**The EEG component of this dataset is available as part of R1.0.0 at https://legacy.openfmri.org/dataset/ds000116/** Data acquisition methods have been described in detail in Walz et al. (2013, J Neurosci), but we outline them here and provide file information for clarity and ease of use. ———————————————————————————————— ODDBALL PARADIGM ———————————————————————————————— 3 runs each of separate auditory and visual tasks 20% target stimuli (requiring button response), 80% standard stimuli (to be ignored) stimulus duration 200 ms ITI 2-3 sec, uniformly distributed first 2 stimuli constrained to be standards each run consisted of 125 total stimuli AUDITORY target - broadband “laser gun” sound standard - 390 Hz tone VISUAL target - large red circle on isoluminant grey background (3.45 degree visual angle) standard - small green circle on isoluminant grey background (1.15 degree visual angle) ———————————————————————————————— EEG DATA ———————————————————————————————— custom built MR-compatible EEG system with differential amplifier custom cap configuration using bipolar electrode pairs and twisted leads 1000 Hz sampling rate 49 channels start triggered by fMRI scan start and EEG clock synced with scanner clock on each TR. CHANNEL KEY chan 1-43 (or 1-34 in the re-referenced electrode space) EEG (for locations refer to cap diagram and bipolar mapping table, described in SUPPLEMENTARY FILES section below) chan 44 EOG - horizontal chan 45 EOG - vertical chan 46-47 (or chan 35 in the re-referenced electrode space) ECG chan 48 (or chan 36 in the re-referenced electrode space) stimulus event markers baseline 0 value jumps to 25 at start of task value 125 is auditory standards value 150 is auditory targets value 225 is visual standards value 250 is visual targets chan 49 (or chan 37 in the re-referenced electrode space) behavioral event markers baseline value 0 with value 100 indicating response RAW EEG DATA (EEG_raw.mat) This is the completely raw EEG recording, contaminated with gradient and BCG artifacts. range is -8 to +8 V (amplifier gain was set to 10K) GRADIENT-FREE EEG DATA (EEG_noGA.mat) units of uV This is the data following gradient-artifact removal using mean (across TRs for each channel) subtraction method and standard filtering. BCG artifact remains. Note: these are the data we used for classification in our study. RE-REFERENCED GRADIENT-FREE EEG DATA (EEG_rereferenced.mat) units of uV This is the same data as above, but in the 34-channel electrode space. Re-referencing is performed via a basic matrix operation using the shortestpath.m (see SUPPLEMENTARY FILES section below). Noisy channels (subjectively determined by visual inspection) were excluded prior to re-referencing; due to oversampled system design, this does not cause loss of electrodes. subject excluded bipolar pair channels sub001 [8 24] sub002 [5 30] sub003 [29 30 34 36] sub004 [23 24 30] sub005 [8 30] sub006 [9 24 30] sub007 [30] sub008 [28 29 30 33 36] sub009 [29 30 33] sub010 [24 29 33] sub011 [16 24 30] sub012 [30] sub013 [30] sub014 [30] sub015 [8 16 28 34 37] sub016 [16 24 30 28 37] sub017 [7 33] ———————————————————————————————— BOLD fMRI DATA ———————————————————————————————— 3T Philips Achieva MR Scanner single channel send and receive head coil EPI sequence 170 TRs per run 2 s TR 25 ms TE 32 slices 3x3x4 mm resolution no slice gap AC-PC alignment The supplied explanatory variable files - cond001.txt - target stimuli - cond002.txt - standard stimuli - cond003.txt - reaction time contain only trials with correct responses and clean EEG. To include all trials, one can easily create new model files from the behavdata.txt files. IMPORTANT NOTES ABOUT PREPROCESSING * Slices were not acquired in interleaved order. Refer to the slice_order.txt file (single column FSL format) if you choose to perform slice timing correction. * The scanner began the EPI pulse sequence a few seconds prior to the start of recording, so there is no need to discard the first few TRs, as is commonly done with fMRI data. * EEG wires can create field inhomogeneities that are sometimes visible on the images, particularly on the left posterior region closest to where the wires were bunched together as they exited the coil. Bias field correction can optionally be performed using FSL FAST without the need of a field map. ———————————————————————————————— ANATOMICAL MRI DATA ———————————————————————————————— - highresSPGR001.nii.gz - SPGR 1x1x1 mm (present for all subjects except sub010) - highresMPRAGE001.nii.gz - MPRAGE 1x1x1 mm (for sub010 and additionally exists for a few others) - highresEPI001.nii.gz - high resolution single volume EPI 2x2x2 mm ———————————————————————————————— SUPPLEMENTARY FILES ———————————————————————————————— EEG RE-REFERENCING - shortestpath.m - Matlab function used to re-reference EEG data from 43-channel bipolar pair space to 34-channel electrode space. See function help for more info. Dependencies are efmri36mastoids.ced location file (supplied) and EEGLAB readlocs.m function (free download from UCSD). ELECTRODE LOCATIONS - efmri36mastoids.ced - location file for re-referenced data that includes mastoid channels. - locations_34_EEG.ced - location file for 34 electrodes. - EEGfMRI_cap_diagram.png - for easy reference. - EEGfMRI_cap_bipolar_pair_mapping.pdf - table showing mapping from bipolar pair space to electrode space. fMRI ACQUISITION - slice_order.txt - single column FSL format custom slice timing correction file. This is the order in which slices were acquired by the scanner (i.e. not interleaved). ———————————————————————————————— PUBLICATIONS ———————————————————————————————— Walz JM, Goldman RI, Carapezza M, Muraskin J, Brown TR, Sajda P (2015) “Prestimulus EEG Alpha Oscillations Modulate Task-Related fMRI BOLD Responses to Auditory Stimuli,” Neuroimage 113:153-163. doi: 10.1016/j.neuroimage.2015.03.028. Conroy B, Walz JM, Cheung B, Sajda P (2014) “Fast Simultaneous Training of Generalized Linear Models (FaSTGLZ)” arXiv 1307.8430. Walz JM, Goldman RI, Carapezza M, Muraskin J, Brown TR, Sajda P (2013) “Simultaneous EEG-fMRI Reveals Temporal Evolution of Coupling between Supramodal Cortical Attention Networks and the Brainstem,” J Neurosci 33(49):19212-22. doi: 10.1523/JNEUROSCI.2649-13.2013. Walz JM, Goldman RI, Carapezza M, Muraskin J, Brown TR, Sajda P (2013) “Simultaneous EEG-fMRI reveals a temporal cascade of task-related and default-mode activations during a simple target detection task,” Neuroimage. 2013 Aug 17. pii: S1053-8119(13)00868-9. doi: 10.1016/j.neuroimage.2013.08.014. Conroy BR, Walz JM, Sajda P (2013) “Fast bootstrapping and permutation testing for assessing reproducibility and interpretability of multivariate FMRI decoding models,” PLoS One. 2013 Nov 14;8(11):e79271. doi: 10.1371/journal.pone.0079271.