Welcome to my site!
I am an HFSP Cross-Disciplinary Postdoctoral Fellow at EPFL in the Signal Processing Lab 2 led by Pierre Vandergheynst.
Previously, I was a postdoc in the Neuroengineering lab of Pavan Ramdya’s at EPFL. I received my PhD in Applied Mathematics at Imperial College London supervised jointly by Mauricio Barahona and Martin Buck. Before that, I studied Applied Mathematics (Part III) at the University of Cambridge and Mechanical Engineering at University College London.
Keywords
animal behaviour – machine learning – geometry – dynamical systems
Research interests
Complex structures often emerge from simpler components interacting in space and time. In music, notes build phrases that combine into movements. Analogously, the behaviour of an animal foraging for food arises from neurons and muscles generating body posture dynamics that combine into sequences of epochs like walking and turning. How do goal-oriented behaviours like foraging emerge from the interactions of components in the musculoskeletal and nervous systems? How do neuronal networks enable healthy animal behaviours despite intrinsic variability in their connectivity and how does their malfunction lead to motor impairments?
In my research, I use graph theory, dynamical systems theory and machine learning to study how network interactions between neurons and muscles scale up to complex, goal-oriented behaviours. I am particularly interested in leveraging the wealth of data arising in neuroscience thanks to advances in connectomics, functional neuroimaging and computer vision. The common feature of these datasets is that they can be modelled as coupled dynamic signals on networks—neuronal activity on connectomes or joint kinematics on body skeletons—enabling the data-driven discovery of fundamental design principles of neuromechanical control systems and how they go awry in motor disorders.
Contact
The best way to contact me is by e-mail at:
a (dot) surname (at) gmail (dot) com
Adam Gosztolai 