD was focused on the development of models and metrics to quantify pain and stress in premature infants. Based on the investigation of the development processes in the first of my PhD, I combined the information associated to the nonlinear properties of EEG and the heart-rate variability as well as the interaction among EEG signals and between EEG and heart-rate to detect stress or pain in premature infants. These approaches focused on the resting-state activity as well as the specific phenomenon of heart-rate drops, known as bradycardias. The outputs of these analysis were combined in an investigation effort to understand if the accumulation of pain in the intensive care units might lead any difference in maturation trajectories. The results showed greater discontinuity in EEGs were associated to a greater accumulation of pain by preamture infants. The codes associated to each of these studies were developed in Matlab.

My first year as PhD researcher was fully dedicated to the full comprehension of the specific domains of neonatal physiology and development as well as machine learning and graph theory. These tools were necessary to create and design the maturation and stress models for my doctoral studies.