gtas-generative-spike-train-model

What problem does it solve?

GTaS Generative Spike Train Model provides a principled framework to generate correlated neural spike trains by starting from a marginal Poisson process and applying thinning and time-shift operations to induce controllable higher-order temporal dependencies, enabling analytical characterization of spike train statistics.

Core Features & Use Cases

• Generate multi-neuron spike trains with tunable thinning probabilities and shifts to create 2nd- and higher-order correlations.
• Analyze cumulants and higher-order statistics to study how temporal structure affects network dynamics and decoding.
• Use as synthetic data for testing neural decoding algorithms, network simulations, and methodological validation.

Quick Start

Define a Poisson baseline rate, apply thinning with probability p, apply a shift Δ, generate the trains, and compute cumulants for subsequent analysis.