jokra1 / hsmm_ld_eeg

Repository for "Word type and frequency effects on lexical decisions are process-dependent and start early" (Krause, van Rij, & Borst; submitted)

HSMM_LD_EEG

Description:

Code to perform the HsMM-MVPA model selection and trial-level GAMM analysis described in Word type and frequency effects on lexical decisions are process-dependent and start early (Krause, van Rij, & Borst; submitted). The code makes use of the HsMM functions provided by Berberyan et al. (2021) that can be downloaded from here. Additionally, the code requires a copy of the item-level DLP data (Keuleers et al., 2010), which can for example be downloaded here. The vwr R package is required for parts of the trial-level analysis but appears to no longer be on CRAN - the most recent version can still be downloaded from here and then installed locally.

Performing the analysis:

• Clone the repository
• Copy the HsMM functions folder into the repository
• Copy the dlp-items.Rdata and dlp-stimuli.Rdata files into the data folder
• Open Matlab and make sure the repository is the working/current directory
• Install eeglab, the parallel computing, and wavelet toolboxes if they are not yet installed
• Run HSMMMVPA_forward_selection.mat
• Open R and make sure the repository is the working/current directory, (optionally) install the dependencies via renv
• Run trial_level_analyis.Rmd