Biology subjects

Yartsev, M. M.

Publications and source records attributed to Yartsev, M. M..

6 recordsLinked to original sources

bioRxiv2026

A hippocampal digital Twin reveals latent structure in freely flying bats

Concurrent advances in artificial intelligence (AI), large-scale neuronal recording, and behavioral tracking have enabled increasingly powerful predictive models of brain-behavior relationships. Yet predictive accuracy does not automatically yield interpretable principles of neural organization, particularly during natural behavior, where the relevant variables computed are often unknown in advance. Here, we developed a functional digital twin of the hippocampus, trained jointly on large-scale neural activity and self-directed three-dimensional flight in freely flying bats. A multimodal, multitask transformer accurately predicted both neural activity and behavior, outperformed standard approaches, and generalized across recording sessions. Interrogating the model's representations suggested a form of hippocampal generalization across self-selected trajectories that was not apparent in physical coordinates, but was plainly visible in the model's low-dimensional latent space. We confirmed this organization directly in the recorded neural activity: neurons fired at corresponding stages of spatially distinct paths connecting the same origin and destination, preserving their relative firing order even as their firing fields shifted with the path. Analysis of neuron-specific readouts further identified functional relationships among neurons that generalized from flight to waking rest and sleep, brain states never used for training, and differed systematically between the hippocampal subregions CA1 and CA3. Together, these results show how AI-based functional digital twins can complement conventional analysis by generating testable hypotheses about neural organization during natural behavior.

neuroscience↗

bioRxiv2026

A Two-Dimensional Grid-Cell Code for Three-Dimensional Navigation in Freely Flying Bats

Animals have evolved to navigate environments with diverse geometries and dimensionalities, yet how neural circuits for spatial navigation support behavior across such varied demands remains unresolved. In two-dimensional environments, grid cells in the medial entorhinal cortex provide a periodic signal for self-location that is thought to arise from underlying toroidal attractor dynamics at the ensemble level. However, during movement in three-dimensional environments, studies have reported a loss of global grid-cell periodicity. Furthermore, it remains unknown whether, at the ensemble level, a toroidal attractor is preserved across species and, if so, how such a two-dimensional code could support navigation in three-dimensional space. Here we performed large-scale, wireless neural recordings from the medial entorhinal cortex of freely flying bats engaged in spontaneous aerial foraging. We find that grid cells exhibit robust periodic firing during structured flight trajectories, and that co-modular grid-cell ensembles show topological signatures consistent with a two-dimensional toroidal manifold. Behavioral analyses of bats navigating in the wild and in the laboratory revealed that individual flight paths are organized along two-dimensional planes of motion, providing a natural substrate for a two-dimensional code. Accordingly, toroidal phase along flights was consistent with plane-wave patterns on the plane of motion, and single-cell firing was well described by a hexagonal lattice on the same plane. Together, these findings reveal a parsimonious solution whereby a two-dimensional neural code aligns with behaviorally relevant subspaces to support navigation in a three-dimensional world.

neuroscience↗

bioRxiv2023

Effects of deafening on vocal production learning in the Egyptian fruit-bat

Some species have evolved the ability to use the sense of hearing to modify existing vocalizations, or even create new ones. This ability corresponds to various forms of vocal production learning that are all possessed by humans, and independently displayed by distantly related vertebrates. Among mammals, a few species, including the Egyptian fruit-bat, would possess such vocal production learning abilities. Yet the necessity of an intact auditory system for the development of the Egyptian fruit-bat typical vocal repertoire has not been tested. Furthermore, a systematic causal examination of learned and innate aspects of the entire repertoire has never been performed in any vocal learner. Here we addressed these gaps by eliminating pups sense of hearing at birth and assessing its effects on vocal production in adulthood. The deafening treatment enabled us to both causally test these bats vocal learning ability and discern learned from innate aspects of their vocalizations. Leveraging wireless individual audio recordings from freely interacting adults, we show that a subset of the Egyptian fruit-bat vocal repertoire necessitates auditory feedback. Intriguingly, these affected vocalizations belong to different acoustic groups in the vocal repertoire of males and females. These findings open the possibilities for targeted studies of the mammalian neural circuits that enable sexually dimorphic forms of vocal learning.

animal behavior and cognition↗

bioRxiv2022

Vocal learning-associated convergent evolution in mammalian proteins and regulatory elements

Vocal learning, the ability to modify vocal behavior based on experience, is a convergently evolved trait in birds and mammals. To identify genomic elements associated with vocal learning, we integrated new experiments conducted in the brain of the Egyptian fruit bat with analyses of the genomes of 222 placental mammals. We first identified an anatomically specialized region of the bat motor cortex containing direct monosynaptic projections to laryngeal motoneurons. Using wireless neural recordings of this brain region in freely vocalizing bats, we verified that single neuron activity in this region relates to vocal production. We profiled the open chromatin of this vocal-motor region, which we used to train machine learning models to identify enhancers associated with vocal learning across mammals. We found 201 proteins and 45 candidate enhancers that display convergent evolution associated with vocal learning, many of which overlapped loci associated with human speech disability. One such locus contains the neurodevelopmental transcription factors TSHZ3 and ZNF536 and multiple candidate vocal learning-associated enhancers, suggesting the co-evolution of protein and regulatory sequences underlying vocal learning. One-Sentence SummaryAnalyses of bat neural activity and epigenomic data in a brain region involved in vocal behavior were used to identify proteins and regulatory elements associated with vocal learning in mammals.

neuroscience↗

bioRxiv2021

A unifying mechanism governing inter-brain neural relationship during social interactions

A key goal of social neuroscience is to understand the relationship between the neural activity of socially interacting individuals. Decades of research have focused on a single aspect of that relationship: the similarity in neural activity across brains. Here we instead asked how neural activity differs between brains, and how that difference evolves alongside activity patterns shared between brains. Applying this framework to pairs of bats engaged in spontaneous social interactions revealed two complementary phenomena characterizing the inter-brain neural relationship: fast "inter-brain catch-up" unfolding in parallel with slow activity covariation across brains. A model reproduced these observations, generated multiple predictions that we confirmed using experimental data, and provided testable hypotheses for studying the inter-brain relationship in larger social groups. Together, the data and model suggest a parsimonious computational mechanism--opposite feedback to neural activity components reflecting inter-brain difference and similarity--that unifies diverse aspects of the inter-brain neural relationship.

neuroscience↗

bioRxiv2018

Causal contribution and dynamical encoding in the striatum during evidence accumulation

A broad range of decision-making processes involve gradual accumulation of evidence over time, but the neural circuits responsible for this computation are not yet established. Recent data indicates that cortical regions prominently associated with accumulating evidence, such as posterior parietal cortex and the frontal orienting fields, are not necessary for computing it. Which, then, are the regions responsible? Regions directly involved in evidence accumulation should satisfy the criteria of being necessary for accumulation-based decision-making behavior, having a graded neural encoding of accumulated evidence and causal contributing throughout the accumulation process. Here, we investigated the role of the anterior dorsal striatum (ADS) in a rodent auditory evidence accumulation task using a combination of behavioral, pharmacological, optogenetic, electrophysiological and computational approaches. We find that the ADS is the first brain region known to satisfy these criteria. Thus, the ADS may be the first identified node in the network responsible for evidence accumulation.

neuroscience↗