Yusi Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Yusi Chen has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Terrence J. Sejnowski
Computational Neurobiology Laboratory, The Salk Institute for Biological Studies
7 shared publications - Herbert Zheng Wu
Nash Family Department of Neuroscience and Friedman Brain Institute, Icahn School of Medicine at Mount Sinai, New York, USA
3 shared publications Alvi Khan
Nash Family Department of Neuroscience and Friedman Brain Institute, Icahn School of Medicine at Mount Sinai, New York, USA
2 shared publications- Burke Q. Rosen
Neuroscience, Washington Univ. in St. Louis, Saint Louis, MO
2 shared publications - Giri P. Krishnan
Department of Medicine, University of California San Diego, La Jolla, CA 92092, U.S.A.
2 shared publications Jared Winslow
Quantitative Sciences Unit, Department of Medicine, Stanford University, Stanford, USA
2 shared publications
Latest Publications
- Open Access
- 19 August 2026
- Published by Springer Nature in Nature
- Vol. 658 (8134) , 174-185
Asymmetric prefrontal representations for leader–follower dynamics
- 19 August 2026
- Published by Springer Nature in Nature
- Vol. 658 (8134) , 174-185
Open AccessAsymmetric prefrontal representations for leader–follower dynamics
- Open Access
Balancing Stability and Flow in Hippocampal Networks via Inductive Bias and Learned Symmetry Breaking
The hippocampus must balance stable memory representations with internally generated sequential dynamics underlying replay and prediction. How hippocampal circuitry achieves both remains unclear. Here, we show that recurrent neural networks trained on prediction tasks converge to a mixed-symmetry dynamical regime, in which dominant symmetric recurrence stabilizes an attractor while a weaker antisymmetric component induces directed flow. This structure accelerates learning and supports robust replay and prediction. We further show that such symmetry breaking can arise from biologically plausible spike-timing-dependent plasticity (STDP) rules, yielding a tilted Mexican-hat connectivity profile without explicit architectural constraints. Initializing networks with CA3-like structured connectivity biases learning toward this regime, and improves optimization efficiency and performance. These results suggest that hippocampal computation reflects an interaction between biologically constrained circuit structure and task-driven learning, with partial symmetry breaking providing a low-dimensional control principle for balancing stability and flow in sequence generation. Subject Areas: computational neuroscience, systems neuroscience, neural dynamics, learning, memory
Open AccessBalancing Stability and Flow in Hippocampal Networks via Inductive Bias and Learned Symmetry Breaking
The hippocampus must balance stable memory representations with internally generated sequential dynamics underlying replay and prediction. How hippocampal circuitry achieves both remains unclear. Here, we show that recurrent neural networks trained on prediction tasks converge to a mixed-symmetry dynamical regime, in which dominant symmetric recurrence stabilizes an attractor while a weaker antisymmetric component induces directed flow. This structure accelerates learning and supports robust replay and prediction. We further show that such symmetry breaking can arise from biologically plausible spike-timing-dependent plasticity (STDP) rules, yielding a tilted Mexican-hat connectivity profile without explicit architectural constraints. Initializing networks with CA3-like structured connectivity biases learning toward this regime, and improves optimization efficiency and performance. These results suggest that hippocampal computation reflects an interaction between biologically constrained circuit structure and task-driven learning, with partial symmetry breaking providing a low-dimensional control principle for balancing stability and flow in sequence generation. Subject Areas: computational neuroscience, systems neuroscience, neural dynamics, learning, memory
- Open Access
Asymmetric Social Representations in the Prefrontal Cortex for Cooperative Behavior
Cooperation is a hallmark of social species, enabling individuals to achieve goals that are unattainable alone. Across species, cooperative behaviors are often organized by distinct social roles such as leaders and followers, yet the neural mechanisms supporting such role-based coordination remain elusive. Here we introduce a new paradigm for studying cooperation in mice, where pairs of animals engage in a joint spatial foraging task that naturally gives rise to stable leader-follower roles predictive of learning speed. Disruption of medial prefrontal cortex (mPFC) activity, particularly in followers, impairs cooperation and induces reciprocal shifts in how animals weigh self- and partner-related cues for decision-making. Calcium imaging reveals that mPFC encodes both leadership dynamics and an egocentric social value map of the partner’s position, each in an asymmetric, role-specific manner. Combining this behavior with a novel multi-agent inverse reinforcement learning framework, we identify latent value functions that guide cooperative decisions and are decodable from mPFC activity. These findings uncover fundamental neural computations that support cooperation, revealing how social roles shape decision-making in real time. Our work opens new avenues for investigating the cellular and circuit basis of social cognition and collective behavior.
Open AccessAsymmetric Social Representations in the Prefrontal Cortex for Cooperative Behavior
Cooperation is a hallmark of social species, enabling individuals to achieve goals that are unattainable alone. Across species, cooperative behaviors are often organized by distinct social roles such as leaders and followers, yet the neural mechanisms supporting such role-based coordination remain elusive. Here we introduce a new paradigm for studying cooperation in mice, where pairs of animals engage in a joint spatial foraging task that naturally gives rise to stable leader-follower roles predictive of learning speed. Disruption of medial prefrontal cortex (mPFC) activity, particularly in followers, impairs cooperation and induces reciprocal shifts in how animals weigh self- and partner-related cues for decision-making. Calcium imaging reveals that mPFC encodes both leadership dynamics and an egocentric social value map of the partner’s position, each in an asymmetric, role-specific manner. Combining this behavior with a novel multi-agent inverse reinforcement learning framework, we identify latent value functions that guide cooperative decisions and are decodable from mPFC activity. These findings uncover fundamental neural computations that support cooperation, revealing how social roles shape decision-making in real time. Our work opens new avenues for investigating the cellular and circuit basis of social cognition and collective behavior.
- Open Access
Unveiling value functions in social cognition with multi-agent inverse reinforcement learning
Social behavior requires individuals to consider not only their own goals but also those of others. Latent reward functions that encode such goals can be recovered from behavior using inverse reinforcement learning in single-agent settings. However, extending this to multi-agent interactions is challenging, as joint reward functions are defined over joint state spaces that grow exponentially with the number of agents. Existing approaches address this complexity by imposing restrictive assumptions on social interactions, limiting generality and interpretability. Here we show that joint reward functions governing social interactions can be efficiently represented through decomposition into individual reward maps for each agent and low-dimensional interaction terms. Building on this formulation, we introduce a general multi-agent inverse reinforcement learning framework (MAIRL) that enables scalable and interpretable inference of latent reward representations directly from behavior. In simulated social tasks with known ground-truth rewards, MAIRL recovered the rewarded states and imitated expert behavior as accurately as or better than existing multi-agent inverse reinforcement learning methods. Applying MAIRL to mouse and primate social tasks, we recover structured reward maps that reflect distinct social roles and interaction strategies during group behavior. These results establish MAIRL as a general framework for identifying latent objectives underlying multi-agent behavior across species and paradigms.
Open AccessUnveiling value functions in social cognition with multi-agent inverse reinforcement learning
Social behavior requires individuals to consider not only their own goals but also those of others. Latent reward functions that encode such goals can be recovered from behavior using inverse reinforcement learning in single-agent settings. However, extending this to multi-agent interactions is challenging, as joint reward functions are defined over joint state spaces that grow exponentially with the number of agents. Existing approaches address this complexity by imposing restrictive assumptions on social interactions, limiting generality and interpretability. Here we show that joint reward functions governing social interactions can be efficiently represented through decomposition into individual reward maps for each agent and low-dimensional interaction terms. Building on this formulation, we introduce a general multi-agent inverse reinforcement learning framework (MAIRL) that enables scalable and interpretable inference of latent reward representations directly from behavior. In simulated social tasks with known ground-truth rewards, MAIRL recovered the rewarded states and imitated expert behavior as accurately as or better than existing multi-agent inverse reinforcement learning methods. Applying MAIRL to mouse and primate social tasks, we recover structured reward maps that reflect distinct social roles and interaction strategies during group behavior. These results establish MAIRL as a general framework for identifying latent objectives underlying multi-agent behavior across species and paradigms.
- Open Access
- 9 June 2022
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 119 (24)
Dynamical differential covariance recovers directional network structure in multiscale neural systems
Investigating neural interactions is essential to understanding the neural basis of behavior. Many statistical methods have been used for analyzing neural activity, but estimating the direction of network interactions correctly and efficiently remains a difficult problem. Here, we derive dynamical differential covariance (DDC), a method based on dynamical network models that detects directional interactions with low bias and high noise tolerance under nonstationarity conditions. Moreover, DDC scales well with the number of recording sites and the computation required is comparable to that needed for covariance. DDC was validated and compared favorably with other methods on networks with false positive motifs and multiscale neural simulations where the ground-truth connectivity was known. When applied to recordings of resting-state functional magnetic resonance imaging (rs-fMRI), DDC consistently detected regional interactions with strong structural connectivity in over 1,000 individual subjects obtained by diffusion MRI (dMRI). DDC is a promising family of methods for estimating connectivity that can be generalized to a wide range of dynamical models and recording techniques and to other applications where system identification is needed.
- 9 June 2022
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 119 (24)
Open AccessDynamical differential covariance recovers directional network structure in multiscale neural systems
Investigating neural interactions is essential to understanding the neural basis of behavior. Many statistical methods have been used for analyzing neural activity, but estimating the direction of network interactions correctly and efficiently remains a difficult problem. Here, we derive dynamical differential covariance (DDC), a method based on dynamical network models that detects directional interactions with low bias and high noise tolerance under nonstationarity conditions. Moreover, DDC scales well with the number of recording sites and the computation required is comparable to that needed for covariance. DDC was validated and compared favorably with other methods on networks with false positive motifs and multiscale neural simulations where the ground-truth connectivity was known. When applied to recordings of resting-state functional magnetic resonance imaging (rs-fMRI), DDC consistently detected regional interactions with strong structural connectivity in over 1,000 individual subjects obtained by diffusion MRI (dMRI). DDC is a promising family of methods for estimating connectivity that can be generalized to a wide range of dynamical models and recording techniques and to other applications where system identification is needed.
- Open Access
- 1 June 2022
- Published by MIT Press in Network Neuroscience
- Vol. 6 (2) , 614-633
Functional connectivity of fMRI using differential covariance predicts structural connectivity and behavioral reaction times
Recordings from resting-state functional magnetic resonance imaging (rs-fMRI) reflect the influence of pathways between brain areas. A wide range of methods have been proposed to measure this functional connectivity (FC), but the lack of “ground truth” has made it difficult to systematically validate them. Most measures of FC produce connectivity estimates that are symmetrical between brain areas. Differential covariance (dCov) is an algorithm for analyzing FC with directed graph edges. When we applied dCov to rs-fMRI recordings from the human connectome project (HCP) and anesthetized mice, dCov-FC accurately identified strong cortical connections from diffusion magnetic resonance imaging (dMRI) in individual humans and viral tract tracing in mice. In addition, those HCP subjects whose dCov-FCs were more integrated, as assessed by a graph-theoretic measure, tended to have shorter reaction times in several behavioral tests. Thus, dCov-FC was able to identify anatomically verified connectivity that yielded measures of brain integration significantly correlated with behavior.
- 1 June 2022
- Published by MIT Press in Network Neuroscience
- Vol. 6 (2) , 614-633
Open AccessFunctional connectivity of fMRI using differential covariance predicts structural connectivity and behavioral reaction times
Recordings from resting-state functional magnetic resonance imaging (rs-fMRI) reflect the influence of pathways between brain areas. A wide range of methods have been proposed to measure this functional connectivity (FC), but the lack of “ground truth” has made it difficult to systematically validate them. Most measures of FC produce connectivity estimates that are symmetrical between brain areas. Differential covariance (dCov) is an algorithm for analyzing FC with directed graph edges. When we applied dCov to rs-fMRI recordings from the human connectome project (HCP) and anesthetized mice, dCov-FC accurately identified strong cortical connections from diffusion magnetic resonance imaging (dMRI) in individual humans and viral tract tracing in mice. In addition, those HCP subjects whose dCov-FCs were more integrated, as assessed by a graph-theoretic measure, tended to have shorter reaction times in several behavioral tests. Thus, dCov-FC was able to identify anatomically verified connectivity that yielded measures of brain integration significantly correlated with behavior.
- Open Access
Predictive Sequence Learning in the Hippocampal Formation
Summary: The hippocampus receives sequences of sensory inputs from the cortex during exploration and encodes the sequences with millisecond precision. We developed a predictive autoencoder model of the hippocampus including the trisynaptic and monosynaptic circuits from the entorhinal cortex (EC). CA3 was trained as a self-supervised recurrent neural network to predict its next input. We confirmed that CA3 is prediction ahead by analyzing the spike coupling between simultaneously recorded neurons in the dentate gyrus, CA3 and CA1 of the mouse hippocampus. In the model, CA1 neurons signal prediction errors by comparing the prediction from CA3 to the next input directly from the EC. The model exhibits the rapid appearance and the slow fading of CA1 place cells, and displays replay and phase precession from CA3. The model could be learnt in a biologically plausible way with the help of error-encoding neurons. Similarities between the circuits in the hippocampus and thalamocortical circuits suggest that such computation motif could also underlie self-supervised sequence learning in the cortex.
Open AccessPredictive Sequence Learning in the Hippocampal Formation
Summary: The hippocampus receives sequences of sensory inputs from the cortex during exploration and encodes the sequences with millisecond precision. We developed a predictive autoencoder model of the hippocampus including the trisynaptic and monosynaptic circuits from the entorhinal cortex (EC). CA3 was trained as a self-supervised recurrent neural network to predict its next input. We confirmed that CA3 is prediction ahead by analyzing the spike coupling between simultaneously recorded neurons in the dentate gyrus, CA3 and CA1 of the mouse hippocampus. In the model, CA1 neurons signal prediction errors by comparing the prediction from CA3 to the next input directly from the EC. The model exhibits the rapid appearance and the slow fading of CA1 place cells, and displays replay and phase precession from CA3. The model could be learnt in a biologically plausible way with the help of error-encoding neurons. Similarities between the circuits in the hippocampus and thalamocortical circuits suggest that such computation motif could also underlie self-supervised sequence learning in the cortex.
- Open Access
Functional connectivity of fMRI using differential covariance predicts structural connectivity and behavioral reaction times
Recordings from resting state functional magnetic resonance imaging (rs-fMRI) reflect the influence of pathways between brain areas. A wide range of methods have been proposed to measure this functional connectivity (FC), but the lack of “ground truth” has made it difficult to systematically validate them. Most measures of FC produce connectivity estimates that are symmetrical between brain areas. Differential covariance (dCov) is an algorithm for analyzing FC with directed graph edges. Applied to synthetic datasets, dCov-FC was more effective than covariance and partial correlation in reducing false positive connections and more accurately matching the underlying structural connectivity. When we applied dCov-FC to resting state fMRI recordings from the human connectome project (HCP) and anesthetized mice, dCov-FC accurately identified strong cortical connections from diffusion Magnetic Resonance Imaging (dMRI) in individual humans and viral tract tracing in mice. In addition, those HCP subjects whose rs-fMRI were more integrated, as assessed by a graph-theoretic measure, tended to have shorter reaction times in several behavioral tests. Thus, dCov-FC was able to identify anatomically verified connectivity that yielded measures of brain integration causally related to behavior.
Open AccessFunctional connectivity of fMRI using differential covariance predicts structural connectivity and behavioral reaction times
Recordings from resting state functional magnetic resonance imaging (rs-fMRI) reflect the influence of pathways between brain areas. A wide range of methods have been proposed to measure this functional connectivity (FC), but the lack of “ground truth” has made it difficult to systematically validate them. Most measures of FC produce connectivity estimates that are symmetrical between brain areas. Differential covariance (dCov) is an algorithm for analyzing FC with directed graph edges. Applied to synthetic datasets, dCov-FC was more effective than covariance and partial correlation in reducing false positive connections and more accurately matching the underlying structural connectivity. When we applied dCov-FC to resting state fMRI recordings from the human connectome project (HCP) and anesthetized mice, dCov-FC accurately identified strong cortical connections from diffusion Magnetic Resonance Imaging (dMRI) in individual humans and viral tract tracing in mice. In addition, those HCP subjects whose rs-fMRI were more integrated, as assessed by a graph-theoretic measure, tended to have shorter reaction times in several behavioral tests. Thus, dCov-FC was able to identify anatomically verified connectivity that yielded measures of brain integration causally related to behavior.
- Open Access
Dynamical differential covariance recovers directional network structure in multiscale neural systems
Investigating causal neural interactions are essential to understanding subsequent behaviors. Many statistical methods have been used for analyzing neural activity, but efficiently and correctly estimating the direction of network interactions remains difficult (1). Here, we derive dynamical differential covariance (DDC), a new method based on dynamical network models that detects directional interactions with low bias and high noise tolerance without the stationary assumption. The method is first validated on networks with false positive motifs and multiscale neural simulations where the ground truth connectivity is known. Then, applying DDC to recordings of resting-state functional magnetic resonance imaging (rs-fMRI) from over 1,000 individual subjects, DDC consistently detected regional interactions with strong structural connectivity. DDC can be generalized to a wide range of dynamical models and recording techniques.
Open AccessDynamical differential covariance recovers directional network structure in multiscale neural systems
Investigating causal neural interactions are essential to understanding subsequent behaviors. Many statistical methods have been used for analyzing neural activity, but efficiently and correctly estimating the direction of network interactions remains difficult (1). Here, we derive dynamical differential covariance (DDC), a new method based on dynamical network models that detects directional interactions with low bias and high noise tolerance without the stationary assumption. The method is first validated on networks with false positive motifs and multiscale neural simulations where the ground truth connectivity is known. Then, applying DDC to recordings of resting-state functional magnetic resonance imaging (rs-fMRI) from over 1,000 individual subjects, DDC consistently detected regional interactions with strong structural connectivity. DDC can be generalized to a wide range of dynamical models and recording techniques.
- 1 December 2020
- Published by MIT Press in Neural Computation
- Vol. 32 (12) , 2389-2421
Differential Covariance: A New Method to Estimate Functional Connectivity in fMRI
Measuring functional connectivity from fMRI recordings is important in understanding processing in cortical networks. However, because the brain's connection pattern is complex, currently used methods are prone to producing false functional connections. We introduce differential covariance analysis, a new method that uses derivatives of the signal for estimating functional connectivity. We generated neural activities from dynamical causal modeling and a neural network of Hodgkin-Huxley neurons and then converted them to hemodynamic signals using the forward balloon model. The simulated fMRI signals, together with the ground-truth connectivity pattern, were used to benchmark our method with other commonly used methods. Differential covariance achieved better results in complex network simulations. This new method opens an alternative way to estimate functional connectivity. Measuring functional connectivity from fMRI recordings is important in understanding processing in cortical networks. However, because the brain's connection pattern is complex, currently used methods are prone to producing false functional connections. We introduce differential covariance analysis, a new method that uses derivatives of the signal for estimating functional connectivity. We generated neural activities from dynamical causal modeling and a neural network of Hodgkin-Huxley neurons and then converted them to hemodynamic signals using the forward balloon model. The simulated fMRI signals, together with the ground-truth connectivity pattern, were used to benchmark our method with other commonly used methods. Differential covariance achieved better results in complex network simulations. This new method opens an alternative way to estimate functional connectivity. Measuring functional connectivity from fMRI recordings is important in understanding processing in cortical networks. However, because the brain's connection pattern is complex, currently used methods are prone to producing false functional connections. We introduce differential covariance analysis, a new method that uses derivatives of the signal for estimating functional connectivity. We generated neural activities from dynamical causal modeling and a neural network of Hodgkin-Huxley neurons and then converted them to hemodynamic signals using the forward balloon model. The simulated fMRI signals, together with the ground-truth connectivity pattern, were used to benchmark our method with other commonly used methods. Differential covariance achieved better results in complex network simulations. This new method opens an alternative way to estimate functional connectivity.
- 1 December 2020
- Published by MIT Press in Neural Computation
- Vol. 32 (12) , 2389-2421
Differential Covariance: A New Method to Estimate Functional Connectivity in fMRI
Measuring functional connectivity from fMRI recordings is important in understanding processing in cortical networks. However, because the brain's connection pattern is complex, currently used methods are prone to producing false functional connections. We introduce differential covariance analysis, a new method that uses derivatives of the signal for estimating functional connectivity. We generated neural activities from dynamical causal modeling and a neural network of Hodgkin-Huxley neurons and then converted them to hemodynamic signals using the forward balloon model. The simulated fMRI signals, together with the ground-truth connectivity pattern, were used to benchmark our method with other commonly used methods. Differential covariance achieved better results in complex network simulations. This new method opens an alternative way to estimate functional connectivity. Measuring functional connectivity from fMRI recordings is important in understanding processing in cortical networks. However, because the brain's connection pattern is complex, currently used methods are prone to producing false functional connections. We introduce differential covariance analysis, a new method that uses derivatives of the signal for estimating functional connectivity. We generated neural activities from dynamical causal modeling and a neural network of Hodgkin-Huxley neurons and then converted them to hemodynamic signals using the forward balloon model. The simulated fMRI signals, together with the ground-truth connectivity pattern, were used to benchmark our method with other commonly used methods. Differential covariance achieved better results in complex network simulations. This new method opens an alternative way to estimate functional connectivity. Measuring functional connectivity from fMRI recordings is important in understanding processing in cortical networks. However, because the brain's connection pattern is complex, currently used methods are prone to producing false functional connections. We introduce differential covariance analysis, a new method that uses derivatives of the signal for estimating functional connectivity. We generated neural activities from dynamical causal modeling and a neural network of Hodgkin-Huxley neurons and then converted them to hemodynamic signals using the forward balloon model. The simulated fMRI signals, together with the ground-truth connectivity pattern, were used to benchmark our method with other commonly used methods. Differential covariance achieved better results in complex network simulations. This new method opens an alternative way to estimate functional connectivity.