A factor-adjusted multiple testing procedure for ERP data analysis

Abstract

Event-related potentials (ERPs) are now widely collected in psychological research to determine the time courses of mental events. When event-related potentials from treatment conditions are compared, often there is no a priori information on when or how long the differences should occur. Testing simultaneously for differences over the entire set of time points creates a serious multiple comparison problem in which the probability of false positive errors must be controlled, while maintaining reasonable power for correct detection. In this work, we extend the factor-adjusted multiple testing procedure developed by Friguet, Kloareg, and Causeur (Journal of the American Statistical Association, 104, 1406-1415, 2009) to manage the multiplicity problem in ERP data analysis and compare its performance with that of the Benjamini and Hochberg (Journal of the Royal Statistical Society B, 57, 289-300, 1995) false discovery rate procedure, using simulations. The proposed procedure outperformed the latter in detecting more truly significant time points, in addition to reducing the variability of the false discovery rate, suggesting that corrections for mass multiple testings of ERPs can be much improved by modeling the strong local temporal dependencies.

Comparing Self-Regulation-Associated Event Related Potentials in Preschool Children with and without High Levels of Disruptive Behavior

“…It is not currently the universal standard in ERP research to correct for multiple comparisons (for examples see Buss, Dennis, Brooker, & Sippel, 2011; Roos et al, 2014; Stieben et al, 2007; and Wiersema et al, 2006) with some researchers arguing for caution in situations where alpha correction might reduce the sensitivity needed to evaluate hypotheses (Burwell, Malone, & Iocono, 2016). Further, other psychophysiological methodologists have argued that conventional multiple comparison corrections are not appropriate for ERP data in cases where different waveforms are derived from the same task and are thus dependent on each other (Causeur, Chu, Hsieh, & Sheu, 2012). Nonetheless, we employed the False Discovery Rate (FDR) correction (Benjamini & Hochberg, 1995) to provide a more conservative estimate of effects.…”

The time course of individual face recognition: A pattern analysis of ERP signals

“…Concretely, instead of targeting specific components, we subjected ERP amplitudes to a sequence of one-way ANOVAs across participants, separately for each time point and each electrode. The results of these analyses were then corrected for multiple comparisons using the false discovery rate (FDR) (Benjamini and Hochberg, 1995; for its application to EEG see Causeur et al, 2012).…”

A functional generalized F‐test for signal detection with applications to event‐related potentials significance analysis

“…When the number of variables exceeds the sample size, ML estimators of the factor model parameters can be obtained using an EM algorithm (Friguet et al ., ; Causeur et al ., ).…”

“…The maximum likelihood (ML) estimators of φ and θ are obtained by replacing Ψ and L with their ML estimators Ψ  and L  , respectively, in expressions (7) When the number of variables exceeds the sample size, ML estimators of the factor model parameters can be obtained using an EM algorithm (Friguet et al, 2009;Causeur et al, 2012).…”