Skip to main content

BIL&GIN sentence and rest asymmetries

Cite this dataset

Mazoyer, Bernard et al. (2020). BIL&GIN sentence and rest asymmetries [Dataset]. Dryad.


This dataset has been generated from a subsample of 287 participants (150 left-handers) from the BIL&GIN database (Mazoyer et al. Neuroimage 2016, 124:1225-1231). The dataset contains age, gender, self-reported handedness and Edinburgh score, for each participant, as well as its education level. It also includes measures of FMRI task-induced BOLD-signal variations and left minus asymmetry during sentence processing minus word list in 3 conditions: production, listening and reading. These variations are measured for the whole sentence core network as well as for its hubs as defined in the SENSAAS atlas (Labache et al.  Brain Structure and Function 2019, 224:859-882). Both intrahemispheric and interhemispheric measures of resting-state intrinsic connectivity in the same sets of regions are also included. Brain language hemispheric dominance pattern is given based on a multitask multimodal hierarchical classification using as entry variables:1) task-induced asymmetries (2) intrahemispheric resting-state degree connectivity and asymmetry, and 3) interhemispheric connectivity applied. Other brain lateralization patterns for sentence production only based a hemispheric functional asymmetry index (Mazoyer et al. PLoS One 2014, 9(6):e101165) and on a SVM approach (Zago et al. Human Brain Mapping 2017, 12:5871-5889) are also given. Morphometric data are included as well, including total intracranial volume, hemisperic volumes and asymmetry of grey and white matter and  corpus callosum volume area at the mid-sagittal level. Scores to a cognitive battery of 11 tests adjusted for age and education level are provided as well as loadings of the first 4 principal components computed from these scores.


We have extracted hand lateralization, sex , age, eduction level, MRI (structural and functional for language tasks and resting-state), and cognitive skills from 287 participants of the BIL&GIN database. Data acquisition and pre-processing of these data have been detailed in Mazoyer et al. Neuroimage 2016, 124:1225-1231.

Processing of FMRI data (both language-task  and resting-state related) have been detailed in Labache et al.  Brain Structure and Function 2019, 224:859-882.

Language lateralization derived from an hemispheric functional lateralization index during sentence production has been described in Mazoyer et al. PLoS One 2014, 9(6):e101165.

Language dominance pattern assessed with a support vector machine has been detailed in Zago et al. Human Brain Mapping 2017, 12:5871-5889.

Multimodal hierarchical classification was completed with R (R version 3.5.1; R Core Team, 2013), while other statistical analyses were performed using JMP15 (, SAS Institute Inc., 2018).

Usage notes

There are some missing data for the cognitive skill variables:

1 for Residuals_Maze, 2 for Residuals_ravenQI, and 3 for PC1_SPA, PC2_MEM, PC3_PHONO, and PC4_VERB