Haozhe Shan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Haozhe Shan has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Peggy Mason
The University of Chicago
5 shared publications - Inbal Ben-Ami Bartal
Tel Aviv University, School of Psychological Sciences
3 shared publications - Haim Sompolinsky
Center for Brain Science, Harvard University
2 shared publications - Jasper Z. Williams
Cell Design Institute and Department of Cellular and Molecular Pharmacology, University of California, San Francisco, USA
2 shared publications - Jean Decety
Department of Psychology, University of Chicago
2 shared publications - Nora M. R. Molasky
Department of Neurobiology, University of Chicago
2 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 6 February 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (6)
Order parameters and phase transitions of continual learning in deep neural networks
Continual learning (CL) enables animals to learn new tasks without erasing prior knowledge. CL in artificial neural networks (NNs) is challenging due to catastrophic forgetting, where new learning degrades performance on older tasks. While various techniques exist to mitigate forgetting, theoretical insights into when and why CL fails in NNs are lacking. Here, we present a statistical-mechanics theory of CL in deep, wide NNs, which characterizes the network’s input–output mapping as it learns a sequence of tasks. It gives rise to order parameters (OPs) that capture how task relations and network architecture influence forgetting and anterograde interference, as verified by numerical evaluations. For networks with a shared readout for all tasks (single-head CL), the relevant-feature and rule similarity between tasks, respectively measured by two OPs, are sufficient to predict a wide range of CL behaviors on classic benchmark tasks. In addition, the theory predicts that increasing the network depth can effectively reduce interference between tasks, thereby lowering forgetting. For networks with task-specific readouts (multihead CL), the theory identifies a phase transition where CL performance shifts dramatically as tasks become less similar, as measured by another task-similarity OP. While forgetting is relatively mild compared to single-head CL across all tasks, sufficiently low similarity leads to catastrophic anterograde interference, where the network retains old tasks and interpolates new training data perfectly but completely fails to generalize new learning. Our results delineate important factors affecting CL performance and offer theoretical insights into common heuristics for mitigation of forgetting.
- 6 February 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (6)
Top Cited PapersOpen AccessOrder parameters and phase transitions of continual learning in deep neural networks
Continual learning (CL) enables animals to learn new tasks without erasing prior knowledge. CL in artificial neural networks (NNs) is challenging due to catastrophic forgetting, where new learning degrades performance on older tasks. While various techniques exist to mitigate forgetting, theoretical insights into when and why CL fails in NNs are lacking. Here, we present a statistical-mechanics theory of CL in deep, wide NNs, which characterizes the network’s input–output mapping as it learns a sequence of tasks. It gives rise to order parameters (OPs) that capture how task relations and network architecture influence forgetting and anterograde interference, as verified by numerical evaluations. For networks with a shared readout for all tasks (single-head CL), the relevant-feature and rule similarity between tasks, respectively measured by two OPs, are sufficient to predict a wide range of CL behaviors on classic benchmark tasks. In addition, the theory predicts that increasing the network depth can effectively reduce interference between tasks, thereby lowering forgetting. For networks with task-specific readouts (multihead CL), the theory identifies a phase transition where CL performance shifts dramatically as tasks become less similar, as measured by another task-similarity OP. While forgetting is relatively mild compared to single-head CL across all tasks, sufficiently low similarity leads to catastrophic anterograde interference, where the network retains old tasks and interpolates new training data perfectly but completely fails to generalize new learning. Our results delineate important factors affecting CL performance and offer theoretical insights into common heuristics for mitigation of forgetting.
- Open Access
Graph embeddings for identifying symmetries in connectomes
In many circuit models of neural computation, synaptic connections between neurons are organized according to their tuning to the variables being processed. The connectivities of canonical neural network models of head direction, spatial navigation, and orientation selectivity obey this principle and contain symmetries related to the angular and spatial variables they operate on. We develop a graph embedding algorithm to identify such symmetries. The algorithm segregates structure related to cell type from that related to the symmetries we seek to identify, distinguishing it from standard embedding methods. Our method successfully identifies rotational and translational symmetries in heading direction and visual projection neuron circuits using a connectome of the adult {\em Drosophila} brain, and it also identifies a toroidal symmetry in a synthetic connectome of grid cells in the medial entorhinal cortex. Such embedding geometries reveal the latent variables that are processed by a neural circuit and which cell types are responsible for this computation.
Open AccessGraph embeddings for identifying symmetries in connectomes
In many circuit models of neural computation, synaptic connections between neurons are organized according to their tuning to the variables being processed. The connectivities of canonical neural network models of head direction, spatial navigation, and orientation selectivity obey this principle and contain symmetries related to the angular and spatial variables they operate on. We develop a graph embedding algorithm to identify such symmetries. The algorithm segregates structure related to cell type from that related to the symmetries we seek to identify, distinguishing it from standard embedding methods. Our method successfully identifies rotational and translational symmetries in heading direction and visual projection neuron circuits using a connectome of the adult {\em Drosophila} brain, and it also identifies a toroidal symmetry in a synthetic connectome of grid cells in the medial entorhinal cortex. Such embedding geometries reveal the latent variables that are processed by a neural circuit and which cell types are responsible for this computation.
- 12 December 2022
- Published by American Physical Society (APS) in Physical Review E
- Vol. 106 (6) , 064406
Minimum perturbation theory of deep perceptual learning
Perceptual learning (PL) involves long-lasting improvement in perceptual tasks following extensive training and is accompanied by modified neuronal responses in sensory cortical areas in the brain. Understanding the dynamics of PL and the resultant synaptic changes is important for causally connecting PL to the observed neural plasticity. This is theoretically challenging because learning-related changes are distributed across many stages of the sensory hierarchy. In this paper, we modeled the sensory hierarchy as a deep nonlinear neural network and studied PL of fine discrimination, a common and well-studied paradigm of PL. Using tools from statistical physics, we developed a mean-field theory of the network in the limit of a large number of neurons and large number of examples. Our theory suggests that, in this thermodynamic limit, the input-output function of the network can be exactly mapped to that of a deep network, allowing us to characterize the space of solutions for the task. Surprisingly, we found that modifying synaptic weights in the first layer of the hierarchy is both sufficient and necessary for PL. To address the degeneracy of the space of solutions, we postulate that PL dynamics are constrained by a normative minimum perturbation (MP) principle, which favors weight matrices with minimal changes relative to their prelearning values. Interestingly, MP plasticity induces changes to weights and neural representations in all layers of the network, except for the readout weight vector. While weight changes in higher layers are not necessary for learning, they help reduce overall perturbation to the network. In addition, such plasticity can be learned simply through slow learning. We further elucidate the properties of MP changes and compare them against experimental findings. Overall, our statistical mechanics theory of PL provides mechanistic and normative understanding of several important empirical findings of PL.- 12 December 2022
- Published by American Physical Society (APS) in Physical Review E
- Vol. 106 (6) , 064406
Minimum perturbation theory of deep perceptual learning
Perceptual learning (PL) involves long-lasting improvement in perceptual tasks following extensive training and is accompanied by modified neuronal responses in sensory cortical areas in the brain. Understanding the dynamics of PL and the resultant synaptic changes is important for causally connecting PL to the observed neural plasticity. This is theoretically challenging because learning-related changes are distributed across many stages of the sensory hierarchy. In this paper, we modeled the sensory hierarchy as a deep nonlinear neural network and studied PL of fine discrimination, a common and well-studied paradigm of PL. Using tools from statistical physics, we developed a mean-field theory of the network in the limit of a large number of neurons and large number of examples. Our theory suggests that, in this thermodynamic limit, the input-output function of the network can be exactly mapped to that of a deep network, allowing us to characterize the space of solutions for the task. Surprisingly, we found that modifying synaptic weights in the first layer of the hierarchy is both sufficient and necessary for PL. To address the degeneracy of the space of solutions, we postulate that PL dynamics are constrained by a normative minimum perturbation (MP) principle, which favors weight matrices with minimal changes relative to their prelearning values. Interestingly, MP plasticity induces changes to weights and neural representations in all layers of the network, except for the readout weight vector. While weight changes in higher layers are not necessary for learning, they help reduce overall perturbation to the network. In addition, such plasticity can be learned simply through slow learning. We further elucidate the properties of MP changes and compare them against experimental findings. Overall, our statistical mechanics theory of PL provides mechanistic and normative understanding of several important empirical findings of PL.- Open Access
Helping can be driven by non-affective cues in rat
Helping another in distress can be motivated by either affective or cognitive empathy, with the latter commonly believed to be restricted to humans and possibly other apes. Here, we found evidence for rodent helping that occurs in the absence of affective cues. We employed a paradigm in which a free rat can open the door to a restrainer containing a trapped rat. When the trapped rat was treated with the anxiolytic midazolam, the helping behavior exhibited by the free rat was diminished but did not extinguish. Correspondingly, midazolam-treated trapped rats still released themselves when given the opportunity, albeit at longer latencies than controls, evidence that midazolam only partially reduced the distress experienced by trapped rats. To test whether helping could occur for a rat who exhibited no affect, trapped rats were immobilized by general anesthesia or heavy sedation. Surprisingly, rats opened the door to restrainers containing these immobilized rats and pulled the incapacitated rats out of the restrainer, pushing them away from the arena center. The same solicitous behavior was observed when an anesthetized rat was simply placed in the center of the arena, without being confined within a restrainer. We speculate that the cognitive dissonance of immobile rats, at odds with predictive expectations of rat behavior built up over a lifetime, motivated solicitous behavior including helping. To block affective behavioral displays without associated dissonant cues of immobility, metyrapone, a drug that selectively blocks corticosterone synthesis, was administered to trapped rats. Under such circumstances, little helping behavior occurred. In sum, rats may be motivated either by affect or by cognitive dissonance, the latter comprising a rudimentary form of cognitive empathy.
Open AccessHelping can be driven by non-affective cues in rat
Helping another in distress can be motivated by either affective or cognitive empathy, with the latter commonly believed to be restricted to humans and possibly other apes. Here, we found evidence for rodent helping that occurs in the absence of affective cues. We employed a paradigm in which a free rat can open the door to a restrainer containing a trapped rat. When the trapped rat was treated with the anxiolytic midazolam, the helping behavior exhibited by the free rat was diminished but did not extinguish. Correspondingly, midazolam-treated trapped rats still released themselves when given the opportunity, albeit at longer latencies than controls, evidence that midazolam only partially reduced the distress experienced by trapped rats. To test whether helping could occur for a rat who exhibited no affect, trapped rats were immobilized by general anesthesia or heavy sedation. Surprisingly, rats opened the door to restrainers containing these immobilized rats and pulled the incapacitated rats out of the restrainer, pushing them away from the arena center. The same solicitous behavior was observed when an anesthetized rat was simply placed in the center of the arena, without being confined within a restrainer. We speculate that the cognitive dissonance of immobile rats, at odds with predictive expectations of rat behavior built up over a lifetime, motivated solicitous behavior including helping. To block affective behavioral displays without associated dissonant cues of immobility, metyrapone, a drug that selectively blocks corticosterone synthesis, was administered to trapped rats. Under such circumstances, little helping behavior occurred. In sum, rats may be motivated either by affect or by cognitive dissonance, the latter comprising a rudimentary form of cognitive empathy.
- Open Access
A Theory of Neural Tangent Kernel Alignment and Its Influence on Training
The training dynamics and generalization properties of neural networks (NN) can be precisely characterized in function space via the neural tangent kernel (NTK). Structural changes to the NTK during training reflect feature learning and underlie the superior performance of networks outside of the static kernel regime. In this work, we seek to theoretically understand kernel alignment, a prominent and ubiquitous structural change that aligns the NTK with the target function. We first study a toy model of kernel evolution in which the NTK evolves to accelerate training and show that alignment naturally emerges from this demand. We then study alignment mechanism in deep linear networks and two layer ReLU networks. These theories provide good qualitative descriptions of kernel alignment and specialization in practical networks and identify factors in network architecture and data structure that drive kernel alignment. In nonlinear networks with multiple outputs, we identify the phenomenon of kernel specialization, where the kernel function for each output head preferentially aligns to its own target function. Together, our results provide a mechanistic explanation of how kernel alignment emerges during NN training and a normative explanation of how it benefits training.
Open AccessA Theory of Neural Tangent Kernel Alignment and Its Influence on Training
The training dynamics and generalization properties of neural networks (NN) can be precisely characterized in function space via the neural tangent kernel (NTK). Structural changes to the NTK during training reflect feature learning and underlie the superior performance of networks outside of the static kernel regime. In this work, we seek to theoretically understand kernel alignment, a prominent and ubiquitous structural change that aligns the NTK with the target function. We first study a toy model of kernel evolution in which the NTK evolves to accelerate training and show that alignment naturally emerges from this demand. We then study alignment mechanism in deep linear networks and two layer ReLU networks. These theories provide good qualitative descriptions of kernel alignment and specialization in practical networks and identify factors in network architecture and data structure that drive kernel alignment. In nonlinear networks with multiple outputs, we identify the phenomenon of kernel specialization, where the kernel function for each output head preferentially aligns to its own target function. Together, our results provide a mechanistic explanation of how kernel alignment emerges during NN training and a normative explanation of how it benefits training.
- Open Access
A Minimum Perturbation Theory of Deep Perceptual Learning
Perceptual learning (PL) involves long-lasting improvement in perceptual tasks following extensive training and is accompanied by modified neuronal responses in sensory cortical areas in the brain. Understanding the dynamics of PL and the resultant synaptic changes is important for causally connecting PL to the observed neural plasticity. This is theoretically challenging because learning-related changes are distributed across many stages of the sensory hierarchy. In this work, we modeled the sensory hierarchy as a deep nonlinear neural network and studied PL of fine discrimination, a common and well-studied paradigm of PL. Using tools from statistical physics, we developed a mean-field theory of the network in the limit of large number of neurons and large number of examples. Our theory suggests that, in this “thermodynamic” limit, the input-output function of the network can be exactly mapped to that of a deep linear network, allowing us to characterize the space of solutions for the task. Surprisingly, we found that modifications of synaptic weights in lower levels of the hierarchy are both sufficient and necessary for PL. To address the degeneracy of the space of solutions, we postulate that PL dynamics are constrained by a normative “minimum perturbation” (MP) principle, which favors weight matrices with minimal changes relative to their pre-learning values. Interestingly, MP plasticity induces changes to weights and neural representations in all layers of the network, except for the readout weight vector. In addition, such plasticity can be learned simply through slow learning. We further elucidate the properties of MP changes and compare them against experimental findings. Overall, our statistical mechanics theory of PL provides mechanistic and normative understanding of several important empirical findings of PL.
Open AccessA Minimum Perturbation Theory of Deep Perceptual Learning
Perceptual learning (PL) involves long-lasting improvement in perceptual tasks following extensive training and is accompanied by modified neuronal responses in sensory cortical areas in the brain. Understanding the dynamics of PL and the resultant synaptic changes is important for causally connecting PL to the observed neural plasticity. This is theoretically challenging because learning-related changes are distributed across many stages of the sensory hierarchy. In this work, we modeled the sensory hierarchy as a deep nonlinear neural network and studied PL of fine discrimination, a common and well-studied paradigm of PL. Using tools from statistical physics, we developed a mean-field theory of the network in the limit of large number of neurons and large number of examples. Our theory suggests that, in this “thermodynamic” limit, the input-output function of the network can be exactly mapped to that of a deep linear network, allowing us to characterize the space of solutions for the task. Surprisingly, we found that modifications of synaptic weights in lower levels of the hierarchy are both sufficient and necessary for PL. To address the degeneracy of the space of solutions, we postulate that PL dynamics are constrained by a normative “minimum perturbation” (MP) principle, which favors weight matrices with minimal changes relative to their pre-learning values. Interestingly, MP plasticity induces changes to weights and neural representations in all layers of the network, except for the readout weight vector. In addition, such plasticity can be learned simply through slow learning. We further elucidate the properties of MP changes and compare them against experimental findings. Overall, our statistical mechanics theory of PL provides mechanistic and normative understanding of several important empirical findings of PL.
- Open Access
Unsupervised identification of rat behavioral motifs across timescales
Behaviors of several laboratory animals can be modeled as sequences of stereotyped behaviors, or behavioral motifs. However, identifying such motifs is a challenging problem. Behaviors have a multi-scale structure: the animal can be simultaneously performing a small-scale motif and a large-scale one (e.g. \textit{chewing} and \textit{feeding}). Motifs are compositional: a large-scale motif is a chain of smaller-scale ones, folded in (some behavioral) space in a specific manner. We demonstrate an approach which captures these structures, using rat locomotor data as an example. From the same dataset, we used a preprocessing procedure to create different versions, each describing motifs of a different scale. We then trained several Hidden Markov Models (HMMs) in parallel, one for each dataset version. This approach essentially forced each HMM to learn motifs on a different scale, allowing us to capture behavioral structures lost in previous approaches. By comparing HMMs with models representing different null hypotheses, we found that rat locomotion was composed of distinct motifs from second scale to minute scale. We found that transitions between motifs were modulated by rats' location in the environment, leading to non-Markovian transitions. To test the ethological relevance of motifs we discovered, we compared their usage between rats with differences in a high-level trait, prosociality. We found that these rats had distinct motif repertoires, suggesting that motif usage statistics can be used to infer internal states of rats. Our method is therefore an efficient way to discover multi-scale, compositional structures in animal behaviors. It may also be applied as a sensitive assay for internal states.
Open AccessUnsupervised identification of rat behavioral motifs across timescales
Behaviors of several laboratory animals can be modeled as sequences of stereotyped behaviors, or behavioral motifs. However, identifying such motifs is a challenging problem. Behaviors have a multi-scale structure: the animal can be simultaneously performing a small-scale motif and a large-scale one (e.g. \textit{chewing} and \textit{feeding}). Motifs are compositional: a large-scale motif is a chain of smaller-scale ones, folded in (some behavioral) space in a specific manner. We demonstrate an approach which captures these structures, using rat locomotor data as an example. From the same dataset, we used a preprocessing procedure to create different versions, each describing motifs of a different scale. We then trained several Hidden Markov Models (HMMs) in parallel, one for each dataset version. This approach essentially forced each HMM to learn motifs on a different scale, allowing us to capture behavioral structures lost in previous approaches. By comparing HMMs with models representing different null hypotheses, we found that rat locomotion was composed of distinct motifs from second scale to minute scale. We found that transitions between motifs were modulated by rats' location in the environment, leading to non-Markovian transitions. To test the ethological relevance of motifs we discovered, we compared their usage between rats with differences in a high-level trait, prosociality. We found that these rats had distinct motif repertoires, suggesting that motif usage statistics can be used to infer internal states of rats. Our method is therefore an efficient way to discover multi-scale, compositional structures in animal behaviors. It may also be applied as a sensitive assay for internal states.
- 20 September 2019
- Published by American Physical Society (APS) in Physical Review E
- Vol. 100 (3) , 032132
Family of closed-form solutions for two-dimensional correlated diffusion processes
Diffusion processes with boundaries are models of transport phenomena with wide applicability across many fields. These processes are described by their probability density functions (PDFs), which often obey Fokker-Planck equations (FPEs). While obtaining analytical solutions is often possible in the absence of boundaries, obtaining closed-form solutions to the FPE is more challenging once absorbing boundaries are present. As a result, analyses of these processes have largely relied on approximations or direct simulations. In this paper, we studied two-dimensional, time-homogeneous, spatially correlated diffusion with linear, axis-aligned, absorbing boundaries. Our main result is the explicit construction of a full family of closed-form solutions for their PDFs using the method of images. We found that such solutions can be built if and only if the correlation coefficient between the two diffusing processes takes one of a numerable set of values. Using a geometric argument, we derived the complete set of 's where such solutions can be found. Solvable 's are given by , where . Solutions were validated in simulations. Qualitative behaviors of the process appear to vary smoothly over , allowing extrapolation from our solutions to cases with unsolvable 's.- 20 September 2019
- Published by American Physical Society (APS) in Physical Review E
- Vol. 100 (3) , 032132
Family of closed-form solutions for two-dimensional correlated diffusion processes
Diffusion processes with boundaries are models of transport phenomena with wide applicability across many fields. These processes are described by their probability density functions (PDFs), which often obey Fokker-Planck equations (FPEs). While obtaining analytical solutions is often possible in the absence of boundaries, obtaining closed-form solutions to the FPE is more challenging once absorbing boundaries are present. As a result, analyses of these processes have largely relied on approximations or direct simulations. In this paper, we studied two-dimensional, time-homogeneous, spatially correlated diffusion with linear, axis-aligned, absorbing boundaries. Our main result is the explicit construction of a full family of closed-form solutions for their PDFs using the method of images. We found that such solutions can be built if and only if the correlation coefficient between the two diffusing processes takes one of a numerable set of values. Using a geometric argument, we derived the complete set of 's where such solutions can be found. Solvable 's are given by , where . Solutions were validated in simulations. Qualitative behaviors of the process appear to vary smoothly over , allowing extrapolation from our solutions to cases with unsolvable 's.- Open Access
A rodent model of social rejection
Rats help conspecifics by releasing them from a restrainer only when they have previous social experience with the strain of the restrained rat. When rats have been cross-fostered with rats of a different strain than their own since birth and have never interacted with rats of their own strain, they do not help rats of their own strain when tested as adults. Here we interrogated whether a cross-fostered rat expressed his lack of motivation to help through any behaviors beyond not-helping. Accordingly, a cross-fostered rat was placed in an arena with a trapped rat and the door to the centrally located restrainer was taped shut. We found that cross-fostered rats moved more slowly and approached the trapped rat less than did control, regularly-raised rats tested under the same conditions. We then asked whether the behavior of the cross-fostered rats influenced the trapped rat. After being restrained with cross-fostered rats, trapped rats showed a decrease in exploratory behavior in an open field test compared to trapped rats who were raised normally. The same decrease in movement was observed after subject rats were allowed to freely interact with cross-fostered rats. These results suggest that rats that do not help demonstrate their disinterest to a trapped rat and that trapped rats exposed to apathetic rats show behavior suggestive of an increase in anxiety. In sum, the paradigm introduced here could serve as a rodent model for social rejection.
Open AccessA rodent model of social rejection
Rats help conspecifics by releasing them from a restrainer only when they have previous social experience with the strain of the restrained rat. When rats have been cross-fostered with rats of a different strain than their own since birth and have never interacted with rats of their own strain, they do not help rats of their own strain when tested as adults. Here we interrogated whether a cross-fostered rat expressed his lack of motivation to help through any behaviors beyond not-helping. Accordingly, a cross-fostered rat was placed in an arena with a trapped rat and the door to the centrally located restrainer was taped shut. We found that cross-fostered rats moved more slowly and approached the trapped rat less than did control, regularly-raised rats tested under the same conditions. We then asked whether the behavior of the cross-fostered rats influenced the trapped rat. After being restrained with cross-fostered rats, trapped rats showed a decrease in exploratory behavior in an open field test compared to trapped rats who were raised normally. The same decrease in movement was observed after subject rats were allowed to freely interact with cross-fostered rats. These results suggest that rats that do not help demonstrate their disinterest to a trapped rat and that trapped rats exposed to apathetic rats show behavior suggestive of an increase in anxiety. In sum, the paradigm introduced here could serve as a rodent model for social rejection.
- Open Access
- 8 June 2016
- Published by Frontiers Media SA in Frontiers in Psychology
- Vol. 7, 850
Anxiolytic Treatment Impairs Helping Behavior in Rats
Despite decades of research with humans, the biological mechanisms that motivate an individual to help others remain poorly understood. In order to investigate the roots of pro-sociality in mammals, we established the helping behavior test, a paradigm in which rats are faced with a conspecific trapped in a restrainer that can only be opened from the outside. Over the course of repeated test sessions, rats exposed to a trapped cagemate learn to open the door to the restrainer, thereby helping the trapped rat to escape (Ben-Ami Bartal et al., 2011). The discovery of this natural behavior provides a unique opportunity to probe the motivation of rodent helping behavior, leading to a deeper understanding of biological influences on human pro-sociality. To determine if an affective response motivates door-opening, rats received midazolam, a benzodiazepine anxiolytic, and were tested in the helping behavior test. Midazolam-treated rats showed less helping behavior than saline-treated rats or rats receiving no injection. Yet, midazolam-treated rats opened a restrainer containing chocolate, highlighting the socially specific effects of the anxiolytic. To determine if midazolam interferes with helping through a sympatholytic effect, the peripherally restricted beta-adrenergic receptor antagonist nadolol was administered; nadolol did not interfere with helping. The corticosterone response of rats exposed to a trapped cagemate was measured and compared to the rats’ subsequent helping behavior. Rats with the greatest corticosterone responses showed the least helping behavior and those with the smallest responses showed the most consistent helping at the shortest latency. These results are discussed in terms of their implications for the interaction between stress and pro-social behavior. Finally, we observed that door-opening appeared to be reinforcing. A novel analytical tool was designed to interrogate the pattern of door-opening for signs that a rat’s behavior on one session influenced his behavior on the next session. Results suggest that helping a trapped rat has a greater motivational value than does chocolate. In sum, this series of experiments clearly demonstrates the fundamental role of affect in motivating pro-social behavior in rodents and the need for a helper to resonate with the affect of a victim.
- 8 June 2016
- Published by Frontiers Media SA in Frontiers in Psychology
- Vol. 7, 850
Open AccessAnxiolytic Treatment Impairs Helping Behavior in Rats
Despite decades of research with humans, the biological mechanisms that motivate an individual to help others remain poorly understood. In order to investigate the roots of pro-sociality in mammals, we established the helping behavior test, a paradigm in which rats are faced with a conspecific trapped in a restrainer that can only be opened from the outside. Over the course of repeated test sessions, rats exposed to a trapped cagemate learn to open the door to the restrainer, thereby helping the trapped rat to escape (Ben-Ami Bartal et al., 2011). The discovery of this natural behavior provides a unique opportunity to probe the motivation of rodent helping behavior, leading to a deeper understanding of biological influences on human pro-sociality. To determine if an affective response motivates door-opening, rats received midazolam, a benzodiazepine anxiolytic, and were tested in the helping behavior test. Midazolam-treated rats showed less helping behavior than saline-treated rats or rats receiving no injection. Yet, midazolam-treated rats opened a restrainer containing chocolate, highlighting the socially specific effects of the anxiolytic. To determine if midazolam interferes with helping through a sympatholytic effect, the peripherally restricted beta-adrenergic receptor antagonist nadolol was administered; nadolol did not interfere with helping. The corticosterone response of rats exposed to a trapped cagemate was measured and compared to the rats’ subsequent helping behavior. Rats with the greatest corticosterone responses showed the least helping behavior and those with the smallest responses showed the most consistent helping at the shortest latency. These results are discussed in terms of their implications for the interaction between stress and pro-social behavior. Finally, we observed that door-opening appeared to be reinforcing. A novel analytical tool was designed to interrogate the pattern of door-opening for signs that a rat’s behavior on one session influenced his behavior on the next session. Results suggest that helping a trapped rat has a greater motivational value than does chocolate. In sum, this series of experiments clearly demonstrates the fundamental role of affect in motivating pro-social behavior in rodents and the need for a helper to resonate with the affect of a victim.