Biology subjects

Schwartz, A. B.

Publications and source records attributed to Schwartz, A. B..

6 recordsLinked to original sources

bioRxiv2025

Hybrid Neural Networks for Volitional Movement

Massive interconnectivity in large-scale neural networks is the key feature underlying their powerful and complex functionality. We have developed hybrid neural network (HNN) models that allow us to find statistical structure in this connectivity. Describing this structure is critical for understanding biological and artificial neural networks. The HNNs are composed of artificial neurons, a subset of which are trained to reproduce the responses of individual neurons recorded experimentally. The experimentally observed firing rates came from populations of neurons recorded in the motor cortices of monkeys performing a reaching task. After training, these networks (recurrent and spiking) underwent the same state transitions as those observed in the empirical data, a result that helps resolve a long-standing question of prescribed vs ongoing control of volitional movement. Because all aspects of the models are exposed, we were able to analyze the dynamic statistics of the connections between neurons. Our results show that the dynamics of extrinsic input to the network changed this connectivity to cause the state transitions. Two processes at the synaptic level were recognized: one in which many different neurons contributed to a buildup of membrane potential and another in which more specific neurons triggered an action potential. HNNs facilitate modeling of realistic neuron-neuron connectivity and provide foundational descriptions of large-scale network functionality.

neuroscience↗

bioRxiv2025

A Generalist Intracortical Motor Decoder

Mapping the relationship between neural activity and motor behavior is a central aim of sensori-motor neuroscience and neurotechnology. While most progress to this end has relied on restricting complexity, the advent of foundation models instead proposes integrating a breadth of data as an alternate avenue for broadly advancing downstream modeling. We quantify this premise for motor decoding from intracortical microelectrode data, pretraining an autoregressive Transformer on 2000 hours of neural population spiking activity paired with diverse motor covariates from over 30 monkeys and humans. The resulting model is broadly useful, benefiting decoding on 8 downstream decoding tasks and generalizing to a variety of neural distribution shifts. However, we also highlight that scaling autoregressive Transformers seems unlikely to resolve limitations stemming from sensor variability and output stereotypy in neural datasets. Code: https://github.com/joel99/ndt3

neuroscience↗

bioRxiv2024

Neutrophils exhibit distinct migration phenotypes that are modulated by transendothelial migration

The extravasation of polymorphonuclear neutrophils (PMNs) is a critical component of the innate immune response that involves transendothelial migration (TEM) and interstitial migration. TEM-mediated interactions between PMNs and vascular endothelial cells (VECs) trigger a cascade of biochemical and mechanobiological signals whose effects on interstitial migration are currently unclear. To address this question, we cultured human VECs on a fibronectin-treated transwell insert to model the endothelium and basement membrane, loaded PMN-like differentiated HL60 (dHL-60) cells in the upper chamber of the insert, and collected the PMNs that crossed the membrane-supported monolayer from the lower chamber. The 3D chemotactic migration of the TEM-conditioned PMNs through collagen matrices was then quantified. Data collected from over 50,000 trajectories showed two distinct migratory phenotypes, i.e., a high-persistence phenotype and a low-persistence phenotype. These phenotypes were conserved across treatment conditions, and their existence was confirmed in human primary PMNs. The high-persistence phenotype was characterized by more straight trajectories and faster migration speeds, whereas the low-persistence one exhibited more frequent sharp turns and loitering periods. A key finding of our study is that TEM induced a phenotypic shift in PMNs from high-persistence migration to low-persistence migration. Changes in the relative proportion of high-persistence and low-persistence populations correlated with GRK2 expression levels. Inhibiting GRK2 hindered the TEM-induced shift in migratory phenotype and impaired the phagocytic function of PMNs. Overall, our study suggests that TEM-mediated GRK2 signaling primes PMNs for a migration phenotype better suited for spatial exploration and inflammation resolution. These observations provide novel insight into the biophysical impacts of TEM that priming PMNs is essential to conduct sentinel functions.

cell biology↗

bioRxiv2018

Stiffness as a control factor for object manipulation

We act on the world by producing forces that move objects. During manipulation, force is exerted with the expectation that an object will move in an intended manner. This prediction is a learned coordination between force and displacement. Mechanically, impedance is a way to describe this coordination. As an efficient control strategy, object interaction could be anticipated by setting impedance before the hand moves the object. We examined this possibility with a paradigm in which subjects moved a handle to a specific target position along a track. The handle was locked in place until the subject exerted enough force to cross a specific threshold; then the handle was abruptly released and could move along the track. We hypothesized that this ballistic-release task would encourage subjects to modify their arm impedance in anticipation of the upcoming movement. If we consider the handle as an object, this paradigm loosely approximates the uncertainty encountered at the end of a reach when contacting a fixed object. We found that one component of arm impedance, stiffness, varied in a way that matched the behavioral demands of the task and we were able to dissociate stiffness from changes in force and displacement. We also found separate components of muscle activity that corresponded to stiffness and to changes in force. Our results show that subjects used a robust and efficient strategy to coordinate force and displacement by modulating muscle activity in a way that was behaviorally relevant in the task.\n\nNew & NoteworthyThe arm can behave like a spring, suggesting the concept of exerting force to move an object by selecting a spring of a certain length and stiffness that, respectively, depend on the movement and force requirements of the task. We show that these spring-like characteristics describe the strategy used to arrest a pre-loaded handle. These results extend our understanding of the arms spring-like behavior to include force and movement constraints, important factors for object interaction.

neuroscience↗

bioRxiv2018

Viral-mediated optical stimulation of peripheral motor nerves in non-human primates

ObjectiveReanimation of muscles paralyzed by disease states such as spinal cord injury remains a much sought after therapeutic goal of neuroprosthetic research. Optogenetic stimulation of peripheral motor nerves expressing light-sensitive opsins is a promising approach to muscle reanimation that may overcome several drawbacks of traditional methods such as functional electrical stimulation (FES). However, the utility of these methods has only been demonstrated in rodents to date, while translation to clinical practice will likely first require demonstration and refinement of these gene therapy techniques in non-human primates.\n\nApproachThree rhesus macaques were injected intramuscularly with either one or both of two optogenetic constructs (AAV6-hSyn-ChR2-eYFP and/or AAV6-hSyn-Chronos-eYFP) to transduce opsin expression in the corresponding nerves. Neuromuscular junctions were targeted for virus delivery using an electrical stimulating injection technique. Functional opsin expression was periodically evaluated up to 13 weeks post-injection by optically stimulating targeted nerves with a 472 nm fiber-coupled laser while recording electromyographic (EMG) responses.\n\nMain ResultsOne monkey demonstrated functional expression of ChR2 at 8 weeks post-injection in each of two injected muscles, while the second monkey briefly exhibited contractions coupled to optical stimulation in a muscle injected with the Chronos construct at 10 weeks. A third monkey injected only in one muscle with the ChR2 construct showed strong optically coupled contractions at 5 [1/2] weeks which then disappeared by 9 weeks. EMG responses to optical stimulation of ChR2-transduced nerves demonstrated graded recruitment relative to both stimulus pulse-width and light intensity, and were able to track stimulus trains up to 16 Hz. In addition, the EMG response to prolonged stimulation showed delayed fatigue over several minutes.\n\nSignificanceThese results demonstrate the feasibility of viral transduction of peripheral motor nerves for functional optical stimulation of motor activity in non-human primates, a variable timeline of opsin expression in a primate model closer to humans, and fundamental EMG response characteristics to optical nerve stimulation. Subsequently, they represent an important step in translating these optogenetic techniques as a clinically viable gene therapy.

bioengineering↗

bioRxiv2016

Deconstructing cell size control into physiological modules in Escherichia coli

It is generally assumed that the allocation and synthesis of total cellular resources in microorganisms are uniquely determined by the growth conditions. Adaptation to a new physiological state leads to a change in cell size via reallocation of cellular resources. However, it has not been understood how cell size is coordinated with biosynthesis and robustly adapts to physiological states. We show that cell size in Escherichia coli can be predicted for any steady-state condition by projecting all biosynthesis into three measurable variables representing replication initiation, replication-division cycle, and the global biosynthesis rate. These variables can be decoupled by selectively controlling their respective core biosynthesis using CRISPR interference and antibiotics, verifying our predictions that different physiological states can result in the same cell size. We performed extensive growth inhibition experiments, and discovered that cell size at replication initiation per origin, namely the initiation mass or \"unit cell,\" is remarkably invariant under perturbations targeting transcription, translation, ribosome content, replication kinetics, fatty acid and cell-wall synthesis, cell division, and cell shape. Based on this invariance and balanced resource allocation, we explain why the total cell size is the sum of all unit cells. These results provide an overarching framework with quantitative predictive power over cell size in bacteria.

microbiology↗