The Annals of Applied Statistics
- Ann. Appl. Stat.
- Volume 10, Number 1 (2016), 219-245.
Accounting for time dependence in large-scale multiple testing of event-related potential data
Event-related potentials (ERPs) are recordings of electrical activity along the scalp time-locked to perceptual, motor and cognitive events. Because ERP signals are often rare and weak, relative to the large between-subject variability, establishing significant associations between ERPs and behavioral (or experimental) variables of interest poses major challenges for statistical analysis.
Noting that ERP time dependence exhibits a block pattern suggesting strong local and long-range autocorrelation components, we propose a flexible factor modeling of dependence. An adaptive factor adjustment procedure is derived from a joint estimation of the signal and noise processes, given a prior knowledge of the noise-alone intervals. A simulation study is presented using known signals embedded in a real dependence structure extracted from authentic ERP measurements. The proposed procedure performs well compared with existing multiple testing procedures and is more powerful at discovering interesting ERP features.
Ann. Appl. Stat., Volume 10, Number 1 (2016), 219-245.
Received: July 2014
Revised: October 2015
First available in Project Euclid: 25 March 2016
Permanent link to this document
Digital Object Identifier
Mathematical Reviews number (MathSciNet)
Zentralblatt MATH identifier
Sheu, Ching-Fan; Perthame, Émeline; Lee, Yuh-shiow; Causeur, David. Accounting for time dependence in large-scale multiple testing of event-related potential data. Ann. Appl. Stat. 10 (2016), no. 1, 219--245. doi:10.1214/15-AOAS888. https://projecteuclid.org/euclid.aoas/1458909914
- Accounting for time dependence in large-scale multiple testing of event-related potential data: Online supplement. The impact of ERP time dependence on multiple testing results. To demonstrate the impact of time dependence on the ability of multiple testing procedures to identify a predetermined true signal, a simulation study is conducted in which ERP data are generated according to model (3.1). This simulation study compares the GB procedure [Guthrie and Buchwald (1991)] and two FDR-controlling procedures: BH [Benjamini and Hochberg (1995)] and BY [Benjamini and Yekutieli (2001)]. The results highlight the instability of multiple testing results when using methods ignoring dependence among tests.