Prediction of Brain Resilience and Identification of Brain Features Predictive of Resilience
Poster No:
874
Submission Type:
Abstract Submission
Authors:
Byeongwook Lee1, Kaustubh Supekar1, Srikanth Ryali1, Harinarayana Mellacheruvu1, Vinod Menon1
Institutions:
1Stanford University, Stanford, CA
First Author:
Co-Author(s):
Introduction:
Unprecedented population aging and escalating prevalence of neurodegenerative disorders pose profound challenges to public health and society. Deepening our understanding of the complex mechanisms underlying brain resilience to aging and neurodegeneration could inform strategies to enhance brain health and resilience. We utilized advanced AI, neuroimaging, and multi-scale biological and environmental data to investigate individual brain resilience to aging and identify its key determinants across circuit, molecular, cellular, and environmental scales (Figure 1A)
·Figure 1
Methods:
Spatiotemporal deep neural networks (stDNN) model with explainable AI approaches were applied to predict individual brain resilience to aging and identify functional brain features from fMRI data (Figure 1B). The model was trained to predict age in healthy individuals with residuals (actual minus predicted ages) serving as a measure of resilience to aging. Integrated gradients (IG) analysis was used to identify dynamic functional brain circuit features contributing to age-related brain resilience. Finally, we explored relationships between model-derived age-related brain resilience measures and cognitive, molecular, cellular, and exposomic factors.
Results:
Application of stDNN to fMRI timeseries data from 262 healthy older adults (ages 60+) in the NKI cohort, predicted ages accurately (Pearson r = 0.9, Figure 2A). Notably, individual differences between actual and predicted ages revealed greater brain resilience to aging in some individuals. IG analysis identified brain areas-the hippocampus, angular gyrus nodes of the DMN, and DLPFC, PPC nodes of the FPN along with M1, Thalamus, and Caudate-that significantly contribute to resilience against aging (Figure 2B). Furthermore, the model trained on NKI data accurately predicted ages (Pearson r = 0.72) in 398 older adults (60+) from the HCP-Aging cohort without further training, identifying similar predictive features in the DMN and FPN. These results demonstrate the model's ability to accurately estimate individual brain resilience to aging and identify highly replicable brain features of age-related resilience across multiple cohorts.
Further analysis revealed that individuals with the most brain resilience to aging, as measured by a larger positive difference in the model predicted age and the actual age, had superior cognitive functioning and executive function6, as well as lower impulsivity (Figure 2C). We also identified that receptor profiles, especially the spatial distribution of NMDA receptors, play a significant role in shaping the IG-derived functional brain features predictive of age (Figure 2D). Additionally, we revealed that elevated quinolinic acid, a neurotoxin and NMDA receptor agonist that is closely associated with neurodegenerative diseases, was uniquely associated with poor resilience (Figure 2E), independent of biological age. Finally, we identified that higher resilience to aging is correlated with lower odds of allergies, Type 2 diabetes, high lead blood level, sleep disturbance, and substance use (Figure 2F). These findings highlight the multi-scale factors contributing to age-related brain resilience.
Further analysis revealed that individuals with the most brain resilience to aging, as measured by a larger positive difference in the model predicted age and the actual age, had superior cognitive functioning and executive function6, as well as lower impulsivity (Figure 2C). We also identified that receptor profiles, especially the spatial distribution of NMDA receptors, play a significant role in shaping the IG-derived functional brain features predictive of age (Figure 2D). Additionally, we revealed that elevated quinolinic acid, a neurotoxin and NMDA receptor agonist that is closely associated with neurodegenerative diseases, was uniquely associated with poor resilience (Figure 2E), independent of biological age. Finally, we identified that higher resilience to aging is correlated with lower odds of allergies, Type 2 diabetes, high lead blood level, sleep disturbance, and substance use (Figure 2F). These findings highlight the multi-scale factors contributing to age-related brain resilience.
·Figure 2
Conclusions:
Existing approaches to studying brain aging and AD often focus on single levels of analysis, such as brain structure or molecular pathology, limiting our ability to capture the complex, multi-scale nature of resilience mechanisms. Our study addresses this critical gap by leveraging cutting-edge AI techniques to integrate data across scales, from molecular and cellular processes to brain networks and behavior. This multi-scale approach enabled us to identify key drivers of brain resilience and their interactions, providing a more comprehensive understanding of the factors that promote healthy brain aging.
Lifespan Development:
Aging 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
fMRI Connectivity and Network Modeling
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
Aging
FUNCTIONAL MRI
Modeling
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?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
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.
Hou, Y. et al. Ageing as a risk factor for neurodegenerative disease. Nature Reviews Neurology 15, 565-581 (2019).
Ryali, S., Zhang, Y., de Los Angeles, C., Supekar, K. & Menon, V. Deep learning models reveal replicable, generalizable, and behaviorally relevant sex differences in human functional brain organization. Proceedings of the National Academy of Sciences 121, e2310012121 (2024).
Supekar, K. et al. Deep learning identifies robust gender differences in functional brain organization and their dissociable links to clinical symptoms in autism. The British Journal of Psychiatry 220, 202-209 (2022).
Supekar, K. et al. Robust, generalizable, and interpretable artificial intelligence–derived brain fingerprints of autism and social communication symptom severity. Biological Psychiatry 92, 643-653 (2022).
Nooner, K. B. et al. The NKI-Rockland sample: a model for accelerating the pace of discovery science in psychiatry. Frontiers in neuroscience 6, 152 (2012).
Lugo-Huitrón, R. et al. Quinolinic acid: an endogenous neurotoxin with multiple targets. Oxid Med Cell Longev 2013, 104024 (2013). https://doi.org:10.1155/2013/104024
HEYES, M. P. et al. Quinolinic acid and kynurenine pathway metabolism in inflammatory and non-inflammatory neurological disease. Brain 115, 1249-1273 (1992).
Moresco, R. et al. Quinolinic acid induced neurodegeneration in the striatum: a combined in vivo and in vitro analysis of receptor changes and microglia activation. European journal of nuclear medicine and molecular imaging 35, 704-715 (2008).