loiclabache / labache_2022_ao

Code to reproduce results from Labache, L., et al. 2023. Language network lateralization is reflected throughout the macroscale functional organization of cortex. DOI: 10.1038/s41467-023-39131-y

Language network lateralization is reflected throughout the macroscale functional organization of cortex

---

Contents

• Background
• Reference
• Code Release
• Data
• Atlas Used
• Other related papers that might interest you
• Questions

---

Background

Hemispheric specialization is a fundamental feature of human brain organization. However, it is not yet clear to what extent the lateralization of specific cognitive processes may be evident throughout the broad functional architecture of cortex. While the majority of people exhibit left-hemispheric language dominance, a substantial minority of the population shows reverse lateralization. Using twin and family data from the Human Connectome Project, we provide evidence that atypical language dominance is associated with global shifts in cortical organization. Individuals with atypical language organization exhibit corresponding hemispheric differences in the macroscale functional gradients that situate discrete large-scale networks along a continuous spectrum, extending from unimodal through association territories. Analyses reveal that both language lateralization and gradient asymmetries are, in part, driven by genetic factors. These findings pave the way for a deeper understanding of the origins and relationships linking population-level variability in hemispheric specialization and global properties of cortical organization.

---

Reference

For usage of the manuscript, please cite:

• Labache, L., Ge, T., Yeo, B.T.T. et al. Language network lateralization is reflected throughout the macroscale functional organization of cortex. Nat Commun 14, 3405 (2023). DOI: 10.1038/s41467-023-39131-y

For usage of the associated code, please also cite:

• Labache, L., Ge, T., Yeo, B.T. T., Holmes, A. J. (2023). Language network lateralization is reflected throughout the macroscale functional organization of cortex. loiclabache/Labache_2022_AO. DOI: 10.5281/zenodo.7869040

---

Code Release

The Script folder contains 2 sub-folders: Analysis and Visualization.

The Analysis folder contains the scripts to download resting-state and language task fMRI data (from the Amazon Web Services: AWS, S3 bucket), compute language lateralization metrics and gradients, and analyze data.

• step_1_download_rest_fMRI_HCP_data.R: R script to download resting-state data for chosen participants. Requires to create AWS credentials through the ConnectomeDB website. Only the rs-BOLD time series from the language network (SENT_CORE) are downloaded.
• step_2_download_task_fMRI_HCP_data.R: R script to download language task data for chosen participants (aka files named LANGUAGE_level2_hp200_s4.dscalar.nii on HCP server). Requires to create AWS credentials through the ConnectomeDB website.
• step_3_splitting_areal_tfMRI.sh: shell script to split the fMRI activations of the language task between the left and right hemispheres. Requires Connectome Workbench.
• step_4_language_metrics_computation.R: R script to compute the 5 language lateralization metrics for each participant (data available in the Data folder: 995participants_language_metrics_HCP.xlsx):
  • average asymmetry of activation (left-right hemisphere) of the story-math fMRI contrast at the network level,
  • average asymmetry of activation (left-right hemisphere) of the story-math fMRI contrast at the hub level,
  • average strength sum (left+right hemisphere) during rs-fMRI of the language network,
  • average strength asymmetry (left-right hemisphere) during rs-fMRI of the language network,
  • average inter-hemispheric homotopic connectivity strength during rs-fMRI of the language network.
• step_5_participants_classification.R: R script to classify the 995 participants. Each participant is characterized by the 5 language lateralization metrics.
• step_6_whole_brain_fc_matrices.R: R script to compute the whole brain connectivity matrix (384 x 384, AICHA atlas) for each participant.
• step_7_1_gradient_computation.py: Python script to compute the first 3 functional gradients as defined by (Margulies, D., et al. 2016), using the Python library BrainSpace. The script requires group level correlation matrix, available there: Data/groupLevel_correlationMatrix.txt, and the group level gradient values, available there:Data/groupLevel_gradient.csv.
• step_7_2_gradient_computation.R: R script to create a compiled file of all the gradient values for all participant. The normalized gradients at the network level for each participant can be found there: Data/994participants_gradients_network.xlsx. The correspondence between the AICHA regions and the 7 networks from Yeo, B.T. T., et al. 2011 can be found there: Atlas/AICHA/overlap_AICHA_Yeo_7Network_maskYeo_SchaeferAtlas.csv.
• step_8_1_prepare_data_for_heritability_analysis.R: R script to create an compatible csv file for Solar, a software allowing to perform heritability analysis.
• step_8_2_heritability_gradient_network.R: R script to perform the heritability analysis.

The Visualization folder contains R files (FigX_script.R) used to generate each figures included in the paper. Each script corresponds to a figure or a panel. The brain renderings in the paper require a customized version of Surf Ice that we will be happy to share on demand.

---

Data

All the data used are brain fMRI data from the Human Connectome Project (HCP; n=995). The HCP data requires data access permission and if you have this, we are happy to share the processed data used here.

• The primary data used are the 5 functional features characterizing the high-order language network (see SENSAAS). These 5 features have been previously shown to accurately determine the language network lateralization at the individual level (Labache, L., et al. 2020). The 5 language lateralization metrics are availabe there: Data\995participants_language_metrics_HCP.xlsx (participants identifiers are anonymized).

• The second set of data used are the region-level functional connectivity matrices (384 × 384 AICHA brain regions). Those matrices have been used to compute the first 3 functional gradients (Margulies, D., et al. 2016). The individual normalized network gradient values are available there: Data/994participants_gradients_network.xlsx.

All data computed in the paper and to generate figures are provided in Data folder.

The file Data\infos_995participants.csv does not correspond to the HCP identifiers of each participant, but an exemple file to run the scripts.

The data to reproduce Figure 1C, 3A, 3B, and 3C are available in the excel file Data/Source Data.xlsx.

---

Atlas Used

The atlas used in te paper are available in the Atlas folder. This folder contains 2 sub-folders: SENSAAS and AICHA.

• SENSAAS provide an atlas in standardized MNI volume space of 32 sentence-related areas based on a 3-step method combining the analysis of activation and asymmetry during multiple language tasks with hierarchical clustering of resting-state connectivity and graph analyses. The temporal correlations at rest between these 32 regions made it possible to detect their belonging to 3 networks. Among these networks, one, including 18 regions, contains the essential language areas (SENT_CORE network), i.e. those whose lesion would cause an alteration in the understanding of speech. Full description of the language atlas can be found there: SENSAAS, and the related paper there: Labache, L., et al. 2019.
  • The volumetric (in the MNI ICBM 152 space) and area (32k_fs_LR space) atlas are available in the Atlas/SENSAAS folder. The sub-folder Volumetric contains the volumetric SENSAAS atlas: SENSAAS_MNI_ICBM_152_2mm.nii, and a CSV file containing a full description of each language areas: SENSAAS_description.csv. The sub-folder Area contains the area SENSAAS atlas in the left (S1200_binarySentCore_L_surface.shape.gii) and right hemisphere (S1200_binarySentCore_R_surface.shape.gii), as well as the hub atlas (i.e. regions STS3, STS4 and F3t only) in the left (S1200_binaryHubsSentCore_L_surface.shape.gii) and right hemisphere (S1200_binaryHubsSentCore_R_surface.shape.gii).
    • Briefly, the hub language network atlas corresponded to the inferior frontal gyrus (Broca’s area, F3t) and to the posterior aspect of the superior temporal sulcus (corresponding to Wernicke’s area, STS3 and STS4).

• AICHA is a functional brain ROIs atlas (384 brain regions) based on resting-state fMRI data acquired in 281 individuals. AICHA ROIs cover the whole cerebrum, each having 1) homogeneity of its constituting voxels intrinsic activity, and 2) a unique homotopic contralateral counterpart with which it has maximal intrinsic connectivity. Full description of the atlas can be found there: AICHA, and the related paper there: Joliot, M., et al. 2015.
  • The version of AICHA used in the paper is available in the Atlas\AICHA folder: AICHA.nii (MNI ICBM 152 space). AICHA_vol3.txt is a description of each atlas’ regions. Readme_AICHA.pdf is the user manual.

---

Other related papers that might interest you

• Sentence Supramodal Areas Atlas: SENSAAS
• Typical and Atypical Language Brain Organization: Labache, L., et al. 2020. DOI: 10.7554/eLife.58722

---

Questions

Please contact me (Loïc Labache) as [email protected] and/or [email protected]