ERPs to transitions in facial expression: experiment 2
Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
These are 32 channel EEG raw data files (with blinks removed) recorded from 16 (anonymous) young adult participants. Full details of methods are given in a paper accepted for publication in PLoS One, which compares ERPs and recognition scores to dynamic changes in face images of 10 individuals (JAFFE dataset, link below). In Expt. 2, changes in happy facial expressions (happy/neutral or neutral/happy) are compared with changes in frightened facial expressions (frightened/neutral or neutral/frightened).
The hypotheses were 1) that there are ERP differences between neutral-happy and neutral frightened transitions, 2) that there are ERP differences between happy-neutral and frightened-neutral transitions and 3) that there are directional differences in responses to transitions towards and away from neutral. Hypotheses 1) and 3) were supported, but 2 was not.
The .cnt files are in Neuroscan Scan 3.3 format. Electrode locations are standard 10/20 locations for the 32 channel quik-cap. Stimulus events coincide with the instantaneous transition between two 500ms duration face images.The event codes are as follows: 46 = neutral-happy; 47 = neutral-frightened; 55 = happy-neutral, 56 = frightened-neutral. Response events are button presses indicating perceived change type: 33 = change to or from happy, 34 = change to or from frightened.
Neuroscan files can be read by open access software such as EEGlab, SPM or Brainstorm (links below) as well as Neuroscan proprietary software.
Excluded data for 5 further participants is available in a separate zip file: three were rejected for an incorrect 1Hz high pass filter setting during acquisition, two for noise issues.
The excel file happyfear_differences.xls contains paired-sample t-values (group level t-map) generated between pairs of experimental conditions, and corresponds with fig 5. Columns represent electrodes and rows represent time samples (downsampled to 15 ms/sample by spline interpolation).
The formulae and uncorrected p-values as formatted for FDR correction are included. The actual FDR correction to obtain corrected critical values of t was done using an excel spreadsheet avalilable from the following source.
Pike N. Using false discovery rates for multiple comparisons in ecology and evolution. Method Ecol Evol 2010; 2: 278-282.