GBLM-SMOOTH

GBLM-SMOOTH is a model, implemented in MATLAB, that can track long-term and short-term synaptic plasticity simultaneously. This repository contains four folders.

1. "core"

Three sub-folders are included:

(i) "diagnose": tools to evaluate fitting.
(ii) "modelFit": implementation of the model. The basic model (without tuning Q in adaptive smoothing) starts from smooth_gblm.m. The Q-tuned versions are generally called tune_smooth_gblm_xd_xxxx.m. "xd"" means whether it is 2D optimization or 1D approximation. The last four letters denote optimization methods, i.e. "_grad" for gradient descent and "_grid" for grid search. (all "v2" turn on the estimation of history effects)
(iii) "simulation": code for generating neural spikes.

2. "demo"

This folder provides code for generating all figures in the paper. Generally, "_caseDemo" denotes code for simulations and model fitting, while "_caseAnalysis" denotes code for plotting.

3. "helper"

It contains helper functions for simulations and model fitting.

4. "plot"

This folder contains all the plots used in the paper.