Linear dynamical neural population models through nonlinear embeddings
نویسندگان
چکیده
A body of recent work in modeling neural activity focuses on recovering lowdimensional latent features that capture the statistical structure of large-scale neural populations. Most such approaches have focused on linear generative models, where inference is computationally tractable. Here, we propose fLDS, a general class of nonlinear generative models that permits the firing rate of each neuron to vary as an arbitrary smooth function of a latent, linear dynamical state. This extra flexibility allows the model to capture a richer set of neural variability than a purely linear model, but retains an easily visualizable low-dimensional latent space. To fit this class of non-conjugate models we propose a variational inference scheme, along with a novel approximate posterior capable of capturing rich temporal correlations across time. 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, fLDS captures a much larger proportion of neural variability with a small number of latent dimensions, providing superior predictive performance and interpretability.
منابع مشابه
A Linear Dynamical System Model for Text
Low dimensional representations of words allow accurate NLP models to be trained on limited annotated data. While most representations ignore words’ local context, a natural way to induce context-dependent representations is to perform inference in a probabilistic latent-variable sequence model. Given the recent success of continuous vector space word representations, we provide such an inferen...
متن کاملDelay Differential Analysis of Time Series
Nonlinear dynamical system analysis based on embedding theory has been used for modeling and prediction, but it also has applications to signal detection and classification of time series. An embedding creates a multidimensional geometrical object from a single time series. Traditionally either delay or derivative embeddings have been used. The delay embedding is composed of delayed versions of...
متن کاملWhich Methodology is Better for Combining Linear and Nonlinear Models for Time Series Forecasting?
Both theoretical and empirical findings have suggested that combining different models can be an effective way to improve the predictive performance of each individual model. It is especially occurred when the models in the ensemble are quite different. Hybrid techniques that decompose a time series into its linear and nonlinear components are one of the most important kinds of the hybrid model...
متن کاملExistence and uniqueness of positive solutions of a system of nonlinear algebraic equations
Nonlinear or linear algebraic systems appear as steady-state equations in continuous and discrete dynamical models (e.g., reaction-diffusion equations [14,19], neural networks [5,6,15,22] compartmental systems [2,4,11,12,16,17], population models [13,21]). Next we mention some typical models. Compartmental systems are used to model many processes in pharmacokinetics, metabolism, epidemiology an...
متن کاملUsing bundle embeddings to predict daily cortisol levels in human subjects
BACKGROUND Many biological variables sampled from human subjects show a diurnal pattern, which poses special demands on the techniques used to analyze such data. Furthermore, most biological variables belong to nonlinear dynamical systems, which may make linear statistical techniques less suitable to analyze their dynamics. The current study investigates the usefulness of two analysis technique...
متن کامل