About

John Pearson is an Associate Professor of Neurobiology at Duke University, with additional roles as an Assistant Research Professor in Neurobiology, an Assistant Professor in the Department of Electrical and Computer Engineering, and an Associate Professor of Psychology and Neuroscience. He is a member of the Center for Cognitive Neuroscience. His research and academic activities are based in the Bryan Research Building in Durham, North Carolina. Pearson's professional focus involves neurobiological studies, and he is actively engaged in teaching and research within these interdisciplinary fields.

Research topics

• Artificial Intelligence
• Computer Science
• Psychology
• Neuroscience
• Machine Learning
• Biology
• Cognitive science
• Cognitive psychology

Selected publications

• Flow Matching for Count Data

  PubMed Central · 2026-05-08

  preprintOpen accessSenior author

  High-dimensional count data arise in applications such as single-cell RNA sequencing and neural spike trains, where mapping between distributions across successive batches or time points form critical components of data analysis. The recent success of diffusion- and flow-based deep generative models for images, video, and text motivates extending these ideas to count-valued settings, but many existing methods either treat each count as a categorical state or transform counts into a continuous space, neither of which is natural or efficient when the count range is large. We propose count-FM, a flow-matching framework for count data based on a continuous-time birth-death process with local unit jumps. Count-FM learns marginal transitions efficiently in count space through simulation-free training of conditional transition rates, allowing transport between arbitrary count-distributed source and target populations. In simulation, count-FM achieves better sample quality than representative baselines while using substantially fewer parameters. We further apply count-FM to scRNA-seq and neural spike-train data for unconditional generation, transport, and conditional generation. Across these tasks, count-FM yields improved sample quality, greater modeling efficiency, and interpretable transport paths.

  PublisherDOI

• Dynamic Compression Flows for Neuroscience Data

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-02-13

  articleOpen accessSenior author

  While neuroscience experiments have repeatedly demonstrated the involvement of large populations of neurons in even simple behaviors, these studies have just as often reported that the collective dynamics of neural activity are approximately low-dimensional. As a result, methods for identifying low-dimensional latent representations of time series data have become increasingly prominent in neuroscience. However, most existing methods either ignore temporal structure or model time evolution using latent dynamical systems approaches. In the first case, dynamics may be distorted or even scrambled in the latent space, while in the second, many possible latent dynamics may give rise to the same data. Here, we address these challenges using a novel flow-matching approach in which data are generated by a pair of flow fields, one governing time evolution, the other a mapping between data and a low-dimensional latent space. Importantly, the dimension-reducing flow is trained to minimize distortions of the temporal dynamics, learning an identifiable low-dimensional representation that preserves temporal relations in the original data. Additionally, we constrain our latent spaces to have low-dimensional support in a soft, parameterized manner, taking inspiration from ideas on nested dropout. Across both neural and behavioral data, we show that this dual flow approach produces both more interpretable dynamics and higher-quality reconstructions than competing models, including in noise-dominated data sets where conventional approaches fail.

  PublisherDOI

• Flow Matching for Count Data

  ArXiv.org · 2026-05-08

  articleOpen accessSenior author

  High-dimensional count data arise in applications such as single-cell RNA sequencing and neural spike trains, where mapping between distributions across successive batches or time points form critical components of data analysis. The recent success of diffusion- and flow-based deep generative models for images, video, and text motivates extending these ideas to count-valued settings, but many existing methods either treat each count as a categorical state or transform counts into a continuous space, neither of which is natural or efficient when the count range is large. We propose count-FM, a flow-matching framework for count data based on a continuous-time birth-death process with local unit jumps. Count-FM learns marginal transitions efficiently in count space through simulation-free training of conditional transition rates, allowing transport between arbitrary count-distributed source and target populations. In simulation, count-FM achieves better sample quality than representative baselines while using substantially fewer parameters. We further apply count-FM to scRNA-seq and neural spike-train data for unconditional generation, transport, and conditional generation. Across these tasks, count-FM yields improved sample quality, greater modeling efficiency, and interpretable transport paths.

  PublisherOA PDF

• Flow Matching for Count Data.

  PubMed · 2026-05-08

  articleSenior author

  High-dimensional count data arise in applications such as single-cell RNA sequencing and neural spike trains, where mapping between distributions across successive batches or time points form critical components of data analysis. The recent success of diffusion- and flow-based deep generative models for images, video, and text motivates extending these ideas to count-valued settings, but many existing methods either treat each count as a categorical state or transform counts into a continuous space, neither of which is natural or efficient when the count range is large. We propose count-FM, a flow-matching framework for count data based on a continuous-time birth-death process with local unit jumps. Count-FM learns marginal transitions efficiently in count space through simulation-free training of conditional transition rates, allowing transport between arbitrary count-distributed source and target populations. In simulation, count-FM achieves better sample quality than representative baselines while using substantially fewer parameters. We further apply count-FM to scRNA-seq and neural spike-train data for unconditional generation, transport, and conditional generation. Across these tasks, count-FM yields improved sample quality, greater modeling efficiency, and interpretable transport paths.

  Publisher

• JuliaStats/Distributions.jl: v0.25.123

  Zenodo (CERN European Organization for Nuclear Research) · 2026-01-04 · 1 citations

  otherOpen access

  Distributions v0.25.123 Diff since v0.25.122 Merged pull requests: Fix von Mises-Fisher sampler (#1930) (@devmotion) Add doctests (#1980) (@abhro) Make MatrixNormal sampling non-allocating (#2012) (@projekter) Update language tags and code output in starting.md (#2013) (@abhro) Update CI workflows for building docs (#2014) (@abhro) Fix syntax for @docs block in types.md (#2015) (@abhro) Bump actions/checkout from 5 to 6 (#2016) (@dependabot[bot]) Bump version from 0.25.122 to 0.25.123 (#2020) (@devmotion) Closed issues: Sampling from von Mises Fisher yields NaN (#1423) inv(Σ) fails with StaticArrays (#1826) Issue with Von Mises-Fisher sampler returning NaN points (#2018)

  PublisherDOI

• A synaptic locus of song learning

  Nature · 2026-05-13

  articleOpen access
  PublisherOA PDFDOI

• Representational similarity and pattern classification of fifteen emotional states induced by movie clips and text scenarios

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-10-17

  preprintOpen access

  Theoretical models emphasize that categorical factors, dimensional factors, or their combination may define the semantic space organization of emotion representations. While recent behavioral work has applied innovative multivariate methods for testing these theories, neuroscientific assessments remain limited due to a focus on a small number of emotions, single theoretical perspectives, univariate methods, or small sample sizes. We overcame these limitations in a comprehensive functional magnetic resonance imaging (fMRI) study in which participants (N = 136) viewed 150 movie clips and 150 text scenarios that reliably induced 15 different emotional states spanning positive, negative, and neutral valence. For the movie inductions, representational similarity analysis yielded a correspondence between the categorical behavioral responses and the brain activity patterns, and partial least squares discriminant analysis achieved strong decoding performance for all 15 emotions from whole-brain fMRI data, with importance maps encompassing cortical, limbic, and subcortical regions. Classification error analyses and a Bayesian model comparison supported the categorical nature of the emotion representations relative to a 2-dimensional arousal-valence model. Hierarchical clustering of the representational dissimilarity matrices revealed that the 15 emotions were organized into similarly meaningful clusters at both the subjective and neural levels. Results from the scenario inductions, while demonstrating similar behavioral effects and behavioral hierarchical structures, were more difficult to decode from the fMRI data. Overall, these findings provide novel insights into how emotions are organized and represented in the human brain and evidence a relationship between our subjective experience of and brain responses to emotional inductions. Significance Statement: We found a robust relationship between participants' categorical endorsement of and their neural responses to emotionally evocative movie clips using an unsupervised machine learning approach. Subsequently, we were also able to decode the emotional content of these movie clips using a supervised classification technique. Finally, we found that these emotions are organized in a very similar ways both in terms of participants' self-report and their brain responses to these stimuli. Together, these data-driven and computational modeling based findings from our study significantly advance our understanding of how emotions are organized and represented in our mind and brain.

  PublisherOA PDFDOI

• Modeling of control over task switching and cross-task interference supports a two-dimensional model of cognitive stability and flexibility

  Psychonomic Bulletin & Review · 2025-06-02

  reviewOpen access
  PublisherOA PDFDOI

• Selective encoding of priors for flexible categorization but not Bayesian inference in the frontal eye field

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-01-02

  preprintOpen access

  Abstract Expectations or prior beliefs about the world have been shown to modulate sensory processing both at the behavioral and neural levels. Bayesian models predict that such priors compensate for input uncertainty to optimize sensory judgments. Although Bayesian behavior is prevalent across sensorimotor systems, the relationship between priors and Bayesian inference is not obligatory. Priors may simply shift one’s internal decision boundaries without interacting with sensory uncertainty at all. We recently showed that humans and monkeys use both Bayesian and non-Bayesian strategies when reporting judgments of visual stability across saccades, despite using priors in both cases. While they increased prior use to compensate for internal, movement-driven sensory uncertainty in a Bayesian manner, they decreased prior use when faced with external, visual image uncertainty. The latter, “anti-Bayesian” pattern was best explained by a model in which category boundaries were adjusted by the prior but susceptible to image noise. Here, we recorded neural activity in the frontal eye field (FEF), a prefrontal region important for visuosaccadic behavior, while toggling between subjects’ prior use for Bayesian and anti-Bayesian behavior via trial-by-trial manipulation of the two uncertainty conditions. First, we found that FEF activity signaled the priors in both conditions. The prior-related modulation of activity, however, predicted only the anti-Bayesian, categorization behavior. The results suggest that neural activity in the FEF reflects the use of a flexible decision boundary for the perception of visual stability and, more generally, that neural mechanisms for Bayesian inference and visual categorization are dissociable and distributed in the primate brain. Significance Appreciating a visual scene depends not only on retinal input, but also on priors about the world. Foreknowledge interacts with visual inputs to improve reactions and decisions. One way the brain combines priors and inputs is by using Bayes’ rule to model optimal outcomes. A simpler way is by categorizing inputs with prior-adjusted boundaries. Here, we tested how neurons in primate frontal cortex use priors: for Bayes’ rule, or for flexible categorization? A key feature of the study was to use a single perceptual task that was varied trial-by-trial to yield either Bayesian or categorization behaviors. We could then establish which behavior the neurons encoded. The implications extend beyond visuomotor behavior to broader neurocomputational mechanisms of prior use for cognition.

  PublisherDOI

• Dual neuromodulatory dynamics underlie birdsong learning

  Nature · 2025-03-12 · 8 citations

  articleOpen accessCorresponding
  PublisherOA PDFDOI

Recent grants

• Nonparametric Bayes Methods for Big Data in Neuroscience

  NIH · $152k · 2014–2019

• NIH Grant R01AG005690

  NIH · $255k · 1989

• NIH Grant R21NS084176

  NIH · $432k · 2017

• Nonparametric Bayes Methods for Big Data in Neuroscience

  NIH · $580k · 2014–2019

Frequent coauthors

• N. Chamel

  Université Libre de Bruxelles

  71 shared

• Michael L. Platt

  University of Pennsylvania

  66 shared

• S. Goriely

  56 shared

• A. F. Fantina

  GANIL

  55 shared

• Menek Goldstein

  Karolinska Institutet

  44 shared

• Scott A. Huettel

  43 shared

• Leslie Brandeis

  36 shared

• Benjamin Y. Hayden

  Baylor College of Medicine

  34 shared

Labs

• Pearson LabPI

Education

• PhD, Physics

  Princeton University

  2004

• BS, Physics, Mathematics

  University of Kentucky

  1999