Seminars

Navigation and memory require coordination between map-like allocentric representations and body-centered egocentric actions. Aaron Wilber examines how the parietal–retrosplenial–anterior thalamic–hippocampal network performs these transformations, and how disrupted coordination contributes to spatial deficits in aging and Alzheimer’s disease. Rat sequence-task experiments reveal bidirectional hippocampal–parietal interactions across route-centered, place and egocentric representations. Because that task does not isolate reference-frame use, the laboratory developed a freely available task separating egocentric, allocentric and transformation conditions. Validation indicates that parietal cortex and anterior thalamus are required for transformation; ongoing work examines population states associated with each condition. The talk then connects impaired hippocampal–cortical communication during sleep to navigation deficits in 3xTg-AD mice, reporting recovery of circuit function and cognition after early, circuit-targeted 40 Hz entrainment. Related investigations examine this network during waking navigation. In TgF344-AD rats, egocentric impairments appear before transformation deficits, addressing the shortage of rodent studies of reference-frame coordination in Alzheimer’s disease.

This webinar uses research on Chagas disease to explain how predicted protein structures can support drug discovery when experimental structures are unavailable. In the example study, AlphaFold models of Trypanosoma cruzi proteins supported a virtual screen of roughly 30,000 compounds; subsequent experiments identified antiparasitic activity in two approved drugs. The session introduces AlphaFold confidence measures and ways to obtain models, then follows a screening workflow through binding-pocket prediction, chemical-library selection, docking, and interpretation of results. It discusses difficulties encountered when docking against predicted structures and considerations for adapting the workflow to other targets. The material is intended for doctoral students, early-career researchers, and other drug-discovery professionals.

Jan Kosinski explains how AlphaFold-style methods predict protein complexes and why their reliance on evolutionary information makes interactions between host and pathogen proteins especially difficult. The talk reviews large-scale studies, interpreting both successful predictions and failures rather than treating reported success rates as universal performance. It then examines ways to improve predictions through broader sampling, altered sequence alignments and experimental restraints, including crosslinking mass spectrometry, with influenza A virus as a worked research example. The session is suitable for students and researchers interested in host-pathogen interactions and structural bioinformatics; basic knowledge of protein structures and sequence alignments is useful, but prior AlphaFold experience is unnecessary.

Matthias Vigl examines whether the computational scaling strategies behind modern machine learning can benefit particle physics. Empirical scaling laws relate model performance to computing resources and help balance model size against training-data volume. For a fixed dataset, increasing model capacity beyond the interpolation threshold can improve generalization through double descent, but gains eventually encounter limits imposed by the data. The talk then considers scaling data and model size together, deriving compute-optimal relations for transformer-based jet taggers trained on as many as billions of simulated jets. It studies how training hyperparameters and input representations change these relations. The observed behaviour suggests that larger computing budgets could yield useful gains in particle physics, especially with a move toward more general-purpose foundation models.

This public discussion examines the factors that help preserve brain health during aging. Researchers working on neuroscience, brain stimulation, the microbiome and brain health consider how stress, everyday behavior and social and environmental conditions can affect resilience over a lifetime. The panel discusses emerging evidence about early vulnerability, possible interventions through lifestyle changes and stress management, and more targeted approaches. It also addresses the limits of current knowledge and the uncertainties that remain when translating these findings into ways to protect cognitive health.

Paul S. Weiss (UCLA) explores low-energy interactions and balanced cycling in natural systems as models for efficient synthetic devices and recycling. Because the relevant energies fall below visible-photon energies, tunnelling spectroscopy and imaging are used to probe these processes. Replicating biochemical cycling and low-energy transport could reduce the energy needed for devices and materials recovery. The work seeks missing physical principles, including spin conservation in chiral molecules and polarizability in biological and synthetic systems. This perspective has led to thermal-control advances in scale, speed and magnitude. Atomic-resolution imaging, spectroscopy and functional measurements help build precise structures while revealing heterogeneous nanoscale behaviour. The talk relates these advances to the physical, electronic, mechanical, thermal and chemical connections between materials and their surroundings that determine future nanomaterial measurements and devices.

Deep networks are often thought of as black boxes. Their ability to encompass vast swathes of knowledge indeed makes them hard to explain. Yet many of their behaviours — generalization under overparametrization, grokking, OOD failures, neural scaling laws — recur across architectures and scales, and each, however surprising, can be reproduced and explained in controlled settings. I will review these efforts to identify and explain the universal phenomena of deep learning, and suggest that an LLM may amount to a sum of such tractable sub-phenomena, interpolative in nature. Finally, leaving scientific rigor aside, I'll argue that what separates this prosaic picture from the apparent magic of LLMs may well be the industrial scale of compute and human labour behind it, and that AGI in its deeper extrapolative sense may be much further away than claimed.

Jie Yang presents BRIDGE, a multilingual evaluation benchmark built from real clinical tasks and more than one million samples derived from electronic health records. The benchmark addresses the gap between simplified examination-style tests and the complexity of clinical data, while enabling comparison of rapidly changing language models. Evaluation of 95 models involved over 24,000 experiments and 39 million predictions. Results vary substantially by model family, task and language, with some open-source systems matching proprietary models. The study also finds that chain-of-thought prompting frequently reduces accuracy on these tasks and examines stigmatizing language produced during model reasoning at large scale.

Tom Nowakowski explains lineage barcoding and single-cell sequencing approaches for reconstructing neural stem-cell family trees and early wiring programs in developing mouse and human brains.

Lindsay N. Oberman reviews evidence for transcranial magnetic stimulation protocols in autism spectrum disorder. The seminar first explains how TMS works and how it can probe the brain mechanisms associated with autism. It then presents NIH research using repetitive TMS as an experimental therapeutic tool, examining its effects on both brain function and behaviour to guide the development of potential interventions. Online via Duke Zoom. Wednesday 7 October 2026, 16:00 EDT (America/New_York; UTC−4). Register through the event-specific Register Here link on the Duke Autism Seminar Series page to receive remote access. The presentation will not be recorded. Accommodation requests should reach the organizer at least ten days before the event. The speaker directs NIMH’s Developmental Clinical Neurophysiology and Neurostimulation Research Program.

Shikha Bhutani proves the outstanding positive-characteristic cases of the Borisov–Alexeev–Borisov conjecture for geometrically integral ε-klt del Pezzo surfaces over arbitrary fields. Earlier work by Bernasconi and Martin established boundedness except in characteristics 2, 3 and 5, reducing those cases to a uniform estimate for irregularity. The talk establishes that estimate and thereby completes the boundedness result across all positive characteristics.

Clarice D. Aiello examines evidence that quantum effects influence biological function, including magnetic sensing used in animal navigation, cellular metabolism and enzyme activity, and the capture of light energy in photosynthesis. The talk reviews proposed biophysical mechanisms and their implications for human, plant and environmental biology, while considering the difficulty of connecting observations from nanometre scales to whole organisms. A central question is whether quantum mechanics can be established or ruled out as the explanation for physiologically relevant effects, and whether those effects can be controlled for useful applications. Potential directions include biomimetic electromagnetic probes, quantum-computing architectures that operate at room temperature, improved solar-energy devices and new therapies. Progress will require work across disciplines and scales, with theory and experiment developed together.

How can learning systems recommend actions that prevent an undesirable predicted outcome? Rehearsal learning approaches this problem through influence relations, but existing parametric assumptions can limit its use in complex systems. This talk presents a non-parametric method based on conditional mean embeddings. It expresses the objective in a reproducing kernel Hilbert space, replaces a discontinuous desirability indicator with a smooth Probit surrogate, and uses nested kernel ridge regression to estimate outcome distributions conditional on actions. The resulting estimator can be identified from observational data and has finite-sample error bounds and consistency guarantees. Synthetic and semi-synthetic experiments assess its effectiveness and flexibility.

Nicole Yunger Halpern (NIST, QuICS and University of Maryland) examines how to describe nonequilibrium matter in the early universe and particle colliders from the Standard Model. Classical lattice-gauge calculations have limited reach, motivating quantum simulation. Local Gauss-law constraints create an intricate Hilbert-space structure that complicates thermodynamic definitions for systems coupled to reservoirs, both at equilibrium and during nonequilibrium processes. The talk develops definitions of work and heat through strong-coupling thermodynamics, applying them to instantaneous quenches that quantum simulators can implement. A Z2 lattice gauge theory coupled to matter in one spatial and one temporal dimension provides the example: computed work and heat vary with the quench parameter and reveal the expected phase transition. The framework connects gauge constraints, reservoir coupling and experimentally relevant thermodynamic quantities.

Susanne Yelin explains how modest control of analog quantum platforms can reveal observables and dynamics of strongly correlated fermions, including pairing, long-range interactions and cooling. The seminar then examines universal quantum computation under global control, optimal-control methods for engineering effective interactions and topological dynamics, and approximate error correction designed for analog systems. Together, these approaches treat large quantum devices as adaptable computational resources whose capabilities extend beyond simulating a single model. Hybrid attendance: free in-person participation at Michigan Memorial Phoenix Project, Room 2000, 2301 Bonisteel Boulevard, Ann Arbor, MI 48109, USA, or join remotely through the Zoom link in the organizer page’s main event description. Thursday 8 October, 11:00–12:00 EDT (America/Detroit).

Junjie Guo investigates how an intronic nucleotide-repeat expansion in C9orf72 causes familial and some sporadic cases of amyotrophic lateral sclerosis and frontotemporal dementia. The talk presents progress in understanding how the expansion changes gene expression, alongside functional-genomics approaches that identify potential therapeutic targets controlling expression of the pathogenic mutation. It also considers which mechanisms and modifiers extend to other disease-causing repeat expansions. Guo is Associate Professor of Neuroscience at Yale School of Medicine; the seminar is hosted by J. Wren Kim. Online via CUNY ASRC Zoom. Thursday 8 October 2026, 12:00–13:00 EDT (America/New_York; UTC−4). Follow the Zoom link on the official seminar page; meeting details are also supplied in its abstract PDF. Organized by the Advanced Science Research Center at the CUNY Graduate Center, Neuroscience Initiative. This listing advertises remote attendance.

Takaki Komiyama will discuss how cortical circuits maintain information about the value of behavioral options across trials and transform past experience into future choices. The talk draws on longitudinal imaging and circuit manipulation in mice to examine distinct neuronal populations and local and long-range interactions in history-dependent decision-making.

Takaki Komiyama presents longitudinal imaging and circuit-manipulation studies of history-dependent decision-making. The work identifies retrosplenial-cortex populations and local and long-range interactions that maintain option values across trials, update them after outcomes and translate past experience into future choices.

This seminar investigates Anderson localization in subwavelength elastic periodic systems with random perturbations. For the unperturbed structure, layer-potential methods recast the eigenvalue problem as boundary integral equations, yielding asymptotic expressions for subwavelength eigenvalues and establishing a band gap above the subwavelength band. A Floquet transform then brings the perturbed problem into a periodic formulation and produces equations for resonant frequencies under general perturbations. Numerical experiments on monomer and dimer structures test agreement with the analysis. Increasing the strength and number of random perturbations demonstrates localization, providing a mathematical basis for elastic metamaterial design.

Astrocytes help sustain neural signaling and tissue stability, but injury and implanted devices cause them to change their functions. Tim O'Shea examines how the extent and timing of this response determine tissue protection, regeneration and implant durability. Limited renewal of adult astrocytes contrasts with the greater proliferation and migration of immature cells in newborn mammals. Around implants, continuing movement, infection and other stimuli can prolong the response and affect device performance. The talk uses astrocyte-specific transcriptomic assays to investigate this reprogramming and considers how to improve repair without sacrificing normal astrocyte functions. It also presents biomaterial and cell-transplant approaches being studied in preclinical stroke, spinal cord injury and implant-response models.