Linear dynamical neural population models through nonlinear embeddings

Authors: Yuanjun Gao, Evan W. Archer, Liam Paninski, John P. Cunningham

NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show that our techniques permit inference in a wide class of generative models.We also show in application to two neural datasets that, compared to state-of-the-art neural population models, f LDS captures a much larger proportion of neural variability with a small number of latent dimensions, providing superior predictive performance and interpretability. and 5 Experiments
Researcher Affiliation Academia Columbia University New York, NY, United States
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes A Python/Theano [26, 27] implementation of our algorithms is available at http://github.com/earcher/vilds.
Open Datasets Yes Macaque V1 with drifting grating stimulus with single orientation: The dataset consists of 148 neurons simultaneously recorded from the primary visual cortex (area V1) of an anesthetized macaque, as described in [20] (array 5). and Macaque center-out reaching data: We analyzed the neural population data recorded from the macaque motor cortex(G20040123), details of which can be found in [11, 1].
Dataset Splits Yes For each orientation, we divide the data into 120 training trials and 30 testing trials. For Pf LDS we further divide the 120 training trials into 110 trials for fitting and 10 trials for validation (we use the ELBO on validation set to determine when to stop training).
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU models, CPU models, or cloud instance types) used for running the experiments.
Software Dependencies No The paper mentions 'Python/Theano' but does not specify version numbers for these software components.
Experiment Setup Yes When training a model using the AEVB algorithm, we run 500 epochs before stopping. and We analyze the spike activity from 300ms to 1200ms after stimulus onset. We discretize the data at t = 10ms, resulting in T = 90 timepoints per trial.