Abstract:Understanding how the dynamics of a neural network is shaped by the network structure and, consequently, how the network structure facilitates the functions implemented by the neural system is at the core of using mathematical models to elucidate brain functions. This study investigates the tracking dynamics of continuous attractor neural networks (CANNs). Due to the translational invariance of neuronal recurrent interactions, CANNs can hold a continuous family of stationary states. They form a continuous manifold in which the neural system is neutrally stable. We systematically explore how this property facilitates the tracking performance of a CANN, which is believed to have clear correspondence with brain functions. By using the wave functions of the quantum harmonic oscillator as the basis, we demonstrate how the dynamics of a CANN is decomposed into different motion modes, corresponding to distortions in the amplitude, position, width, or skewness of the network state. We then develop a perturbation approach that utilizes the dominating movement of the network's stationary states in the state space. This method allows us to approximate the network dynamics up to an arbitrary accuracy depending on the order of perturbation used. We quantify the distortions of a gaussian bump during tracking and study their effects on tracking performance. Results are obtained on the maximum speed for a moving stimulus to be trackable and the reaction time for the network to catch up with an abrupt change in the stimulus.

Learning A Continuous Attractor Neural Network from Real Images

Continuous Attractor Neural Networks: Candidate of a Canonical Model for Neural Information Representation

Continuous Attractors of Lotka-Volterra Recurrent Neural Networks

Tracking Changing Stimuli in Continuous Attractor Neural Networks

Spiking Continuous Attractor Neural Networks with Spike Frequency Adaptation for Anticipative Tracking

Continuous attractor neural network model of multisensory integration

Dynamics of Adaptive Continuous Attractor Neural Networks

Dynamics Of Neural Networks With Continuous Attractors

Deep CovDenseSNN: A Hierarchical Event-Driven Dynamic Framework with Spiking Neurons in Noisy Environment

Continuous Attractors of Discrete-Time Recurrent Neural Networks

Slow and Weak Attractor Computation Embedded in Fast and Strong E-I Balanced Neural Dynamics.

Efficient Spiking Neural Networks with Sparse Selective Activation for Continual Learning

Information content in continuous attractor neural networks is preserved in the presence of moderate disordered background connectivity

CSNN: an Augmented Spiking Based Framework with Perceptron-Inception

A Moving Bump in a Continuous Manifold: a Comprehensive Study of the Tracking Dynamics of Continuous Attractor Neural Networks.

Deep Continuous Networks

Representations Of Continuous Attractors Of Discrete-Time Cellular Neural Networks

Learning, Storing, and Disentangling Correlated Patterns in Neural Networks.

Self-organising Continuous Attractor Networks with Multiple Activity Packets, and the Representation of Space

Dynamical Properties of Continuous Attractor Neural Network with Background Tuning

Continuous Attractors Of A Class Of Recurrent Neural Networks Without Lateral Inhibition