Longitudinal Stability and Change in Brain-Cognition Relationship in the ABCD Study
Poster No:
961
Submission Type:
Abstract Submission
Authors:
Yapei Xie1, Shaoshi Zhang1, Leon Ooi1, Csaba Orban1, Ru Kong1, Dorothea Floris2, Zuo Xi-Nian3, Elvisha Dhamala4, Avram Holmes5, Lucina Uddin6, Thomas Nichols7, Adriana Martino8, Thomas Yeo1
Institutions:
1National University of Singapore, Singapore, 2University of Zurich, Zurich, Zurich, 3Beijing Normal University, Beijing, Beijing, 4Feinstein Institutes for Medical Research, Glen Oaks, NY, 5Rutgers University, New Brunswick, NJ, 6Department of Psychology, University of California Los Angeles, Los Angeles, CA, 7University of Oxford, Oxford, Oxfordshire, 8Child Mind Institute, New York, NY
First Author:
Co-Author(s):
Introduction:
The transition from childhood to adolescence is a critical neurodevelopment period, but how brain networks and cognition co-evolve over time during this period remains unclear. Previous longitudinal studies mostly do not disentangle longitudinal effects from cross-sectional ones [1-2], leading to potential misinterpretations [3-5]. Here, we address this gap by comprehensively investigating the co-evolution of functional connectivity (FC) and cognition using longitudinal resting-state fMRI and cognitive performance data at baseline and Year 2 of the ABCD study [6].
Methods:
We included 2,949 unrelated participants from the ABCD study, with resting-state fMRI and seven cognitive tests at baseline (age: 9.93 ± 0.62 years) and Year 2 (age: 11.94 ± 0.65 years) across 22 sites. A 419 × 419 FC matrix was constructed by combining the 400-region Yan parcellation [7] with 19 subcortical ROIs [8] and calculating Pearson correlation coefficients between the average time series of each ROI pair. Overall cognitive ability was represented by the first principal component (PC1) of the seven cognitive measures. Cognitive change and stability were assessed by correlating baseline cognition with its rate of change and Year 2 scores, respectively. FC change and stability were calculated similarly. We applied a Kernel ridge regression (KRR) framework as in our previous studies [9-10] to explore the co-evolution between FC and cognition during the transition to adolescence.
Results:
The Spearman's correlations between baseline cognition and cognitive change rates were negative (ρ: -0.61 to -0.26) and significant for all 8 cognitive measures (FDR q < 0.05; Figure 1A), indicating a "catch-up" effect where children with lower baseline cognition showed greater gains. However, the significant positive correlations (ρ: 0.38 to 0.80) between baseline and Year 2 cognition (FDR q < 0.05; Figure 1B) suggest that children with higher baseline cognition retained their advantage, despite catch-up growth. For FC, regional variation in the catch-up effect was observed along the sensory-association axis, with sensory-motor and visual systems showing the greatest catch-up effect and heteromodal association cortex the least (ρ range: -0.71 to -0.2; Figure 1C&D). FC stability showed the opposite pattern (ρ range: 0.01 to 0.86; Figure 1E&F). We also observed high overall stability in the FC matrix between the two time points for each individual (ranged from 0.34 to 0.88; Figure 1G). Notably, the FC matrix from one timepoint could reliably identify individuals at the other timepoint with high accuracy (Figure 1H).
Turning to the relationship between FC and cognitive performance, we found that baseline FC predicted cognitive performance better at Year 2 than at baseline; Year 2 FC's prediction of Year 2 cognition was better than baseline FC's prediction of Year 2 cognitive measures (Figure 2A&B). The predictive network features of the three models were highly similar, with correlations ranging from 0.91 to 0.99 (Figure 2C). Models trained on baseline FC to predict baseline cognition also improved in accuracy when applied to predict Year 2 cognition with Year 2 FC (Figure 2D). Intriguingly, baseline FC emerged as a stronger predictor of Year 2 cognition than longitudinal FC rate of change (Figure 2E), while the ability of both baseline FC and longitudinal FC changes to predict cognitive growth rates remained limited (Figure 2F).
Turning to the relationship between FC and cognitive performance, we found that baseline FC predicted cognitive performance better at Year 2 than at baseline; Year 2 FC's prediction of Year 2 cognition was better than baseline FC's prediction of Year 2 cognitive measures (Figure 2A&B). The predictive network features of the three models were highly similar, with correlations ranging from 0.91 to 0.99 (Figure 2C). Models trained on baseline FC to predict baseline cognition also improved in accuracy when applied to predict Year 2 cognition with Year 2 FC (Figure 2D). Intriguingly, baseline FC emerged as a stronger predictor of Year 2 cognition than longitudinal FC rate of change (Figure 2E), while the ability of both baseline FC and longitudinal FC changes to predict cognitive growth rates remained limited (Figure 2F).
Conclusions:
Our study reveals a strong and stable relationship between cognition and brain functional network architecture that persists and even strengthens during the transition from childhood to adolescence, although dynamic developmental changes in cognition and functional connectome remain largely decoupled (See the conceptual model in Figure 2G).
Lifespan Development:
Early life, Adolescence, Aging 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
fMRI Connectivity and Network Modeling
Multivariate Approaches
Keywords:
Cognition
Development
FUNCTIONAL MRI
Machine Learning
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
2. Dong, H., et al. (2024). Ventral attention network connectivity is linked to cortical maturation and cognitive ability in childhood. Nature Neuroscience, 27(9), 1234–1245.
3. Guillaume, B., et al. (2014). Fast and accurate modelling of longitudinal and repeated measures neuroimaging data. NeuroImage, 94, 287–302.
4. Sorensen, L., et al. (2021). A recipe for accurate estimation of lifespan brain trajectories, distinguishing longitudinal and cohort effects. NeuroImage, 226, 117596.
5. Kang, X., et al. (2024). Study design features increase replicability in brain-wide association studies. Nature, 589(7843), 123–130.
6. Casey, B. J., et al. (2018). The Adolescent Brain Cognitive Development (ABCD) study: Imaging acquisition across 21 sites. Developmental Cognitive Neuroscience, 32, 43–54.
7. Yan, X., (2023). Homotopic local-global parcellation of the human cerebral cortex from resting-state functional connectivity. NeuroImage, 273, 120010.
8. Fischl, B., (2002). Whole brain segmentation: Automated labeling of neuroanatomical structures in the human brain. Neuron, 33(3), 341–355.
9. Chen, J., et al. (2022). Shared and unique brain network features predict cognitive, personality, and mental health scores in the ABCD study. Nature Communications, 13, 2217.
10. Ooi, L. Q. R., et al. (2022). Comparison of individualized behavioral predictions across anatomical, diffusion, and functional connectivity MRI. NeuroImage, 263, 119603.