Spatial-Temporal Clustering of Neural Data Using Linked-Mixtures of Hidden Markov Models

Shalom Darmanjian,Jose Principe

doi:10.1155/2009/892461

Shalom Darmanjian, Jose Principe

Open Access

https://doi.org/10.1155/2009/892461

Copy DOI

Abstract

This paper builds upon the previous Brain Machine Interface (BMI) signal processing models that require apriori knowledge about the patient's arm kinematics. Specifically, we propose an unsupervised hierarchical clustering model that attempts to discover both the interdependencies between neural channels and the self-organized clusters represented in the spatial-temporal neural data. Results from both synthetic data generated with a realistic neural model and real BMI data are used to quantify the performance of the proposed methodology. Since BMIs must work with disabled patients who lack arm kinematic information, the clustering work described within this paper is relevant for future BMIs.

Highlights

Thousands of people have suffered tragic accidents or debilitating diseases that have either partially or fully removed their ability to effectively interact in the external world
This Achilles heel plagues most Brain Machine Interface (BMI) algorithms since they require kinematic training data to find a mapping to the neural data
Since there are no ground truths to label real BMI neural data, simulations on plausible artificial data will help support the results found by the clustering framework on real data

Summary

Introduction

Thousands of people have suffered tragic accidents or debilitating diseases that have either partially or fully removed their ability to effectively interact in the external world. Most of the behaving animals engaged in BMI experiments are not paralyzed, allowing the kinematic information to be used in the training on the models. This Achilles heel plagues most BMI algorithms since they require kinematic training data to find a mapping to the neural data. The goal is to find a model that can learn these temporal-spatial structures or clusters and segment the EURASIP Journal on Advances in Signal Processing neural data without kinematic clues (i.e., unsupervised). The methodology described will explain how the model learns the parameters and structure of the neural data in order to provide a final set of class or cluster labels for segmentation

Generative Model and Clustering Framework

Simulations

Experimental Animal Data

Experimental Results

Discussion

LM-HMM Framework