About

Chandramouli Chandrasekaran is an Assistant Professor in the Department of Anatomy & Neurobiology and the Department of Psychological and Brain Sciences at Boston University. He is also affiliated with the Center for Systems Neuroscience and the Department of Biomedical Engineering. His research focuses on understanding neural mechanisms underlying behavior and cognition, utilizing systems neuroscience approaches. Chandrasekaran's work involves investigating neural circuits and data analysis related to brain function, contributing to the broader understanding of neurobiological processes.

Research topics

• Artificial Intelligence
• Computer Science
• Machine Learning
• Neuroscience
• Cell biology
• Biology
• Algorithm
• Psychology
• Medicine
• Cognitive psychology

Selected publications

• A multimodal approach for visualizing and identifying electrophysiological cell types in vivo

  Nature Communications · 2026-04-15 · 1 citations

  articleOpen accessSenior authorCorresponding

  Neurons of different types perform diverse computations and coordinate their activity during sensation, perception, and action. While electrophysiological recordings can measure the activity of many neurons simultaneously, identifying cell types during these experiments remains difficult. Here we present PhysMAP, a framework adapted from multiomics data analysis that weights multiple electrophysiological modalities simultaneously to obtain interpretable multimodal representations. We apply PhysMAP to seven datasets and demonstrate that these multimodal representations are better aligned with known transcriptomically-defined cell types than any single modality alone. We then show that this alignment allows PhysMAP to better identify putative cell types in the absence of ground truth. We also demonstrate how annotated datasets can transfer labels to unannotated recordings and confirm that inferred cell types exhibit properties consistent with ground truth. Crucially, we show that PhysMAP can also be used to iteratively detect batch effects which confound classification. Together, these results establish PhysMAP as a tool for studying multiple cell types simultaneously and gaining insight into neural circuit dynamics. In this study, the authors develop PhysMAP, which combines multiple electrophysiological features to identify neuronal cell types in the neural circuit, advancing the ability to investigate circuit dynamics without genetic or optical access during behavior.

  PublisherDOI

• Neuropixels reveal laminar microcircuit organization in monkey V1 in vivo

  Proceedings of the National Academy of Sciences · 2026-02-18 · 1 citations

  articleOpen accessSenior authorCorresponding

  The relationship between different cell populations in monkey primary visual cortex and their role in visual function is not fully resolved. We combined high-density Neuropixels recordings across layers of macaque V1, and a state-of-the-art nonlinear dimensionality reduction approach on waveform shape to delineate nine putative cell classes: 4 narrow-spiking (NS), 4 broad-spiking (BS), and 1 triphasic (TP). Then, we performed targeted analyses of laminar organization, spike amplitude, multichannel spatial features, functional properties, and network connectivity of these cell classes. These analyses have uncovered four fundamental aspects of V1 laminar microcircuitry never fully demonstrated before in vivo. First, NS neurons were most concentrated in layer 4 and outnumbered parvalbumin positive neurons, consistent with findings on potassium channel expression in excitatory neurons in V1. Second, a large-amplitude NS cell class in layer 4b was strongly direction selective, with multichannel waveforms suggestive of a stellate morphology, a likely functional correlate of anatomical descriptions of neurons projecting from V1 to MT. Third, another NS cell class in layer 4b showed robust bursting activity and strong orientation selectivity. Finally, cross-correlation analysis revealed distinct feedforward interactions between cell classes in layer 4 and layer 5/6 and those in layer 2/3. These results demonstrate how high-resolution electrophysiology can reveal links between laminar organization and in vivo function of neurons. Our findings offer key insights for biologically realistic microcircuit models of primate V1 and may generalize to other regions.

  PublisherDOI

• Distinct neural dynamics in prefrontal and premotor cortex during flexible perceptual decisions

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-02-04 · 1 citations

  articleOpen accessSenior authorCorresponding

  The neural mechanisms underlying flexible perceptual decisions, where the mapping between sensory input and motor action changes depending on context, remain unclear. Here we show that prefrontal and premotor cortex use distinct neural dynamics to implement different computations during flexible decisions. We trained monkeys to discriminate the dominant color of a red-green checkerboard and report their decision by choosing one of two targets 1;2 . By randomizing the target configuration on a trial-by-trial basis, we ensured a flexible mapping between color (red vs. green) and action choice (left vs. right), necessitating a nonlinear exclusive- or (XOR) computation 3–5 . We found that neural dynamics in dorsolateral prefrontal cortex (DLPFC) led to higher-dimensional population representations than those in dorsal premotor cortex (PMd). Neural activity in DLPFC first separated by target configuration, then by color choice and action choice after stimulus onset, reflecting the XOR computation. In contrast, neural dynamics in PMd led to lower-dimensional representations that only reflected action choice, the output of the XOR computation. These higher-dimensional representations in DLPFC enabled earlier decoding of both color choice and action choice compared to PMd, and were strongest in anterior and ventral DLPFC 6 . These findings reveal distinct computations by neural dynamics: prefrontal cortex implements flexible sensorimotor mappings through high-dimensional representations while dorsal premotor cortex reflects only the selected action.

  PublisherOA PDFDOI

• A. J. Major et al. reply

  Nature Neuroscience · 2025-12-12 · 1 citations

  letter
  PublisherDOI

• The dynamics and geometry of choice in the premotor cortex

  Nature · 2025-06-25 · 24 citations

  articleOpen access

  Abstract The brain represents sensory variables in the coordinated activity of neural populations, in which tuning curves of single neurons define the geometry of the population code 1,2 . Whether the same coding principle holds for dynamic cognitive variables remains unknown because internal cognitive processes unfold with a unique time course on single trials observed only in the irregular spiking of heterogeneous neural populations 3–8 . Here we show the existence of such a population code for the dynamics of choice formation in the primate premotor cortex. We developed an approach to simultaneously infer population dynamics and tuning functions of single neurons to the population state. Applied to spike data recorded during decision-making, our model revealed that populations of neurons encoded the same dynamic variable predicting choices, and heterogeneous firing rates resulted from the diverse tuning of single neurons to this decision variable. The inferred dynamics indicated an attractor mechanism for decision computation. Our results reveal a unifying geometric principle for neural encoding of sensory and dynamic cognitive variables.

  PublisherOA PDFDOI

• Author response: The information bottleneck as a principle underlying multi-area cortical representations during decision-making

  2025-05-28

  peer-reviewOpen access

  Decision-making emerges from distributed computations across multiple brain areas, but it is unclear why the brain distributes the computation. In deep learning, artificial neural networks use multiple areas (or layers) and form optimal representations of task inputs. These optimal representations are sufficient to perform the task well, but minimal so they are invariant to other irrelevant variables. We recorded single neurons and multiunits in dorsolateral prefrontal cortex (DLPFC) and dorsal premotor cortex (PMd) in monkeys during a perceptual decision-making task. We found that while DLPFC represents task-related inputs required to compute the choice, the downstream PMd contains a minimal sufficient, or optimal, representation of the choice. To identify a mechanism for how cortex may form these optimal representations, we trained a multi-area recurrent neural network (RNN) to perform the task. Remarkably, DLPFC and PMd resembling representations emerged in the early and late areas of the multi-area RNN, respectively. The DLPFC-resembling area partially orthogonalized choice information and task inputs and this choice information was preferentially propagated to downstream areas through selective alignment with inter-area connections, while remaining task information was not. Our results suggest that cortex uses multi-area computation to form minimal sufficient representations by preferential propagation of relevant information between areas.The brain uses multiple areas for cognition, decision-making, and action, but it is unclear why cortical activity differs by brain area. Machine learning and information theory suggests that one benefit of multiple areas is that it provides an “information bottleneck” that compresses inputs into an optimal representation that is minimal and sufficient to solve the task. Combining experimental recordings from behaving animals and computational simulations, we show that later brain areas have a tendency to form such minimal sufficient representations of task inputs through preferential propagation of task-relevant information present in earlier areas. Our results thus provide insight into one possible reason why the brain uses multiple brain areas for supporting decision-making and action.

  PublisherDOI

• The information bottleneck as a principle underlying multi-area cortical representations during decision-making

  eLife · 2025-05-28 · 2 citations

  preprintOpen access

  Abstract Decision-making emerges from distributed computations across multiple brain areas, but it is unclear why the brain distributes the computation. In deep learning, artificial neural networks use multiple areas (or layers) and form optimal representations of task inputs. These optimal representations are sufficient to perform the task well, but minimal so they are invariant to other irrelevant variables. We recorded single neurons and multiunits in dorsolateral prefrontal cortex (DLPFC) and dorsal premotor cortex (PMd) in monkeys during a perceptual decision-making task. We found that while DLPFC represents task-related inputs required to compute the choice, the downstream PMd contains a minimal sufficient, or optimal, representation of the choice. To identify a mechanism for how cortex may form these optimal representations, we trained a multi-area recurrent neural network (RNN) to perform the task. Remarkably, DLPFC and PMd resembling representations emerged in the early and late areas of the multi-area RNN, respectively. The DLPFC-resembling area partially orthogonalized choice information and task inputs and this choice information was preferentially propagated to downstream areas through selective alignment with inter-area connections, while remaining task information was not. Our results suggest that cortex uses multi-area computation to form minimal sufficient representations by preferential propagation of relevant information between areas.

  PublisherDOI

• Neuropixels reveal structure-function relationships in monkey V1 <i>in vivo</i>

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-05-18 · 3 citations

  preprintOpen accessSenior authorCorresponding

  Abstract The relationship between structural properties of diverse neuronal populations in monkey primary visual cortex (V1) and their in vivo functional responses is not fully understood. We combined high-density Neuropixels recordings across cortical layers of macaque V1 with non-linear dimensionality reduction on waveform shape to delineate nine putative cell classes: 4 narrow-spiking (NS), 4 broad-spiking (BS) and 1 tri-phasic (TP). Using targeted analyses of laminar organization, spike amplitude, multichannel waveforms, functional properties, and network connectivity of these cell classes, we demonstrate four aspects of the V1 microcircuit predicted by anatomical studies but never fully demonstrated in vivo . First, NS neurons were concentrated in layer 4. Second, a large-amplitude NS cell class in layer 4B showed strong direction selectivity. Third, another layer 4B NS class exhibited robust bursting and orientation selectivity. Finally, cross-correlation analysis revealed functional interactions between cells in different layers. Our results highlight how high-resolution electrophysiology can reveal novel relationships between in vivo function of neurons and the underlying circuit. Teaser High-resolution electrophysiology used with machine learning reveals links between function and the underlying neural circuitry.

  PublisherOA PDFDOI

• A multimodal approach for visualization and identification of electrophysiological cell types <i>in vivo</i>

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-07-31

  preprintOpen accessSenior authorCorresponding

  Neurons of different types perform diverse computations and coordinate their activity during sensation, perception, and action. While electrophysiological recordings can measure the activity of many neurons simultaneously, identifying cell types during these experiments remains difficult. To identify cell types, we developed PhysMAP, a framework that weighs multiple electrophysiological modalities simultaneously to obtain interpretable multimodal representations. We apply PhysMAP to seven datasets and demonstrate that these multimodal representations are better aligned with known transcriptomically-defined cell types than any single modality alone. We then show that such alignment allows PhysMAP to better identify putative cell types in the absence of ground truth. We also demonstrate how annotated datasets can be used to infer multiple cell types simultaneously in unannotated datasets and show that the properties of inferred types are consistent with the known properties of these cell types. Finally, we provide a first-of-its-kind demonstration of how PhysMAP can help understand how multiple cell types interact to drive circuit dynamics. Collectively, these results demonstrate that multimodal representations from PhysMAP enable the study of multiple cell types simultaneously, thus providing insight into neural circuit dynamics.

  PublisherOA PDFDOI

• Monolithic three-dimensional neural probes from deterministic rolling of soft electronics

  Nature Electronics · 2025-08-11 · 5 citations

  articleOpen access
  PublisherOA PDFDOI

Recent grants

• Organization and Dynamics of Premotor and Prefrontal Cortical Circuits Mediating Goal-Directed Behavior

  NIH · $175k · 2015–2017

• Multimodal Characterization of Prefrontal and Premotor Circuits Underlying Perceptual Decision Making in Therhesus Monkey

  NIH · $3.1M · 2021–2026

Frequent coauthors

• Krishna V. Shenoy

  Howard Hughes Medical Institute

  59 shared

• William T. Newsome

  26 shared

• Diogo Peixoto

  Stanford University

  22 shared

• Jonathan C. Kao

  19 shared

• Stephen I. Ryu

  Korea Research Institute of Bioscience and Biotechnology

  15 shared

• Asif A. Ghazanfar

  Princeton University

  13 shared

• Paul Nuyujukian

  Stanford University

  10 shared

• Matthias Gondan

  Universität Innsbruck

  10 shared

Labs

• Chand LabPI

Education

• Ph.D.

  Princeton University