ShiNung Ching

This paper develops a computational framework to model how task-induced changes in brain dynamics relate to cognitive individual differences using rsfMRI and tfMRI data.

This paper shows that astrocytes process neuromodulatory inputs into distinct temporal signals that modulate synaptic pathways and learning, with dynamics similar to neural network critic mechanisms.

The paper models neural dynamics on an attractor landscape to show how attentional states and contexts alter the geometry of brain activity trajectories.

The paper defines criticality in low-dimensional dynamical systems and shows it enhances information processing in neural models.

This study shows that ventral striatal astrocytes influence reinforcement learning by modulating decision noisiness and reward-guided choice through calcium dynamics.

This paper introduces a modulated MINDy model to capture individualized and nonstationary brain dynamics, validated on synthetic and EEG data, showing accuracy and biological interpretability.

The paper develops a mathematical framework to explain how neural adaptation in C. elegans leads to flexible behavioral responses, including habituation and paradoxical inversion.

The paper constructs dynamical models of resting-state brain networks from fMRI data, revealing anatomically reliable attractor landscapes that go beyond conventional functional connectivity analyses.

This paper develops a framework to model brain dynamics from time-series data, revealing individual differences in alpha and beta oscillations through excitation-inhibition balance.

This paper improves task fMRI preprocessing by filtering out pre-event intrinsic activity using individualized MINDy models, enhancing statistical power and behavioral prediction.

The paper presents a method to efficiently compute all fixed points of discrete systems on large networks using graph partitioning and decomposition.

This paper introduces MINDy, a method to create individualized dynamic brain models from resting state fMRI data, showing they are predictive and better at uncovering mechanisms than functional connectivity methods.

This paper introduces surrogate deconvolution to estimate hidden biological systems from indirect measurements, showing it is scalable and competitive with Kalman Filters in neuroscience applications.

This paper uses optimization-based modeling to study how sensory networks, specifically in the olfactory system, track low-dimensional latent representations of stimuli through temporal dynamics.

This paper introduces a geometric classification method for brain network dynamics that achieves high temporal resolution and anatomical inference using conic derivative discriminants.