"Which Part of the Brain Ages First" is Imprinted in Our Genes? The Blueprint of Local Brain Age

1) The Weakness of "Brain Age" is that it Becomes an "Average"

The concept of "Brain Age" involves estimating the "apparent age" from an MRI and measuring the progression of aging by the gap between this and the actual age. While convenient, many traditional methods summarize the entire brain into a single number. However, in reality, the brain ages and deteriorates at different speeds and vulnerabilities in different areas, such as the frontal lobe, temporal lobe, and cingulate gyrus.The speed of aging and susceptibility to damage differ by location.

In response, the current study introduces the concept of **Local Brain Age (LBA)**. This approach views the brain as a "map" and reads "which parts are aging and by how much" by region. Springer Link

2) Research Framework: Creating "Local Brain Age" with AI and Exploring Genetic Factors with GWAS

The research team (Kim et al.) applied a deep learning model to the T1-weighted MRIs of 41,708 cognitively normal adults from the UK Biobank to estimate LBA by voxel/region. Furthermore, they created a **Local Brain Age Gap (LBAG)** by subtracting the actual age (CA) from the estimated brain age and conducted a GWAS (Genome-Wide Association Study) using this as a phenotype. Springer Link

The analysis associated LBA estimates with 148 brain regions and tested their association with 662,971 SNPs, ultimately identifying 1,212 SNPs significantly associated with LBAG in at least one region. Springer Link

The important point here is that "local brain aging" emerged not as a hit-or-miss of a single gene, but as a **polygenic** effect, where small effects accumulate. hit-or-miss of a single gene. Springer Link

3) "Which Parts Age Easily" Was Quite Biased

The paper notes that while mutations associated with LBAG are widely distributed across the cortex, the "hit count" varies by region. For example, in the left hemisphere, there are many associated SNPs in the paracentral lobule/sulcus, while in the right hemisphere, the paracentral lobule/sulcus also has many, but the right occipital pole has few. Springer Link

This "bias" will be significant when overlaid with a "disease map" in the future. Researchers suggest that LBA-associated mutations form spatial clusters in default mode, limbic, and motor networks, potentially paralleling regional vulnerabilities seen in Alzheimer's disease (AD) and frontotemporal dementia (FTD). Springer Link

4) Notable Genes: KCNK2 and NUAK1 Had Opposite Effects

Delving into individual SNPs, for example, **two SNPs near KCNK2 (rs864736, rs59084003)** are treated as independent signals, widely associated with regions such as the cingulate gyrus and parietal cortex, and it is described that the more reference alleles there are, the higher the LBAG (leaning towards "aging"). Springer Link

On the other hand, the study also presents an organization where "a group including NUAK1 acts protectively (negative LBAG)." Along with NUAK1, LPAR1, ROCK1, and others are in the same "group," showing a noticeable negative β (direction to lower aging gap) in the anterior to middle cingulate gyrus, insular cortex, and paracentral cortex. Springer Link

5) The Key Point: Grouping SNPs into "Three Routes"

The "SNS-friendly" point of this paper is that it doesn't just end with "we found 1,212," but rather reduces the dimensions of how each SNP works (the pattern of which regions it is associated with) and groups them into three clusters. Springer Link

The discussion section of the paper roughly organizes it as follows. Springer Link

• A: Morphogenetic / Metabolic Systems
  An "aging route" that easily affects the hubs of the default mode.

• B: Cytoskeletal / Signaling Systems
  Including NUAK1 and ROCK1, potentially involved in "resilience" in the limbic to paracentral regions.

• C: Immuno-Epigenetic Systems
  Histone-related and immune signal-related groups, likely to localize around the frontal insula to perisylvian area.

The "three-route hypothesis" suggests a perspective of reinterpreting brain aging not as "random atrophy" but as a trajectory partially predictable, prepared from the developmental period and progressing while interacting with metabolic and immune states. Springer Link

6) Can Local Brain Age Serve as an Intermediate Indicator Connecting "DNA→Brain→Disease"?

The research team emphasizes LBA not as a mere "anatomical description," but as a **"genetically informed intermediate phenotype"** that can be placed between DNA and disease. Springer Link

For example, the paper touches on the potential of polygenic scores based on LBA to aid in assessing vulnerability/resilience to AD long before symptoms appear. Springer Link

However, this is prone to misunderstanding. It's not about **"having this SNP means you'll get dementia."** Most GWAS findings are worlds visible only through the sum of small effects, and individual predictions need to be handled carefully.

7) The Limitations Are Clearly Stated: What's Needed Next is "Multi-Population × Longitudinal"

The paper clearly outlines its limitations. Since the main participants are primarily of European descent, it states that validation with multi-ancestral groups is necessary. Springer Link
Moreover, as the design is a cross-sectional study, it requires longitudinal data validation to determine how LBA changes over time and whether the identified SNPs are predictors of future changes or reflect stable morphological characteristics. Springer Link

SNS Reactions (Within the Range Confirmed + Interpretation)

① "The Spread Has Not Yet Begun" Can Be Said from Primary Information

According to Springer's metrics, this paper shows Accesses 13, Altmetric 0, Mentions 0, Citations 0 (last updated: 2025-12-13 UTC). In other words, at least within the range Altmetric captures, mentions in SNS and news have not been observed at this point. Springer Link

② The "Temperature" of the Research Community Can Be Seen Through Another Route

While not necessarily direct mentions of the paper, a LinkedIn post by author Andrei Irimia highlights "local brain age measures help to identify persons at significantly higher risk for Alzheimer’s disease" as a poster highlight at OHBM 2025. This reflects the researchers' interest in whether local brain age might aid in risk stratification. LinkedIn

③ Points Likely to "Grow" in General SNS in the Future (※ This is a Trend Analysis)

When this type of research reaches general SNS, the following three points tend to generate excitement.

• "Which Part of Your Brain Ages First?" (Self-Diagnostic Content)

• "Is It Determined by Genetics? Can It Be Changed by Lifestyle?" (Including Reactions Against Genetic Determinism)

• "Will It Be Used for Insurance or Employment?" (Concerns About Ethics, Inequality, and Data Usage)

This study, with its focus on "mapping" and "genetics," has the potential to quickly gain traction depending on how it is received. However, at present, on primary metrics, it is still in the "upcoming" stage. Springer Link

Summary

The novelty of this research lies in expanding "brain age" from a "single number" to a "local map" and organizing the genetic program behind this map into 1,212 SNPs + 3 biological routes. Springer Link

In the future, when reproducibility in multi-ancestral groups and longitudinal tracking are achieved, the key question will be whether local brain age becomes a tool for discussing "ultra-early disease risk" or remains a "topographical map of lifelong vulnerability." Spring