Eric Sembrat's Test Bonanza

Abstract

Quantification of behavior is one of the primary requirements to study animal behavior scientifically. Traditionally, behavior has been quantified by manually observing the focal animal(s) across a spatio-temporal scale and recording the occurrences of behavioral events. These events are generally deduced from the movement of different body parts of the animal, typical body postures as well as its overall movement. While collecting data by manual observation has several advantages, it is prone to disadvantages like human bias and being imprecise. Though modern videography has improved the observation and recording of behavior, extracting behavioral data from these video data remained challenging until now.

In this talk, I will discuss how the recent progress in machine learning tools has enabled me, a biologist interested in social insects, to extract behavioral data from videos. I will talk in detail about such a tool, called DeepLabCut, which tracks the movement of individual parts of an animal with minimal human input. I will end the talk with an example of the application of this tool in my current projects, which is understanding the evolution of cooperation and foraging strategies in the Carpenter ants.

Abstract

A cell’s interactions with the environment are mediated by its cellular membrane and its ultrastructures. Since cells contain and are surrounded by fluids, the hydrodynamics involved during these interactions are critical. They determine the transport of ions and molecules, locomotion, and sensing. Overall, they mediate a cells’ homeostasis. In this talk, I will present investigations into the fluid-structure interactions of membrane-bound unicellular organisms, specifically fluid slip close to and transmission of shear forces across cellular membranes, and flow enhancement due to fibrous flagellar hairs during swimming. Optical tweezers are used to both apply and measure local forces on free-standing membranes and tethered algal cells. Through these measurements, we find that fluid slip and shear force transmission are highly dependent on the membrane’s lipid composition, and that, contrary to previous claims, the fibrous flagellar hairs of algae do not increase the flagella's effective area while swimming. Our studies therefore contribute towards building a fundamental understanding of the physical principles governing the transfer of hydrodynamic forces by and through the membrane and its ultrastructures.

Abstract

Yield-stress fluids can be engineered from a diverse range of microstructures including polymeric gels, colloidal glasses, emulsions, pastes, foams, and more. This talk will discuss our research on design thinking for yield-stress fluids (Nelson et al. Curr Opin Solid St M, 2019), addressing the rheology-to-structure inverse problem of the many ways to achieve a yield stress and important secondary properties including thixotropy and extensibility, e.g. for direct-write 3D printing. Microstructural considerations include conceptual models relating formulation to properties, quantitative models of formulation-structure-property relations, and chemical transformation strategies for converting effective yield-stress fluids to be more useful solid engineering materials. Future research directions are suggested at the intersection of chemistry, soft-matter physics, and material science in the context of our desire to design useful rheologically-complex functional materials.

Bio

Randy H. Ewoldt, Associate Professor, Department of Mechanical Science and Engineering, University of Illinois at Urbana-Champaign, conducts fundamental research in fluid mechanics and rheology. This includes new measurement methods and paradigms for design with materials such as yield stress fluids, polymer gels, and biological materials. His work has been recognized by awards from NSF (PECASE), ASME, 3M, DuPont, TA Instruments, and The Society of Rheology. His teaching has been recognized by awards from the College of Engineering and alumni of the University of Illinois. In 2018 he was a Guest Professor at ETH-Zürich in the Department of Materials.

Abstract

The observation of gravitational waves has opened a new, unexplored window onto the Universe. Among the sources of gravitational wave transients, compact objects such as neutron stars (NSs) and black holes (BHs) play the most important role. In this talk, I will focus on the expected gravitational wave signal when two compact objects (NS-NS and NS-BH) in a binary merge. These events are believed to be accompanied by a strong electromagnetic signature in gamma-rays, followed by longer-wavelength radiation.  I will discuss what can be learned from the complementary observations of the electromagnetic and the gravitational wave signals during these events.

Abstract: Fast neutrons from fission or spallation, which are created with MeV energies, can be cooled to < 300 neV to become ultracold neutrons (UCNs). This 12 orders-of-magnitude energy reduction allows them to be trapped in material bottles or magnetic traps and be studied for tens of minutes. High-precision UCN experiments using UCNs have impact in nuclear physics, particle physics, fundamental symmetries, cosmology, and condensed matter physics. A flagship experiment is the search for a permanent neutron electric dipole moment (nEDM). Its detection would be an indication of a new beyond-standard-model source of time-reversal symmetry violation, one that is needed to explain the excess of matter to anti-matter in the Universe from baryogenesis. Since the 1980s the most precise measurements have come from using Ramsey’s
separated oscillatory fields technique on UCNs. Our collaboration will perform a new experiment at the Spallation Neutron Source (SNS) at Oak Ridge National Lab using a novel technique that exploits unique properties of combining polarized UCNs and polarized 3He in a superfluid 4He bath at 0.4 K. Two distinct measurement modes can be performed with our setup: double free precession and critical spin dressing. Our sensitivity goal is a two orders-of-magnitude improvement on the current world limit in both statistics and systematics. This seminar will give a brief overview of UCNs and then focus on the nEDM@SNS experiment, which is envisioned to come online in the next few years. Overlaps with AMO physics will be highlighted where possible.

Abstract

We typically think of memories as occurring in magnetic storage devices or neural networks, but memory effects abound in a wide range of materials. Rubbers and rocks can remember the largest loading applied to them; glasses may remember their past relaxation; shape-memory alloys remember a programmed shape. Many more examples exist in physics, biology, and chemistry. Despite the ubiquity of memory formation in condensed matter, there is presently no overarching framework for categorizing and describing such material memories. I will describe a budding effort to change this [1]. After an overview of several known memory behaviors, I will describe some qualitative connections and tentative classifications that attempt to bring together disparate systems. I will pay particular attention to an unusual kind of memory where learning and forgetting are intertwined, observed in sheared non-Brownian suspensions, traveling charge-density waves, and a model of a worn path. Such a comprehensive study of memory has not been pursued before. If it is successful in discerning universal principles, they could apply across scales, from microphysical to astrophysical.

[1] “Memory formation in matter.” NC Keim, JD Paulsen, Z Zeravcic, S Sastry, and SR Nagel. Rev. Mod. Phys. 91, 035002 (2019).

Abstract

The ability to move is a trait of all animals. Yet how do animals, including ourselves, get around in this complex and uncertain world with an ease and agility we find hard to recreate in engineered systems? Using an organismal physics approach my group explores the physical and physiological mechanisms that enable agile movement in living systems. I will discuss studies at three scales in animal movement. We first examine flight maneuvers as emergent dynamics from the underlying complex physiological systems. Using robophysical models of flowers, we probed how maneuverable, hovering hawk moths adjust to light levels that vary by 7 orders of magnitude from early afternoon to late dusk. Using simple one-parameter models from control theory we discover the brain slows its visual processing to accomplish temporal summation in dim light – like increasing the exposure time on a camera. Perhaps more surprisingly the resulting dynamics are empirically linear, invariant under experimental scaling and superposition. This allows us to predict behavior to novel stimulus combinations. We next investigate the neural basis of these behaviors by recording a comprehensive motor program – all the electrical signals that the moth’s brain sends to it wings. The brains of small flapping insects are still complex, with 10⁵-10⁶ neurons, but we find that they control wingstroke dynamics with as few as 10¹ motor units that mostly utilize millisecond precise timing to inform torque. Finally we explored how muscle’s macroscopic properties during locomotion are shaped by its unusual multiscale structure. High-speed x-ray diffraction through living muscles shows that muscle is active crystalline matter – the regular arrangement of actin and myosin filaments produces a lattice that dynamically changes spacing as a muscle contracts. As a result, muscle has a time-varying poisson ratio including an auxetic regime and these dynamics result in flow that assists the convective delivery of molecules. Moreover a single nanometer difference in muscle lattice spacing can account for how one muscle acts like a motor while another acts like a brake. We cannot yet emulate the motility seen in nature, nor derive behavior, but the emergent dynamics of animal locomotion is an exciting opportunity to explore the how complexity gives way to function and the physical and physiological mechanisms that are the enablers of this performance.