Data for: Intracranial entrainment reveals statistical learning across levels of abstraction
Data files
Jun 08, 2023 version files 9.40 GB
Abstract
The following submission contains the data reported from the manuscript "Intracranial entrainment reveals statistical learning across levels of abstraction". The dataset was obtained from 8 neurosurgical patients who had intracranially implanted electrodes for seizure monitoring. Intracranial EEG (iEEG) data were recorded while the patients viewed a rapid stream of scene images. In the Category-Level Structured condition, patients viewed a series of trial-unique scene images, in which the categories of scenes were paired across repetitions (e.g., images of beach always followed by images of canyons). In the Exemplar-level Structured condition, participants viewed a sequence of 6 repeating scene images (from non-overlapping scene categories), which were paired across repetitions (e.g., image A always followed by image B). In the baseline Random condition, participants again viewed a sequence of 6 repeating scene images, but the images were presented in a random temporal order. Each image was presented for 250 ms, followed by a 250 ms inter-stimulus-interval period.
The data presented here contain the raw iEEG data from the 8 patients during these task conditions, as well as information about the anatomical placement of their electrodes.
Methods
iEEG data were recorded from 8 patients who had electrodes surgically implanted for seizure monitoring. EEG data were recorded on a NATUS NeuroWorks EEG recording system. Data were collected at a sampling rate of 4096 Hz. Signals were referenced to an electrode chosen by the clinical team to minimize noise in the recording. To synchronize EEG signals with the experimental task, a custom-configured data acquisition system (DAQ) was used to convert signals from the research computer to 8-bit "triggers" that were inserted into an open EEG channel (the TRIG channel).
The iEEG data presented here are the raw data. The raw data have not been processed, except that the date information in the metadata has been randomized to preserve anonymity.
Electrode contact locations were identified using post-operative CT and MRI scans. Reconstructions were completed in BioImage Suite and were subsequently registered to the patient's pre-operative MRI scan, resulting in contact locations projected into the patient's pre-operative space. The resulting files were converted from the Bioimagesuite format (.MGRID) into native space coordinates using FieldTrip functions. The coordinates were then used to create a mask in FSL in each participant's native space, with the coordinates of each contact occupying one voxel in the mask. We then registered each patient's pre-operative anatomical scan to the MNI T1 2mm standard brain using linear registration. We then used this registration matrix to transform each electrode mask into standard space. The electrode locations reported here are in standard space.
Usage notes
The iEEG files are in .edf (European Data Format). They can be loaded in by programs such as MATLAB (using the Fieldtrip or EEGlab toolboxes) or Python (using libraries such as MNE).
The iEEG files also contain accompanying files (.json and .tsv) which can be read in by standard text editors.
The electrode location files is a .csv which can be read in by standard text editors or spreadsheet programs, such as Excel.