Computational Models integrating Neurobiology and Social determinants of Brain Health

Brain clocks, which measure deviations between chronological age and predicted brain age (brain age gaps, or BAGs), reveal whether an individual's brain appears older or younger than expected. These models can characterize individual aging trajectories and identify pathological deviations in conditions like Alzheimer's Disease (AD) and behavioral variant frontotemporal dementia (bvFTD). However, challenges such as heterogeneous findings, limited computational methods, and a lack of diverse representation hinder progress in the field. Emerging biophysical whole-brain models offer a promising solution, enabling robust analyses with moderate sample sizes and shedding light on unexplored causal mechanisms. In parallel, creative and artistic experiences have been proposed as interventions to support brain health across the lifespan. Here we combine source space connectivity (EEG) with generative brain modeling in healthy controls (HCs) from both the global south and north, alongside Alzheimer’s disease (AD) and behavioral variant frontotemporal dementia (bvFTD) patients (N=1,399). Additionally, we analyzed how different types of creative experiences can protect the brain from accelerated brain aging via neural plasticity mechanisms, including a subsample of expert and matched non-expert participants in dance, music, visual arts, and video games, along with a pre/post-learning study (N=227). Using source-space EEG connectivity and Support Vector Machine (SVM) regression models, we examined how BAGs reflect disparities at both aggregate (e.g., geographic and income differences), individual levels (e.g., education and sex), and how can be used to characterize the effects of creativity on brain health (expertise and learning). The BAGs in aging were modulated by diversity-related factors inducing accelerated aging, including geography (south>north), income (GPD, low > high), sex (female>male), and education (low > high). A larger BAG was observed in patients, with sex further increasing the effects in AD (female>male). In creative experiences, we observed delayed brain age across all domains, and scalable effects (expertise>learning). The higher the level of expertise and performance, the greater the delay in brain age. Biophysical modeling shows that BAGs are related to specific mechanisms: global hyperexcitability and reduced connectivity (structural disintegration) were implicated in aging’s BAGs. Hypoexcitability and severe disintegration were related to dementia. On the other hand, creativity is associated with reduced hyperexcitability, increased network efficiency, and increased structural integrity. Our work sheds light on biophysical mechanisms of accelerated aging in diverse and underserved populations and provides domain-independent evidence of creativity’s positive impact on brain health. 

Allostatic-interoception, the anticipation and sensing of bodily signals, is linked to sociocognitive function and cardiovascular health. Impairments in this process may arise in behavioral-variant frontotemporal dementia (bvFTD) and Alzheimer’s disease (AD) due to structural and functional disruptions in the allostatic-interoception network (AIN). Here, we integrate three complementary studies to investigate the AIN's role in sociocognitive processes and cardiovascular risk in neurodegeneration. In Study 1, we conducted PRISMA Activated Likelihood Estimate (ALE) meta-analyses on 170 studies (N=7,032), including structural and functional neuroimaging data. We identified key AIN regions (e.g., insula, amygdala, orbitofrontal cortex, anterior cingulate cortex, thalamus, hippocampus) across interoception, emotion, and social cognition domains in all neurodegenerative diseases combined, with more pronounced involvement in bvFTD in patient specific analyses. Study 2 examined multimodal neuroimaging in 1,501 participants (bvFTD: N=304; AD: N=512; controls: N=685) and found that higher cardiovascular risk (e.g., BMI, hypertension, diabetes, smoking) correlated with reduced structural integrity and functional connectivity within the AIN. Functional disruptions were widespread in bvFTD (bilateral insula, cingulate, orbitofrontal cortex, thalamus, hippocampus) and localized in AD (hippocampus, parahippocampus). In Study 3, we analyzed intrinsic neural timescales of interoception using EEG (N=112; bvFTD: N=31; AD: N=35; controls: N=46). Longer autocorrelation windows (ACW) during interoception were observed in bvFTD (frontotemporal/parietal clusters) and AD (central/occipital-parietal clusters). Longer ACW correlated with poorer sociocognitive performance and AIN structural changes in bvFTD. Our findings reveal consistent AIN involvement in bvFTD and partial involvement in AD, supporting predictive coding theories of allostatic-interoception. Increased cardiovascular risk exacerbates AIN impairments, highlighting the need to address systemic health factors in dementia care. This work underscores the potential benefits of interventions targeting allostatic-interoception to mitigate behavioral impairments in neurodegenerative diseases, particularly in bvFTD.