An Experimental Sensitivity Analysis of Gaussian and Non-Gaussian Based Methods for Dynamic Modeling in EEG Signal Processing

An Experimental Sensitivity Analysis of Gaussian and Non-Gaussian Based Methods for Dynamic Modeling in EEG Signal Processing

Gonzalo Safont (Universitat Politècnica de València, Spain), Addisson Salazar (Universitat Politècnica de València, Spain), Alberto Rodriguez (Universidad Miguel Hernández de Elche, Spain) and Luis Vergara (Universitat Politècnica de València, Spain)
Copyright: © 2015 |Pages: 14
DOI: 10.4018/978-1-4666-5888-2.ch397
OnDemand PDF Download:
$30.00
List Price: $37.50

Chapter Preview

Top

Background

Sensitivity analysis is generally defined as the study of how the outputs of a method change under a given change of its inputs. It has many applications, such as testing the stability of a method or testing its robustness when the main assumptions no longer hold true. Sensitivity analysis has been used in many different fields, such as risk assessment, algorithm characterization, and medical studies (Saltelli, et al., 2008). Within medical research, there are many applications of sensitivity analysis on EEG signals, methods and equipment, such as: testing the influence of different factors (the scalp, electrodes…) on the EEG (Rice, Rorden, Little, & Parra, 2013); exploring the behavior of head conductivity and EEG models (Vallaghe, 2009; Dannhauer, Lanfer, Wolters, & Knösche, 2011); or performing meta-analysis studies (Gao & Raine, 2009).

Key Terms in this Chapter

Hypnogram: Representation of the stages of sleep during a period of time.

Sensitivity analysis: Study of the uncertainty of a model with respect to changes in its assumptions.

BSS: Estimation of a set of unknown signals which have been mixed to produce a set of observed signals.

BN: Directed graphical model that expresses conditional independence between variables.

Classification: Separation of observed data into several mutually-exclusive categories.

Electroencephalogram: Neurophysiologic recording of electrical brain activity by electrodes placed on the scalp.

SICAMM: ICA mixture model in which the classes show sequential dependence.

Complete Chapter List

Search this Book:
Reset