An integrated heterogeneous Poisson model for neuron functions in hand movement during reaching and grasp

TL;DR

This paper introduces a heterogeneous Poisson regression model to better analyze neuronal firing data during reach-to-grasp movements, addressing limitations of traditional models in neuroscience research.

Contribution

It proposes an integrated heterogeneous Poisson model tailored for neural activity data, improving upon traditional Poisson regression methods.

Findings

The model effectively categorizes neural activities during motor tasks.

It overcomes disadvantages of traditional Poisson regression.

Applicable to various neural data analysis scenarios.

Abstract

To understand potential encoding mechanism of motor cortical neurons for control commands during reach-to-grasp movements, experiments to record neuronal activities from primary motor cortical regions have been conducted in many research laboratories (for example, (7), (17)). The most popular approach in neuroscience community is to fit the Analysis of Variance (ANOVA) model using the firing rates of individual neurons. In addition to consider neural firing counts but also temporal intervals, (5) proposed to apply Analysis of Covariance (ANCOVA) model. Due to the nature of the data, in this paper we propose to apply an integrated method, called heterogeneous Poisson regression model, to categorize different neural activities. Three scenarios are discussed to show that the proposed heterogeneous Poisson regression model can overcome some disadvantages of the traditional Poisson regression model.