Biology subjects

Sohn, H.

Publications and source records attributed to Sohn, H..

13 recordsLinked to original sources

bioRxiv2026

Physical sensorimotor contingencies facilitate learning in monkeys

Animal studies in neuroscience routinely use virtual tasks that deviate significantly from the animals' natural behavioral repertoire. This choice is often necessary for conducting careful and controlled experimentation. However, this approach raises a fundamental question: do virtual tasks expose neural computations and mechanisms associated with naturalistic tasks? In this study, we explore how naturalistic settings influence learning timescale and generalization behavior in rhesus monkeys. By training animals in a maze-solving task and comparing their behavior across two environments, we observe significant differences. Specifically, we found that a realistic setup using physical objects enabled learning that was orders of magnitude faster than the learning timescale observed in a comparable computerized virtual environment. Furthermore, we noted limited transfer of performance between the physical and virtual settings. These findings underscore the need for understanding how task environments shape learning and generalization, with potential implications for the design of future neuroscience and AI experiments.

neuroscience↗

bioRxiv2026

Hypothesis-Driven Identification of Neural Algorithms With Dynamical Structure-Preserving Manifolds.

Neuroscience aims to uncover the algorithms by which the brain builds and manipulates rich internal representations of objects, agents, and places. Existing reverse-engineering frameworks remain ineffective for identifying these algorithms. Here, we present Dynamical Structure-Preserving Manifolds (dSPMs)---a novel multilevel reverse-engineering framework that programs, without training, recurrent neural networks to enable hypothesis-driven identification of the brain's algorithms that generate symbolic, structure-preserving representations of the world around. We apply dSPM to identify an algorithm for physical scene prediction in the dorsomedial frontal cortex (DMFC) of macaques intercepting a ball on a Pong-like virtual gameboard. dSPM better explained DMFC than alternatives, made predictions confirmed in neural data, and revealed an algorithm of DMFC that we term 'manifold-based simulation': perceived initial conditions rapidly collapse neural activity onto a low-dimensional manifold whose geometry encodes a physics-based representation of the scene, supporting far-out predictions without traditional next-step simulation. dSPM supports effective exploration of the computational foundations of biological intelligence.

neuroscience↗

bioRxiv2026

A number simplex in the human medial temporal lobe

Humans handle numbers nimbly, suggesting a richer neural manifold structure than the prevalent mental number line model. In populations of medial temporal lobe (MTL) neurons in humans performing two simple tasks (dot counting and arithmetic), we find robust neural coding of numerosity that results in high dimensional, simplex-shaped manifolds. This shape affords more flexibility than a linear manifold due to its high shattering dimensionality and expressibility. Dot arrays and Arabic numerals evoked distinct simplicial population codes, yet they were linked by a linearly transferable latent structure within the same task. We find similar simplicial geometry of number representations in large language models (LLMs). Moreover, subjects internally computed arithmetic results were decodable during the calculation period, with decoding accuracy correlating with individual mathematical capacity. Finally, linear transformations of simplicial operand representations modeled the brains conversion of operands into decodable results, suggesting that the brains arithmetic procedures have some resemblance to the attention architecture of LLMs. Together, these findings establish a high dimensional representational foundation for numerical cognition in the brain.

neuroscience↗

bioRxiv2026

Tuned inhibitory control of neuronal firing thresholds explains predictive sensorimotor behavior

Prior expectations guide sensorimotor behavior when sensory information is uncertain, yet the cellular mechanisms underlying this integration remain elusive. Here, we investigate how priors in sensory motion direction shape neural population dynamics by combining recurrent neural network (RNN) modeling with macaque smooth pursuit behavior and middle temporal area (area MT) electrophysiology. Our RNN model reveals that prior expectations are implemented by elevating firing thresholds in neurons tuned away from the expected direction. This selective inhibition sharpens population tuning and reduces behavioral variability under weak sensory information conditions. We validated this prediction in vivo: delta-band local field potentials in area MT, where the phase reflects neural excitability states, exhibited direction-specific phase shifts that scaled with the deviation from the expected direction. These findings demonstrate that prior expectations enhance behavioral reliability through tuned inhibitory control of neuronal excitability, providing a mechanistic link between Bayesian inference and cortical circuit dynamics.

neuroscience↗

bioRxiv2025

ACLY promotes NK cell effector function by regulating glycolysis and histone acetylation

Natural Killer (NK) cells are innate immune lymphocytes important for host viral and tumor immunity. We investigated the requirement for ATP citrate lyase (ACLY) in NK cell function using an inducible genetic mouse model. ACLY regulates the citrate-malate shuttle, generating cytosolic acetyl-coenzyme A that is primarily used for acetylation or lipid synthesis. ACLY-deficient NK cells upon IL-15 activation exhibited significant defects in glycolysis, proliferation, cytokine production, and cytotoxicity, without decreased intracellular lipids. Notably, ACLY-deficiency specifically resulted in reduced NK cell responses to activating receptors associated with the adapter proteins DAP10 or DAP12. This is due to decreased DAP12 and increased DAP10 transcript and protein, coupled with epigenetic profiling that demonstrated altered histone acetylation of these genes in ACLY KO. Supplementation of ACLY-deficient NK cells with acetate was sufficient to overcome most functional defects, including restoring DAP10/12 expression and activating receptor function, emphasizing the importance of ACLY-generated cytosolic acetyl-coenzyme A for NK effector functions.

immunology↗

bioRxiv2024

Dokdo sea lion Zalophus japonicus genome reveals its evolutionary trajectory before extinction

BackgroundThe Dokdo sea lion (Zalophus japonicus), commonly referred to as Gangchi in Korea and known as the Japanese sea lion, was endemic to the Northwest Pacific coast before becoming extinct in the 1950s. Little is known about its origins, speciation compared to other Otariidae species or how the rapid decline affected the species genetic diversity. ResultsTo raise the Dokdo sea lion from this relative obscurity, we sequenced DNA from 16 Z. japonicus bone fragments, obtained from Dokdo and Ulleungdo islands in Korea. Our analysis establishes Z. japonicus as the earliest diverged species within its genus, significantly redefining its evolutionary relationship with the California (Z. californianus) and Galapagos (Z. wollebaeki) sea lions. This genome-scale analysis clarifies the phylogeny of Z. japonicus, shedding light on its speciation and the evolutionary pathways that shaped its genetic diversity before its extinction. In addition, we discovered, population decline of the Z. japonicus started already 1,000 years ago, however, Z. japonicus genome maintained a relatively high heterozygosity despite, nearing extinction. ConclusionsOur genome-scale analysis eliminated ambiguity in Z. japonicus phylogeny, and shed light on the evolutionary pathways underlying its speciation. This study highlights the importance of the genome-scale analysis for extinct species to understand their complex evolutionary histories and conservation status.

genomics↗

bioRxiv2024

Ribonanza: deep learning of RNA structure through dual crowdsourcing

Prediction of RNA structure from sequence remains an unsolved problem, and progress has been slowed by a paucity of experimental data. Here, we present Ribonanza, a dataset of chemical mapping measurements on two million diverse RNA sequences collected through Eterna and other crowdsourced initiatives. Ribonanza measurements enabled solicitation, training, and prospective evaluation of diverse deep neural networks through a Kaggle challenge, followed by distillation into a single, self-contained model called RibonanzaNet. When fine tuned on auxiliary datasets, RibonanzaNet achieves state-of-the-art performance in modeling experimental sequence dropout, RNA hydrolytic degradation, and RNA secondary structure, with implications for modeling RNA tertiary structure.

biophysics↗

bioRxiv2022

Dynamic tracking of objects in the macaque dorsomedial frontal cortex

Cognitive neuroscience posits that humans perform physical inferences using mental simulations. Here, we test this hypothesis by analyzing neural activity in the monkeys frontal cortex in a ball interception task. We find a low-dimensional neural embedding of the ball position that tracks the ball both when it is visible and invisible. This embedding may serve as a neural substrate for mental simulation.

neuroscience↗

bioRxiv2021

Parametric control of flexible timing through low-dimensional neural manifolds

Biological brains possess an unparalleled ability to adapt behavioral responses to changing stimuli and environments. How neural processes enable this capacity is a fundamental open question. Previous works have identified two candidate mechanisms: a low-dimensional organization of neural activity and a modulation by contextual inputs. We hypothesized that combining the two might facilitate generalization and adaptation in complex tasks. We tested this hypothesis in the framework of flexible timing tasks where dynamics play a key role. Examining trained recurrent neural networks, we found that confining the dynamics to a low-dimensional subspace allowed tonic inputs to parametrically control the overall input-output transform, enabling generalization to novel inputs and adaptation to changing conditions. Reverse-engineering and theoretical analyses demonstrated that this parametric control relies on a mechanism where tonic inputs modulate the dynamics along non-linear manifolds in activity space while preserving their geometry. Comparisons with data from behaving monkeys confirmed the behavioral and neural signatures of this mechanism.

neuroscience↗

bioRxiv2021

A precise and adaptive neural mechanism for predictive temporal processing in the frontal cortex

The theory of predictive processing posits that the nervous system uses expectations to process information predictively. Direct empirical evidence in support of this theory however has been scarce and largely limited to sensory areas. Here, we report a precise and adaptive neural mechanism in the frontal cortex of non-human primates consistent with predictive processing of temporal events. We found that the speed at which neural states evolve over time is inversely proportional to the statistical mean of the temporal distribution of an expected stimulus. This lawful relationship was evident across multiple experiments and held true during learning: when temporal statistics underwent covert changes, neural responses underwent predictable changes that reflected the new mean. Together, these results highlight a precise mathematical relationship between temporal statistics in the environment and neural activity in the frontal cortex that could serve as a mechanistic foundation for predictive temporal processing.

neuroscience↗

bioRxiv2021

A large-scale neural network training framework for generalized estimation of single-trial population dynamics

Recent technical advances have enabled recording of increasingly large populations of neural activity, even during natural, unstructured behavior. Deep sequential autoencoders are the current state-of-the-art for uncovering dynamics from these datasets. However, these highly complex models include many non-trainable hyperparameters (HPs) that are typically hand tuned with reference to supervisory information (e.g., behavioral data). This process is cumbersome and time consuming and biases model selection toward models with good representations of individual supervisory variables. Additionally, it cannot be applied to cognitive areas or unstructured tasks for which supervisory information is unavailable. Here we demonstrate AutoLFADS, an automated model-tuning framework that can characterize dynamics using only neural data, without the need for supervisory information. This enables inference of dynamics out-of-the-box in diverse brain areas and behaviors, which we demonstrate on several datasets: motor cortex during free-paced reaching, somatosensory cortex during reaching with perturbations, and dorsomedial frontal cortex during cognitive timing tasks. We also provide a cloud software package and comprehensive tutorials that enable new users to apply the method without dedicated computing resources.

neuroscience↗

bioRxiv2020

Metamers of Bayesian computation

There are two sharply debated views on how humans make decisions under uncertainty. Bayesian decision theory posits that humans optimize their behavior by establishing and integrating internal models of past sensory experiences (priors) and decision outcomes (cost functions). An alternative model-free hypothesis posits that decisions are optimized through trial and error without explicit internal models for priors and cost functions. To distinguish between these possibilities, we introduce a novel paradigm that probes sensitivity of humans to transitions between prior-cost pairs that demand the same optimal policy (metamers) but distinct internal models. We demonstrate the utility of our approach in two experiments that were classically explained by model-based Bayesian theory. Our approach validates the model-based strategy in an interval timing task but not in a visuomotor rotation task. More generally, our work provides a domain-general approach for testing the circumstances under which humans implement model-based Bayesian computations.

neuroscience↗

bioRxiv2018

Bayesian computation through cortical latent dynamics

Statistical regularities in the environment create prior beliefs that we rely on to optimize our behavior when sensory information is uncertain. Bayesian theory formalizes how prior beliefs can be leveraged, and has had a major impact on models of perception 1, sensorimotor function 2,3, and cognition 4. However, it is not known how recurrent interactions among neurons mediate Bayesian integration. Using a time interval reproduction task in monkeys, we found that prior statistics warp the underlying structure of population activity in the frontal cortex allowing the mapping of sensory inputs to motor outputs to be biased in accordance with Bayesian inference. Analysis of neural network models performing the task revealed that this warping was mediated by a low-dimensional curved manifold, and allowed us to further probe the potential causal underpinnings of this computational strategy. These results uncover a simple and general principle whereby prior beliefs exert their influence on behavior by sculpting cortical latent dynamics.

neuroscience↗