Elizabeth Jun (Nuyujukian Lab) & Lavonna Mark (Giocomo Lab)
Today's seminar features two talks from MBCT Student Members, Stanford PhD Students Elizabeth Jun and Lavonna Mark
Join our mailing list for MBCT Seminar announcements and updates about future events
Investigating preserved neural motifs across and within tasks in premotor cortex
Abstract
Motor systems neuroscience studies how populations of motor cortical neurons coordinate to perform motor computations, but less is known about how large repertoires of movements are flexibly generated from a fixed set of fundamental motor computations. There is an emerging idea of conserved neural “modules”, or recurring neural activity patterns deploying computations through dynamics, that may serve as elementary building blocks for performing the same fundamental computation across a wide array of contexts. However, exactly what feature of neural dynamics is “recurring” and fundamentally conserved, if any, when performing the same motor computation within a wide range of reach sequence and kinematic contexts has yet to be shown and characterized empirically.
To investigate this, we trained an animal implanted with a 96-channel Utah array in premotor cortex to perform 4 visually guided arm reaching motor tasks that are composed of unique combinations of 5 distinct motor processes (preparation, initiation, execution, hold, error correction) implemented in various reach conditions, where the same process can reappear across tasks and reappear again within tasks. This design allows us to test the hypothesis that premotor dynamics reuse conserved neural dynamics for performing the same motor computational process, but potentially implemented in different locations and orientations in population state space when reperformed across tasks, repeated in a sequence within tasks, and in different kinematic conditions. In this talk, I will explore systematically testing this hypothesis for all 5 motor processes to understand whether there are general organizing principles for premotor dynamics associated with deploying the same fundamental motor computation in a range of tasks, sequences, and kinematic contexts.
Meet the Speaker
Elizabeth Jun, PhD Student, Nuyujukian Lab, Stanford University
Elizabeth Jun is a neuroscience PhD Candidate at Stanford University and NIH F31 Predoctoral Fellow funded through National Institute of Neurological Disorders and Stroke (NINDS). Her research interests lie in understanding how large networks of neurons coordinate together to perform computations and support a wide range of behaviors in different contexts. Specifically in her thesis work, advised by Dr. Paul Nuyujukian in the Brain Interfacing Laboratory, she aims to understand how the brain can flexibly support a wide range of complex movements that are supported by a fixed set of fundamental motor computations. She is interested in connecting features of high-dimensional population data to similar behaviors in different contexts, and exploring in general how high-dimensional population features relate back to properties of the neural circuit.
State-dependent grid cell spatial maps
Abstract
To navigate, the brain requires an internal map of external space. The medial entorhinal cortex (MEC) plays an important role in generating this map, with spatial cells that encode speed, head-direction, and position. This map is not static, however: MEC neurons can shift the position or rate of their firing fields, a process termed remapping. Remapping occurs in response to external changes to the environment (its color, shape, size, or reward features) but was recently found to occur also spontaneously in an unchanging environment (Low et al., 2021). Why spatial maps reorganize in a stable environment remains unclear.
In this talk, I will address two questions. First, what drives spontaneous remapping? I will show that transitions between grid cell maps track changes in the animal’s internal arousal state, and that optogenetically driving state transitions can evoke spatial remapping. Second, how does internal state shape spatial dynamics? Modeling the toroidal grid cell attractor dynamics in spontaneous and optogenetically-evoked map transitions reveals that each state’s map reflects a distinct strength and phase at which landmarks pin grid cell dynamics. Together, these results suggest the presence of state-dependent spatial maps in the MEC, in which internal state tunes how landmarks entrain intrinsic attractor dynamics.
Meet the Speaker
Lavonna Mark, PhD Student, Giocomo Lab, Stanford University
Lavonna Mark is a PhD candidate in the Giocomo Lab. She is interested in how neural circuits can dynamically switch their coding properties across internal states. Prior to graduate school, she studied how arousal states influence visual coding and perception in the Cardin and Higley Labs at the Yale School of Medicine. As a PhD candidate, she is interested in how downstream navigational circuits integrate state-dependent signals to coordinate adaptive behavior. Her research combines Neuropixels recordings in mice navigating virtual reality, behavior videography, optogenetics, and fiber photometry with computational approaches (continuous attractor and state space models), the latter in collaboration with Scott Linderman.
Continue the conversation: All attendees are invited to digest the seminar with a complimentary dinner with the speaker in the Theory Center after the seminar.