About

Naoshige Uchida is the Jeff C. Tarr Professor of Molecular and Cellular Biology at Harvard University, affiliated with the Department of Molecular & Cellular Biology. His laboratory focuses on neuronal processes by which sensory information and memory about previous experiences guide animal behavior. His research investigates how odor information is coded and processed by neuronal ensembles, the circuit dynamics underlying decision-making processes, and the mechanisms for learning based on rewards and punishments. Uchida has developed behavioral paradigms such as odor-guided perceptual decision tasks in rats and mice, which are combined with multi-electrode recording techniques to monitor neuronal activity during behavior. His work emphasizes establishing causal links between specific neuronal circuit activity and the dynamics of behavior and learning, utilizing molecular and genetic tools available in mice to test hypotheses experimentally.

Research topics

• Computer Science
• Neuroscience
• Psychology
• Artificial Intelligence
• Cognitive psychology
• Machine Learning
• Social psychology
• Statistics
• Biology
• Mathematics
• Physics

Selected publications

• Phasic dopamine drives conditioned responding beyond its role in learning

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-03-25

  articleOpen access

  Animals exposed to pairings of a neutral stimulus with reward acquire a conditioned response to the neutral stimulus. A prominent hypothesis, formalized in the Temporal Difference (TD) learning algorithm, is that animals learn to predict the future reward associated with the neutral stimulus ("value"). Though the TD algorithm does not explicitly specify what drives conditioned responding, a typical assumption is that it reflects the animal's estimate of value. In TD learning, value estimates are updated using reward prediction error (RPE, the discrepancy between observed and predicted reward), and are thought to be signaled by the phasic activity of midbrain dopamine neurons. This hypothesis posits that dopamine's effects on conditioned responding are mediated entirely by its effects on learning. However, recent experimental and theoretical evidence suggests that dopamine may play a more direct role in modulating conditioned responding. We use a combination of data analysis and computational modeling to probe the relationship between dopamine and conditioned responding. Our results suggest that dopamine directly modulates conditioned responding, in addition to its role in learning. These findings can be captured by a model in which dopamine RPE acts both indirectly (via learning) and directly on conditioned responding.

  PublisherOA PDFDOI

• Dopamine in the ventral and tail of striatum supports global and local evaluation in reward-threat conflict

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-05-05

  articleOpen access

  Survival requires balancing reward seeking and threat avoidance, yet how distinct dopamine systems coordinate to support this remains unclear. Using a naturalistic foraging paradigm in which mice pursue water reward under threat from a monster object, we examined roles of dopamine projections to the ventral striatum (VS) and tail of the striatum (TS). Ablation of VS- projecting dopamine neurons impaired both distal reward pursuit and threat avoidance, with the impairment in threat avoidance paralleling effects of TS dopamine ablation. However, simultaneous recordings revealed different activity rules: VS dopamine tracked radial velocity towards the current goal as animals changed goals (reward or shelter), consistent with a temporal-difference error of spatial value, while TS dopamine encoded proximity and orientation to the threat, reflecting immediate sensory experience. Taken together, VS and TS dopamine evaluates distinct state information for avoidance. VS dopamine facilitates allocentric, goal- directed navigation, while TS dopamine facilitates egocentric, stimulus-driven threat responses.

  PublisherDOI

• Prospective contingency explains behavior and dopamine signals during associative learning

  Nature Neuroscience · 2025-03-18 · 20 citations

  articleSenior authorCorresponding
  PublisherDOI

• Emergence of rapid value inference through meta-reinforcement learning

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-12-02

  articleOpen accessSenior authorCorresponding

  The ability to estimate the value associated with a specific stimulus or action is essential for adaptive behavior. Value can be updated either incrementally through experience or rapidly by inference based on latent environmental structure. Yet, how the brain implements and transitions between these modes of value computation remains unclear. To address this question, we examined the neuronal mechanisms underlying reversal learning. Mice were trained in an odor-outcome association task either with stable or dynamically changing contingencies. Mice trained on stable contingencies formed long-term value representations that depended on synaptic plasticity in the basolateral amygdala (BLA). In contrast, mice exposed to repeated reversals acquired the ability to infer values, independent from plasticity in BLA, enabling faster learning but with more rapid memory decay. Recurrent neural network models (RNNs) trained with continuous weight updates recapitulated this transition, shifting from plasticity-based to dynamics-based value computation. Neural activity in the BLA encoded both value and contextual information necessary for computing value based on latent task structure, similar to those found in the RNNs. Disrupting BLA activity before cue delivery preferentially impaired dynamics-based value updating. Furthermore, mice could learn distinct correlation structures that enabled structure-specific value inference. Together, these findings provide a mechanistic framework for fast value updates via inference, a core feature of intelligent behavior.

  PublisherDOI

• Multi-timescale reinforcement learning in the brain

  Nature · 2025-06-04 · 12 citations

  articleSenior author
  PublisherDOI

• An opponent striatal circuit for distributional reinforcement learning

  Nature · 2025-02-19 · 19 citations

  articleOpen accessSenior author
  PublisherOA PDFDOI

• A hypothalamic circuit underlying the dynamic control of social homeostasis

  Nature · 2025-02-26 · 43 citations

  articleOpen access

  Abstract Social grouping increases survival in many species, including humans 1,2 . By contrast, social isolation generates an aversive state (‘loneliness’) that motivates social seeking and heightens social interaction upon reunion 3–5 . The observed rebound in social interaction triggered by isolation suggests a homeostatic process underlying the control of social need, similar to physiological drives such as hunger, thirst or sleep 3,6 . In this study, we assessed social responses in several mouse strains, among which FVB/NJ mice emerged as highly, and C57BL/6J mice as moderately, sensitive to social isolation. Using both strains, we uncovered two previously uncharacterized neuronal populations in the hypothalamic preoptic nucleus that are activated during either social isolation or social rebound and orchestrate the behaviour display of social need and social satiety, respectively. We identified direct connectivity between these two populations and with brain areas associated with social behaviour, emotional state, reward and physiological needs and showed that mice require touch to assess the presence of others and fulfil their social need. These data show a brain-wide neural system underlying social homeostasis and provide significant mechanistic insights into the nature and function of circuits controlling instinctive social need and for the understanding of healthy and diseased brain states associated with social context.

  PublisherOA PDFDOI

• Hunger modulates exploration through suppression of dopamine signaling in the tail of the striatum

  Neuron · 2025-10-04 · 2 citations

  articleOpen access

  Caloric depletion induces behavioral changes that help an animal find food and restore its homeostatic balance. Hunger increases exploration and risk-taking behavior, allowing an animal to forage for food despite risks; however, it is unknown which neural systems coordinate such behavioral adaptations. Here, we characterize how hunger restructures an animal's spontaneous behavior as well as its directed exploration of a novel object. We show that hunger-induced changes in exploration are accompanied by and result from the modulation of dopamine signaling in the tail of the striatum (TOS). Dopamine signaling in the TOS is in turn modulated by hunger through the activity of agouti-related peptide (AgRP) neurons, putative "hunger neurons" in the arcuate nucleus of the hypothalamus that are polysynaptically connected to the TOS through the lateral hypothalamus. Thus, we delineate how hypothalamic systems modulate dopaminergic circuitry to mediate changes in exploratory behavior in the hungry state.

  PublisherOA PDFDOI

• eLife Assessment: Basal ganglia output coding - entopeduncular nucleus - of contextual kinematics and reward in the freely moving mouse

  2025-01-07

  peer-reviewOpen access1st authorCorresponding

  The entopeduncular nucleus (EPN) is often termed as one of the output nuclei of the basal ganglia owing to their highly convergent anatomy. The rodent EPN has been implicated in reward and value coding whereas the primate analogue internal Globus Pallidus has been found to be modulated by some movements and in some circumstances. In this study we sought to understand how the rodent EPN might be coding kinematic, reward, and difficulty parameters, particularly during locomotion. Furthermore, we aimed to understand the level of movement representation: whole-body or specific body parts. To this end, mice were trained in a freely moving two-alternative forced choice task with two periods of displacement (return and go trajectories) and performed electrophysiological recordings together with video-based tracking. We found 1) robust reward coding but not difficulty. 2) Spatio-temporal variables better explain EPN activity during movement compared to kinematic variables, while both types of variables were more robustly represented in reward-related movement. 3) Reward sensitive units encode kinematics similarly to reward insensitive ones. 4) Population dynamics that best account for differences between these two periods of movement can be explained by allocentric references like distance to reward port. 5) The representation of paw and licks is not mutually exclusive, discarding a somatotopic muscle-level representation of movement in the EPN. Our data suggest that EPN activity represents movements and reward in a complex way: highly multiplexed, influenced by the objective of the displacement, where trajectories that lead to reward better represent spatial and kinematic variables. Interestingly, there are intertwining representations of whole-body movement kinematics with single paw and licking variables. Further, reward sensitive units encode kinematics similarly to reward insensitive ones, challenging the notion of distinct pathways for reward and movement processing.The entopeduncular nucleus is one of the main outputs of the basal ganglia whose activity has been hypothesized to be inversely correlated with movement. This study examines motor and reward coding simultaneously, finding that besides the great level of multiplexing of these variables, spatio-temporal coding is better represented than kinematic coding. The level of movement representation seems to be greatly influenced by the goal of a movement, with spatially biased variables influencing the population dynamics of this nucleus. Further, we uncover the coexistence of EPN modulation by movement at different timescales and body parts. The simple overall activity of this output nucleus cannot explain kinematic coding, challenging leading theories of basal ganglia function.

  PublisherOA PDFDOI

• eLife Assessment: Striatal Crosstalk Between Dopamine and Serotonin Systems

  2025-07-16

  peer-reviewOpen access1st authorCorresponding

  Dopamine (DA) and serotonin (5-HT) are neuromodulators in reward processing, decision- making, and motivated behavior. While often viewed as opposing or complementary systems, how DA and 5-HT release integrate in the striatum remains elusive. Using optogenetics, fiber photometry, and slice electrophysiology, we found that ventral tegmental area (VTA) DA neuron stimulation increased DA release without affecting 5-HT release. Dorsal raphe nucleus (DRN) 5-HT neuron activation, on the other hand, induced serotonin release and a transient increase in DA in the NAc, likely via glutamate co-release onto VTA DA neurons during the initial stimulation phase. These findings indicate that DA and 5-HT operate largely independently in the striatum, with selective circuit-dependent interactions. This work refines our understanding of DA-5HT interactions and provides a foundation for future research into their roles in motivated behaviors and neuropsychiatric disorders.

  PublisherDOI

Recent grants

• Towards a Unified Framework for Dopamine Signaling in the Striatum

  NIH · $32.4M · 2019–2025

• Distributional Reinforcement Learning in the Brain

  NIH · $2.9M · 2020–2026

• Experimental examinations of the mechanisms that generate the responses of midbra

  NIH · $2.1M · 2013–2018

• Towards a Unified Framework for Dopamine Signaling in the Striatum

  NIH · $7.2M · 2019–2024

• Neural circuits that regulate dopamine neuron activity

  NIH · $2.1M · 2012–2017

Frequent coauthors

• Mitsuko Watabe‐Uchida

  Harvard University

  41 shared

• HyungGoo R. Kim

  Sungkyunkwan University

  20 shared

• Masatoshi Takeichi

  RIKEN Center for Biosystems Dynamics Research

  20 shared

• Zachary F. Mainen

  Champalimaud Foundation

  17 shared

• Samuel J. Gershman

  Harvard University

  15 shared

• Athar N. Malik

  Brown University

  15 shared

• Scott W. Linderman

  14 shared

• Bénédicte M. Babayan

  Harvard University

  12 shared

Labs

• Naoshige UchidaPI