Biology subjects

Iigaya, K.

Publications and source records attributed to Iigaya, K..

8 recordsLinked to original sources

bioRxiv2026

Striatal acetylcholine enables latent-state creation during reversal learning

Like dopamine, acetylcholine is modulated in the striatum by reward-predicting cues and outcomes, yet its computational role remains unclear. Here, we manipulated dorsomedial striatal acetylcholine in mice performing a reversal learning task using genetic and physiologically guided optogenetic approaches. Inhibiting phasic acetylcholine modulation impaired reversal learning, while enhancing modulation during non-rewarded trials facilitated reversal learning. Trial-by-trial acetylcholine dynamics were best explained by a reinforcement learning model in which new latent states are created when experience is poorly explained by existing states. A circuit-constrained model further suggested that acetylcholine promotes plasticity when reward-omission is ambiguous, leading to a surprising prediction: making reward omission explicit should reduce the need for acetylcholine. We confirmed this in a new experiment in which an auditory omission cue substantially reduced the reversal-learning deficit caused by genetic acetylcholine knockdown. These findings suggest that striatal acetylcholine supports reversal learning by promoting state construction in reinforcement learning.

neuroscience↗

bioRxiv2026

Building Goal-Directed Cognitive Graphs

Flexible behavior requires building internal structured models from experience to support goal-directed action. Although environmental transition statistics accumulate gradually, how they are used to construct a compact goal-directed graph remains unclear. We introduce the Sparse Cognitive Graph (SCG), a reinforcement-learning framework that separates gradual transition learning from the sparse directed graph that governs valuation and action selection. In the SCG, transition statistics accumulate in a dense predictive representation, while nonlinear selection determines which transitions are expressed as graph edges. Consequently, gradual strengthening can trigger discrete reorganization of graph topology and abrupt behavioral shifts. Across human reward and transition revaluation tasks, the SCG explains bimodal and trimodal behavioral regimes as emergent consequences of distinct graph configurations. Across human and mouse two-step tasks, dynamic graph reconfiguration captures canonical reward-by-transition interactions without requiring mixtures of control systems. In mice, transitions preceding reward strengthened more rapidly, biasing graph topology toward reward-directed paths. Temporally precise optogenetic dopamine stimulation produced behavioral effects consistent with accelerated graph edge formation predicted by the SCG. The model further generates a testable prediction: graph topology determines the geometry of low-dimensional population activity. Directed acyclic graphs yield activity concentrated at graph entry and goal states, whereas cyclic graphs produce periodic, grid-like structure. Together, these findings identify reward-dependent graph reorganization as a computational principle that reconciles stable predictive learning with efficient goal-directed control.

neuroscience↗

bioRxiv2025

The neural computations underlying context dependent attribute-based valuation of complex stimuli.

Adaptive decision making requires value computations to be flexible because we often need to value a stimulus differently as the context changes. A stimulus might be highly desirable in one context but completely unappealing in another. However, it is not known how the brain can support such flexible modulation of overall stimulus value. Here we test a model of flexible value construction whereby individual attributes of a stimulus are converted into contextually dependent attribute-specific value representations before being combined into an overall integrated stimulus value. To test this framework, human participants (online n=95, MRI n=35) provided ratings for 75 unique high-dimensional clothing stimuli under three instructed goal-contexts designed to elicit differences in overall value judgments for the items. Integrated value ratings for each stimulus were found to be goal-context dependent, while individual stimulus-attributes varied markedly in how they contributed to value ratings across goal contexts. In the fMRI data, representations of the absolute levels of particular attributes were revealed in visual areas. In contrast, encoding of individual attributes in value space, alongside integrated overall stimulus value, was present within distinct regions of prefrontal cortex. More specifically, behaviorally relevant attributes in value space and integrated stimulus value were found in vmPFC and dmPFC respectively. These findings indicate that the construction of value for high-dimensional stimuli is achieved through the computation of goal-context-dependent attributes in value space, providing mechanistic insight into how the brain can flexibly modulate stimulus-values as context changes.

neuroscience↗

bioRxiv2024

Computational and neural evidence for altered fast and slow learning from losses in gambling disorder

Learning occurs across multiple timescales, with fast learning crucial for adapting to sudden environmental changes, and slow learning beneficial for extracting robust knowledge from multiple events. Here we asked if miscalibrated fast vs slow learning can lead to maladaptive decision-making in individuals with gambling disorder. Participants with problem gambling and a recreational gambling control group without any symptoms associated with problem gambling performed a probabilistic decision-making task involving reward-learning and loss-avoidance while being scanned with fMRI. Problem gamblers showed impaired reversal learning from losses compared to the control group, with excessive dependence on slow timescales and reduced reliance on fast timescales. fMRI data implicated the putamen, an area associated with habit, and medial prefrontal cortex (PFC) in slow loss-value encoding, with significantly more robust encoding in medial PFC in problem gamblers compared to controls. Problem gamblers also exhibited stronger loss prediction error encoding in the insular cortex. These findings suggest that individuals with problem gambling have an impaired ability to adjust their predictions following losses, manifested by a stronger influence of slow value learning. This impairment could contribute to the behavioral inflexibility of problem gamblers, particularly the persistence in gambling behavior typically observed in those individuals after incurring loss outcomes.

neuroscience↗

bioRxiv2020

Aesthetic preference for art emerges from a weighted integration over hierarchically structured visual features in the brain

It is an open question whether preferences for visual art can be lawfully predicted from the basic constituent elements of a visual image. Moreover, little is known about how such preferences are actually constructed in the brain. Here we developed and tested a computational framework to gain an understanding of how the human brain constructs aesthetic value. We show that it is possible to explain human preferences for a piece of art based on an analysis of features present in the image. This was achieved by analyzing the visual properties of drawings and photographs by multiple means, ranging from image statistics extracted by computer vision tools, subjective human ratings about attributes, to a deep convolutional neural network. Crucially, it is possible to predict subjective value ratings not only within but also across individuals, speaking to the possibility that much of the variance in human visual preference is shared across individuals. Neuroimaging data revealed that preference computations occur in the brain by means of a graded hierarchical representation of lower and higher level features in the visual system. These features are in turn integrated to compute an overall subjective preference in the parietal and prefrontal cortex. Our findings suggest that rather than being idiosyncratic, human preferences for art can be explained at least in part as a product of a systematic neural integration over underlying visual features of an image. This work not only advances our understanding of the brain-wide computations underlying value construction but also brings new mechanistic insights to the study of visual aesthetics and art appreciation.

neuroscience↗

bioRxiv2019

Neuro-computational account of arbitration between imitation and emulation during human observational learning

In observational learning (OL), organisms learn from observing the behavior of others. There are at least two distinct strategies for OL. Imitation involves learning to repeat the previous actions of other agents, while in emulation, learning proceeds from inferring the goals and intentions of others. While putative neural correlates for these forms of learning have been identified, a fundamental question remains unaddressed: how does the brain decides which strategy to use in a given situation? Here we developed a novel computational model in which arbitration between the strategies is determined by the predictive reliability, such that control over behavior is adaptively weighted toward the strategy with the most reliable prediction. To test the theory, we designed a novel behavioral task in which our experimental manipulations produced dissociable effects on the reliability of the two strategies. Participants performed this task while undergoing fMRI in two independent studies (the second a pre-registered replication of the first). Behavior manifested patterns consistent with both emulation and imitation and flexibly changed between the two strategies as expected from the theory. Computational modelling revealed that behavior was best described by an arbitration model, in which the reliability of the emulation strategy determined the relative weights allocated to behavior for each strategy. Emulation reliability - the models arbitration signal - was encoded in the ventrolateral prefrontal cortex, temporoparietal junction and rostral cingulate cortex. Being replicated across two fMRI studies, these findings suggest a neuro-computational mechanism for allocating control between emulation and imitation during observational learning.

neuroscience↗

bioRxiv2019

Hippocampal-midbrain circuit enhances the pleasure of anticipation in the prefrontal cortex

Having something to look forward to is a keystone of well-being. Anticipation of a future reward, like an upcoming vacation, can often be more gratifying than the very experience itself. Theories of anticipation have described how it induces behaviors ranging from beneficial information-seeking through to harmful addiction. However, it remains unclear how neural systems compute an attractive value from anticipation, instead of from the reward itself. To address this gap, we administered a decision-making task to human participants that allowed us to analyze brain activity during receipt of information predictive of future pleasant outcomes. Using a computational model of anticipatory value that captures participants decisions, we show that an anticipatory value signal is orchestrated by influences from three brain regions. Ventromedial prefrontal cortex (vmPFC) tracks the value of anticipation; dopaminergic midbrain responds to information that enhances anticipation, while sustained hippocampal activity provides a functional coupling between these regions. This coordinating function of the hippocampus is consistent with its known role in episodic future thinking. Our findings shed new light on the neural underpinnings of anticipations influence over decision-making, while also unifying a range of phenomena associated with risk and time-delay preference.

neuroscience↗

bioRxiv2017

The Long and the Short of Serotonergic Stimulation: Optogenetic activation of dorsal raphe serotonergic neurons changes the learning rate for rewards

Serotonin plays an influential, but computationally obscure, modulatory role in many aspects of normal and dysfunctional learning and cognition. Here, we studied the impact of optogenetic stimulation of dorsal raphe serotonin neurons in mice performing a non-stationary, reward-driven, foraging task. We report that activation of serotonin neurons significantly boosted learning rates for choices following long inter-trial-intervals that were driven by the recent history of reinforcement.

neuroscience↗