## Materials for lab-based sessions 2, 3, 4 and 6 The preparation of the stimuli and the final materials are available in [**L3_transfer_longitudinal.zip**](https://osf.io/wbjyr). Furthermore, to facilitate the consultation of the individual files, a [**GitHub repository**](https://github.com/pablobernabeu/L3_transfer_longitudinal) contains all files except one that exceeds the GitHub limit of 100 MB. This file is `stimulus_preparation/external_corpora/nowac-1.1.lemma.frek.sort_alf_frek.txt`, and is available in [**L3_transfer_longitudinal.zip**](https://osf.io/wbjyr). <br> ## Lab notes <br> - Lack of marker 3 (denoting the property of verb-object agreement) in the experiment of Session **4** in both sites. This glitch affects data collected with versions **previous to 1.1.0**. Version numbers are reflected in the behavioural data files, within the column `materials_version`. The missing markers *can* be manually fixed by using the behavioural data as a reference for the content of each trial. Specifically, this glitch is to be fixed by editing the vmrk files in a plain text editor, such as [**Editpad**][1] (see illustration below). The key will be to replace only the *appropriate* instances of `S 2` with `S 3`. The appropriate instances will be determined by following the order of the trials in the behavioural file for each participant. It must be done per participant as the order of trials varies across participants. Finally, the corrected vmrk files will contain 144 instances of `S 2` and 144 instances of `S 3`. <a href="https://ibb.co/7zCZH3P"><img src="https://i.ibb.co/z5xB13p/fix.png" alt="fix" border="0"></a> <br> To help track the trials in the vmrk files, two [R functions](https://osf.io/esu7t) are available in the 'utilities' folder. These functions create a temporary file with trial indices beginning from 0. Furthermore, to also correct the column 'grammatical_property_trigger' in the behavioural (csv) files, an [additional function](https://osf.io/esu7t) is available in the 'utilities' folder that automatically corrects the appropriate markers from '2' to '3'. - **30 June 2024:** In **Norway** data, fixed behavioural files and vmrk files of Participants 3 and 5 in Session 4. Specifically, the EEG marker corresponding to the property of verb-object agreement was corrected from `S 2` to `S 3`. <br> #### Norwegian site <br> - In some of the first EEG files for Session 2, the resting-state part and the rest of the sesion are contained in the same files. Each part is perfectly identifiable and segmentable based on the markers that are described in the README file contained in the `session_materials` folder in `L3_transfer_longitudinal.zip`. For subsequent instances of Session 2, the two parts are contained in separate files. - The EEG data corresponding to the experiment in **Session 2 of Participant 12** is split between two files, as the recording was interrupted. These files must be appended in BrainVision Analyzer. <br> - **Session 2 of Participant 2**. The EEG markers were not recorded, so the EEG data could not be analysed. <br> - **Session 3 of Participant 5**. The EEG recording stopped halfway. The file was segmented so that the damaged part could be removed, and the preprocessing proceeded as usual. <br> - **Session 4 of Participant 5**. The EEG recording stopped halfway. The file was segmented so that the damaged part could be removed, and the preprocessing proceeded as usual. <br> - **Session 4 of Participant 7** contains an error in the vmrk file. Specifically, in Trial 17, after sending an appropriate marker (`S 40`) corresponding to the target word, BrainVision sent an incorrect marker (`S 45`) corresponding to the target word of the following trial. The incorrect marker corresponded to the following trial, 18 (see excerpt below from the vmrk file). Mk99=Stimulus,S 5,144082,1,0 Mk100=Stimulus,S 40,145028,1,0 Mk101=Stimulus,S 45,145029,1,0 Mk102=Stimulus,S175,145034,1,0 Mk103=Stimulus,S 1,145039,1,0 Mk104=Stimulus,S101,145045,1,0 Mk105=Stimulus,S 6,147243,1,0 Thereafter, the markers corresponding to the target words are incorrect, as they correspond to the preceding trial. In a similar error, in Trial 188, a sentence trigger (`S231`) was wrongly sent after the grammatical property trigger, as shown below. Mk1126=Stimulus,S 5,1217561,1,0 Mk1127=Stimulus,S 47,1218541,1,0 Mk1128=Stimulus,S195,1218546,1,0 Mk1129=Stimulus,S 2,1218551,1,0 Mk1130=Stimulus,S231,1218556,1,0 Mk1131=Stimulus,S103,1218557,1,0 Mk1132=Stimulus,S 6,1220699,1,0 <br> - **Session 4 of Participant 11** does not have EEG data because the trigger port settings failed. Specifically, they were set to the default setting, with four **S**timulus pins and four **R**esponse pins, whereas we need eight response pins to register our triggers. <br> #### Spanish site <br> - No markers for violation of differential object marking in experiments of Sessions 3 and 4. This glitch occurred because markers were not sent with the third word. The markers *cannot* be manually fixed because there are no markers of any kind for these trials. The behavioural data are unaffected. - The second half of **Session 4 of Participant 5** contains an incomplete trial at the top of the vmrk file (`05_s04_cell_2_temp.vmrk`), as shown below. Mk3=Stimulus,S 5,7806,1,0 Mk4=Stimulus,S 7,10532,1,0 Mk5=Stimulus,S 5,11901,1,0 Further in the same file, there are also errors between marker 155 and marker 157, as they do not follow the appropriate order, as shown below. Mk155=Stimulus,S 5,127004,1,0 Mk156=Stimulus,S 6,129763,1,0 Mk157=Stimulus,S 5,131155,1,0 <br> [1]: https://www.editpad.org/ [2]: https://ibb.co/7zCZH3P