Smoothing Spline Analysis of Variance Models for Electroencephalography Data Analysis

• Activity Number: 165
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Committee on Applied Statisticians
• Author(s): Nathaniel Helwig*+
• Companies: University of Illinois
• Keywords: SSANOVA ; Smoothing ; Spline ; Electroencephalography ; EEG ; ERP

Abstract

Electroencephalography (EEG) data consist of electrical activity recorded over time from electrodes on the scalp. Typical EEG studies collect multi-electrode data from many subjects in attempt to compare differences in brain activity between different subject groups. In this talk, I discuss how smoothing spline analysis of variance (SSANOVA) models can be used to reliably analyze differences in EEG data. First, I provide an overview of the SSANOVA framework and discuss some of the model's large sample and dimensionality issues. Next, I present some approximations for fitting tensor product SSANOVA models to large samples. I conclude by discussing how two- and three-way SSANOVA models can be used to reliably compare differences in single- or multi-electrode EEG data, and I provide examples using real (and open-source) EEG data.