General information ------------------- *PLEASE NOTE THAT THIS INFORMATION CAN ALSO BE FOUND IN THE PDF FILE README.PDF WITH BETTER FORMATTING* This OSF project refers to the paper “Reduced visual attention in heterogeneous textures is reflected in occipital alpha and theta band activity” by Tobias Feldmann-Wüstefeld, Makoto Miyakoshi, Marco Petilli, Anna Schubö, and Scott Makeig, published in PLoS One in November 2017. The experiment was run in E-Prime 2.0 and EEG data was recorded from 64 sites with a BrainProducts actiCap system. Additional information can be found in the article online. For any further information, inquiries etc., please contact the corresponding author Tobias Feldmann-Wüstefeld (tobiasfw@uchicago.edu). Overview -------- There are two main folders. One for the ERP pipeline and one for the ICA pipeline. Should you have any questions how to use the ERP scripts, please correspond Tobias Feldmann-Wüstefeld (tobiasfw@uchicago.edu). The ERP pipeline has nine folders: - experiment: runtime software (E-Prime 2.0 files and image files) - scripts: All MATLAB scripts used for this article - rawEEG: raw EEG data files (*.eeg, *.vhdr, *.vmrk) - rawBehavior: E-Prime output files (*.edat2), exported text files (*.txt) and comma separated values (*.csv) - behavior: MATLAB files with independent variables and (behavioral) dependent variables for each trial and each subject (*_behav.mat); It also contains marker files (*._marker.txt) for EEG analyses - preproc: preprocessed EEG data (*._1.mat) - AV: averaged preprocessed EEG data (AV_BILAT.mat, AV_LAT.mat) - export: exported behavioral data, exported EEG data - statistics: data sheet with behavioral and EEG data, SPSS syntax to analyze data sheet and SPSS output file The ICA pipeline has one folder with all scripts used for the ICA pipeline and a README file for instructions. The raw behavioral and raw EEG files in the ERP folder structure were used for these scripts. Should you have any questions how to use the ICA scripts, please correspond Makoto Miyakoshi (mmiyakoshi@ucsd.edu). Behavioral Processing --------------------- Please note that E-Prime makes automatic processing extremely difficult. Some processing steps were thus done by hand. 1. Exporting E-Prime files to StatView format (.txt) with software E-DataAid (comes with E-Prime). 2. Loading .txt files with Microsoft Excel and saving them as comma separated value files (.csv). .txt files should be in folder rawBehavior. 3. Using script getBehavior.m (in folder scripts) to export raw files to MATLAB. Saves crucial variables for each trial and each subject in folder behavior: dependent variables response time (data.rt) and accuracy (data.accuracy) and independent variables targetPresence (data.targetPresence) and distractor heterogeneity (data.heterog). This script will also extract the correct markers from the behavioral files that are required for EEG analyses (markers were not correctly transmitted due to problems with the parallel port). The marker files will end up in the folder behavior as well. 4. Export behavioral data with exportBehavior.m. This will sort trials according to conditions, average across trials and produce a subjects x conditions matrix for the dependent variables response time (rtMatrix), accuracy (accMatrix), d-prime (dprime) and the natural logarithm of beta (lnbeta). 5. Statistics were run in SPSS. Behavioral data was copied/pasted from file behavResults.mat into SPSS data sheet and then the syntax found in statistics\StatisticsSyntax.sps was run to obtain statistical parameters. EEG processing steps (ERPs) --------------------------- To simply run all EEG processing steps as used in the article, copy the entire folder structure to any folder on your computer and delete all files in the folder preproc and AV. Then use the script EEGPIPELINE.m in scripts. 0.) Preprocessing parameter: use preprocSettings.m in folder scripts; current settings are the ones used in the article. Does not need to be run (will be referred to from runPreproc.m). 1.) Preprocessing: use runPreproc.m (refers to settings from 1.); does the following steps: a. Converting BrainProducts raw files to MATLAB format (required scripts: load_BrainProducts_file.m, parsebvmrk.m, pop_loadbv.m, readbvconf.m) b. Re-reference (rereference.m) c. Segment (segment.m) d. Baseline correction (inline code) Result: preprocessing files for each subject (format e.g. 13_1.mat for subject 13) in folder preproc. Crucially, segmented data can be found in variable “erp.trial.data” (trials x electrodes x time points). 2.) Artifact rejection: use artReject_ERP.m; identifies blinks, saccades Result: erp.arf is added to files in preproc folder. It contains vectors for when participants blinked (erp.arf.noiseBlink), or made saccades (erp.arf.noiseSacc). 3.) Sort trials according to conditions and average across trials without blinks and saccades: use sortCond_ERP_bilatN2.m for bilateral analyses and use sortCond_ERP_N2pc.m for lateral analyses. Results: Two average files (AV_BILAT.mat , AV_LAT.mat) in folder AV, one for bilateral and one for lateral analyses. They contain an unpooled data matrix for bilateral analyses of the format subjects x conditions x electrodes x time points (AV.cleanAV) and a pooled data matrix for lateral analyses of the format subjects x conditions x electrodes x laterality x time points (AV.pooled), respectively. Lateralized refers to contra (=1) versus ipsi (=2). 4.) Export and plot data. Use plotGAV_bilatN2.m for bilateral N2 and plotGAV_N2pc.m for N2pc. For export, this function averages across specified time windows. For plotting, this function averages across participants. Excludes participants with too many artifacts. Result: export files in folder export for bilateral (export_BilatN2_1.mat) and lateral (export_N2pc_1.mat) analyses. Both contain a subjects x conditions matrix with the exported data. 5.) Statistics were done with SPSS version 23. The files (data sheets, syntax and output) can be found in the folder statistics. What you can do --------------- For following the exact processing steps used in the article, see section “Data processing” above. For running analyses different from the ones used in the article, some possibilities are listed here. They are sorted from little to a lot of reprocessing necessary. Please note that changing any processing step requires to execute every subsequent step again. For example, if the artifact rejection is changed, sorting/averaging and exporting needs to be done again. - Export different time windows or plot conditions differently (use files plotGAV_bilatN2.m for bilateral N2 and plotGAV_N2pc.m for lateral analyses). - Sort/Average EEG data differently, e.g., all trials instead of correct trials, collapse across conditions (use files sortCond_ERP_bilatN2.m for bilateral analyses and sortCond_ERP_N2pc.m for lateral analyses) - Work with segmented data to do different analyses, e.g. frequency analyses, decoding, forward encoding etc. - Apply a more or less strict artifact rejection criterion (use file artReject_ERP.m) - Change preprocessing settings, e.g. baseline differently, use different rereferencing, change analysis time window (use file preprocSettings.m) - Use the raw EEG and behavioral data to do anything you like