When Age Tips the Balance: a Dual Mechanism Affecting Hemispheric Specialization for Language

---

Contents

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

---

Background

Aging is accompanied by changes in brain architecture that alter the lateralization of functional networks. In this study, we examined how hemispheric specialization changes across the adult lifespan by analyzing resting-state fMRI and structural MRI data from 728 typical adults aged 18 to 88. Using the Language-and-Memory Network atlas, we quantified regional asymmetries in functional connectivity along the cortex’s principal gradient, and normalized regional volumes across 37 bilateral regions. We identified two distinct age-related asymmetry trajectories: one pattern revealed a bilateralization of language-dominant regions, while the other showed increasing leftward specialization in multimodal regions associated with memory and language. These opposing patterns emerged around midlife and were linked to performance in language production tasks. By integrating connectivity gradients, structural asymmetries, and behavioral data, our findings provide new evidence for a dual mechanism reshaping functional brain lateralization with age and demonstrate the utility of resting-state metrics in tracking these shifts.

---

Reference

For usage of the manuscript, please cite:

• Labache, L. $^{†}$, Roger, E. $^{†}$, Hamlin, N., Kruse, J., Baciu, M., & Doucet, G. E. (2025). When age tips the balance: A dual mechanism affecting hemispheric specialization for language. Imaging Neuroscience. DOI: 10.1162/IMAG.a.63. $^{†}$ these authors contributed equally.

For usage of the associated code, please also cite:

• Labache, L., Roger, E., Hamlin, N., Kruse, J., Baciu, M., & Doucet, G. E. (2023). When age tips the balance: a dual mechanisms affecting hemispheric specialization for language. loiclabache/RogerLabache_2023_LanguAging. DOI: 10.5281/zenodo.10253278
• The implementations of the Generalized Additive Mixed Models and Partition Around Medoids algorithm were adapted from Roe, J., et al. 2021. DOI: 10.1038/s41467-021-21057-y, and related GitHub repository: AgeSym.

---

Code Release

The Script folder includes three R scripts. The three R scripts are designed to facilitate the replication of results as detailed in the Method Section of the manuscript.

• 1_GAMM_hROIs.R: R script to model gradient asymmetry trajectories throughout life using factor-smooth Generalized Additive Mixed Models. The script allows to compute the asymmetry trajectories underlying the interaction Hemisphere×Age and their confidence intervals. This script also assesses the significance of the smooth Hemisphere×Age interaction by testing for a difference in the smooth term of Age between hemispheres. We applied a False Discovery Rate correction to control for the number of tests conducted.
• 2_PAM_Clustering.R: R script to classify regions in the Language-and-Memory network that demonstrate a significant Hemisphere×Age interaction, based on their functional asymmetry skewness profiles. This script also allows to compute the intersection point between the two average clusters curves.
• 3_CCA_BrainCognitionAssociation.R: R script to proceed with the Canonical Correlation Analysis to assess brain–behavior Associations.

The Data folder contains output files generated from the Generalized Additive Mixed Models analysis.

Note that the R scripts also contain the code to reproduce the figures found in the manuscript. The brain renderings in the paper require a customized version of Surf Ice that we will be happy to share on demand.

---

Atlas Used

The atlas used in the paper is available in the Atlas folder.

• The Language-and-Memory atlas provides an atlas in standardized MNI volume space of 74 sentence- and memory-related areas (37 by hemisphere). The Language-and-Memory atlas encompasses the core regions that compose the stable components for language and memory. The Language-and-Memory atlas is composed of multiple brain regions provided by task-fMRI: one cross-sectional study for language (see Labache, L., et al. 2019, Github repository: SENSAAS) and one meta-analysis for memory (see Spaniol, J., et al. 2009). The compilation of the Language-and-Memory atlas was initially undertaken in the following paper: Roger, E., et al. 2020.
  • The file Atlas/language_memory_atlas.nii.gz contains the Volumetric Language-and-Memory atlas (in MNI ICBM 152 space).
  • Atlas/language_memory_atlas.txt: text file containing a full description of each Language-and-Memory areas. The first column Abbreviation is the abbreviation of a region. The second column Region is the full anatomical label of a region. Hemisphere refers to the cerebral hemisphere to which a region belongs (“L” for left, “R” for right). Function indicates if a regions process language (“L”), memory (“M”), or language and memory (“LM”). Index is the index of each region that is used in the NIfTI file. Number of Voxels is the number of voxels of each region for the 2mm version of the atlas. The MNI coordinate (columns Xmm, Ymm, Zmm) of each regions centroid is also provided.

---

Other related papers that might interest you

• Language-and-Memory Network seminal paper: Roger, E., et al. 2020. DOI: 10.1002/hbm.24839
• Influence of Language Lateralisation on Gradient Asymmetry: Labache, L., et al. 2023. DOI: 10.1038/s41467-023-39131-y, and related GitHub repository: Labache_2022_AO
• Sentence Supramodal Areas Atlas; Labache, L., et al. 2019. DOI: 10.1007/s00429-018-1810-2, and related GitHub repository: SENSAAS
• For additional reading on GAMMs, consult Gavin Simpson’s procedure for comparing smooth terms: Comparing smooths in factor-smooth interactions (1/2), and Comparing smooths in factor-smooth interactions (2/2)

---

Questions

Please contact me (Loïc Labache) at: loic.labache@rutgers.edu, or Élise Roger at: elise.roger@umontreal.ca