Kai Chen 陈开

Computational Neuroscientist
Ph.D in Applied Mathematics
Amateur Powerlifter

🌟 I study the brain 🧠, mainly using math and numerical analysis to unravel its complexities.

🌟 I like coding and related technical stuffs 👨🏻‍💻, ideally when they are fun to play with.

🌟 I’m an amateur powerlifter 🏋🏻‍♂️. Heavy objects go up, sometimes on purpose.

🌟 I love road cycling 🚴🏻‍♂️, usually as a way to think about work and life while pretending not to.

I’m currently exploring France 🇫🇷 as my new territory of life, working as a CNRS postdoctoral researcher at the Laboratoire de Neurosciences Cognitives & Computationnelles, École Normale Supérieure (ENS) in Paris, working with Dr. Srdjan Ostojic, also known as the “low-rank” group.

Before that, I accomplished my Ph.D. training in Applied Mathematics at the Lab of Computational Neuroscience, Institute of Natural Science, Shanghai Jiao Tong University, under the supervision of Prof. Douglas Zhou and Prof. Songting Li.

Earlier still, I studied physics for both my master’s and bachelor’s degrees at Shanghai Jiao Tong University.

Research Interests: Computational Neuroscience, Machine Learning, Applied Mathematics.

Network dynamics and computations.
1. How perception and cognition conducted through multi-regional network dynamics in brain.
2. The relationship between network connectivity, network dynamics, and network functions.
3. Biologically plausible learning rules for recurrent neural networks.
Neural coding and data analysis.
1. Linking structural, functional, and effective connectivity using multi-modal neural data.
2. The principle underlying spike temporal coding
Temporal proception in the brain (my current ERC project CHRONOLOGY).
1. How the cortical network represent varying time durations in terms of population dynamics.
2. How the neural dynamics bind independent memory items sequentially onto the neural manifold in the temporal order.
3. Can we parametrically manipulate the network dynamics to systemetically perturb the temporal perception and change the working memory capacity?

news

December 24, 2025	Our new paper Anatomical Connectivity Reconstruction of Biological Neuronal Networks Using Granger Causality has been accepted as a research article by Neural Networks
December 21, 2025	Our recent work Emergent hierarchical representations for multi-task learning in anatomy-constrained RNNs has been accepted as a poster at COSYNE 2026.
October 08, 2025	I am starting my new postdoc position at École Normale Supérieure in Paris.
September 30, 2025	I have been awarded my PhD degree in applied mathematics.
August 22, 2025	I passed my PhD defence.

selected publications

PNAS
Causal Connectivity Measures for Pulse-Output Network Reconstruction: Analysis and Applications

Zhong-qi K. Tian, Kai Chen, Songting Li, David W. McLaughlin, and Douglas Zhou

Proceedings of the National Academy of Sciences, 121 (14), e2305297121, Apr 2024

Abs DOI Bib PDF Supp Poster

The causal connectivity of a network is often inferred to understand network function. It is arguably acknowledged that the inferred causal connectivity relies on the causality measure one applies, and it may differ from the network’s underlying structural connectivity. However, the interpretation of causal connectivity remains to be fully clarified, in particular, how causal connectivity depends on causality measures and how causal connectivity relates to structural connectivity. Here, we focus on nonlinear networks with pulse signals as measured output, e.g., neural networks with spike output, and address the above issues based on four commonly utilized causality measures, i.e., time-delayed correlation coefficient, time-delayed mutual information, Granger causality, and transfer entropy. We theoretically show how these causality measures are related to one another when applied to pulse signals. Taking a simulated Hodgkin–Huxley network and a real mouse brain network as two illustrative examples, we further verify the quantitative relations among the four causality measures and demonstrate that the causal connectivity inferred by any of the four well coincides with the underlying network structural connectivity, therefore illustrating a direct link between the causal and structural connectivity. We stress that the structural connectivity of pulse-output networks can be reconstructed pairwise without conditioning on the global information of all other nodes in a network, thus circumventing the curse of dimensionality. Our framework provides a practical and effective approach for pulse-output network reconstruction.
@article{tian2024causal, title = {Causal Connectivity Measures for Pulse-Output Network Reconstruction: {{Analysis}} and Applications}, author = {Tian, Zhong-qi K. and Chen, Kai and Li, Songting and McLaughlin, David W. and Zhou, Douglas}, year = {2024}, month = apr, journal = {Proceedings of the National Academy of Sciences}, volume = {121}, number = {14}, pages = {e2305297121}, issn = {0027-8424, 1091-6490}, doi = {10.1073/pnas.2305297121}, copyright = {All rights reserved}, dimensions = {true}, }