Source modelling
To investigate the neural underpinnings of the effect of exposure to
the different types of speech input, we conducted unbiased distributed
source estimations of the EEG data. BESA 6.0 software was used to run
Laplacian weighted minimum norm algorithm with two iterations
(LORETA; Pascual-Marqui et al., 1994). Individual subject-level source estimations
were calculated for time-windows where significant ERP effects
were observed (see below for details). These individual source images
were averaged together for a group average image and average source
maxima were used to guide the selection of voxel coordinates in the
standardised Talairach space (Talairach and Tournoux, 1988). A 4-shell ellipsoidal
head model was used in the source reconstruction process.