Internal Seminar: Gioia Unit and De Schutter Unit

Join us for January's first Internal Seminar Series on Jan. 09, from 17:00 to 18:00 in B250. This month's seminars feature the Continuum Physics Unit (Gustavo Gioia) and the Computational Neuroscience Unit (Erik De Schutter).

Continuum Physics Unit (Gustavo Gioia)

Speaker : Gustavo Gioia

Title : Potholes

Abstract : There is a sand-bedded pothole at the foot of Taa Falls, northern Okinawa. That pothole is a favorite spot of summer bathers and revelers. It is also a reduced-size version of larger potholes (variously known as scour trenches, scour pits, blue holes and crevasse lakes) which have long been of concern to hydraulic engineers, hydrologists, geomorphologists and planetary scientists. Regardless of size, a pothole forms where a turbulent flow attains a state of equilibrium with an erodible granular bed. In this talk we use dimensional analysis and similarity methods to derive a theoretical formula for the depth of a pothole. Whereas empirical formulas, of which there are several, contain numerous free exponents, the theoretical formula contains only one, the ``similarity exponent.'' We also show that the similarity exponent can be determined via the phenomenological theory of turbulence, which to our knowledge has never been exploited in prior research on potholes. Our findings are in good accord with, and shed new light upon, experimental data on potholes. What is more, our findings suggest that the phenomenological theory could help develop a physical understanding of bridge pier-induced erosion, mine burial and other processes akin to pothole formation. This research was carried out in collaboration with the Fluid Mechanics Unit and with Fabian Bombardelli of the University of California, Davis.

Computational Neuroscience Unit (Erik De Schutter)

Speaker : Ben Torben-Nielsen

Title : Context-aware modeling of neuronal morphologies

Abstract : Neuronal morphology is pivotal for normal brain functioning: axonal and dendritic morphologies define the circuit connectivity and dendritic morphology sets the computations performed on inputs. At the same time neurons are highly diverse and variant. Diversity between types can be assumed to stem from the genetic make-up of a neuron while the variance is assumed to originate from interactions between developing neurons and the brain substrate. Indeed, a plethora of such microscopic interactions have been demonstrated to influence morphology. But how do combinations of these interactions shape full morphologies and circuits at large?

We argue that to fully understand neuronal morphology it is important to consider neurons in relation to each other and to other actors in the surrounding brain substrate, i.e., their context. We propose a context-aware computational framework in which large numbers of neurons are generated simultaneously according to phenomenological growth rules expressed in terms of repulsion and attraction from/to actors in their context. As a proof of principle, we demonstrate that by using this framework we can generate accurate virtual morphologies of distinct classes both in isolation and as part of neuronal forests. We speculate about the applicability of this framework to investigate morphologies and circuits, to classify healthy and pathological morphologies, and to generate large quantities of morphologies for large-scale modeling.