CFE 2019: Start Registration
View Submission - CMStatistics
Title: Testing time by time differences of EEG signals using the slopes within multiple comparisons procedures Authors:  Jaromil Frossard - University of Geneva (Switzerland) [presenting]
Olivier Renaud - University of Geneva (Switzerland)
Abstract: Electroencephalography (EEG) is widely used in experiments recording brain activity. The researcher would like to compare the EEG signals and to test differences between experimental conditions. Each time point is tested, and a multiple comparisons procedure should be used to control the family wise error rate (FWER). The state-of-the-art statistical analysis is permutation tests in combination with multiple comparisons procedures like the cluster-mass tests or the threshold-free cluster-enhancement. We show how the slopes of the signals may be used to better modelise the researcher's hypothesis and to increase the power of the test while controlling the FWER.