1. Genetics and Genomics
  2. Neuroscience

Population clustering of structural brain aging and its association with brain development

  1. Haojing Duan
  2. Runye Shi
  3. Jujiao Kang
  4. Tobias Banaschewski
  5. Arun LW Bokde
  6. Christian Büchel
  7. Sylvane Desrivières
  8. Herta Flor
  9. Antoine Grigis
  10. Hugh Garavan
  11. Penny A Gowland
  12. Andreas Heinz
  13. Rüdiger Brühl
  14. Jean-Luc Martinot
  15. Marie-Laure Paillère Martinot
  16. Eric Artiges
  17. Frauke Nees
  18. Dimitri Papadopoulos Orfanos
  19. Luise Poustka
  20. Sarah Hohmann
  21. Nathalie Nathalie Holz
  22. Juliane Fröhner
  23. Michael N Smolka
  24. Nilakshi Vaidya
  25. Henrik Walter
  26. Robert Whelan
  27. Gunter Schumann
  28. Xiaolei Lin  Is a corresponding author
  29. Jianfeng Feng  Is a corresponding author
  1. Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, China
  2. Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence (Fudan University), Ministry of Education, China
  3. School of Data Science, Fudan University, China
  4. Department of Child and Adolescent Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, Germany
  5. Discipline of Psychiatry, School of Medicine and Trinity College Institute of Neuroscience, Trinity College Dublin, Ireland
  6. University Medical Centre Hamburg-Eppendorf, Germany
  7. Social Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, United Kingdom
  8. Institute of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, Germany
  9. Department of Psychology, School of Social Sciences, University of Mannheim, Germany
  10. NeuroSpin, CEA, Université Paris-Saclay, France
  11. Departments of Psychiatry and Psychology, University of Vermont, United States
  12. Sir Peter Mansfield Imaging Centre School of Physics and Astronomy, University of Nottingham, United Kingdom
  13. Department of Psychiatry and Psychotherapy CCM, Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Germany
  14. Physikalisch-Technische Bundesanstalt (PTB), Braunschweig and Berlin, Germany
  15. Institut National de la Santé et de la Recherche Médicale, INSERM U1299 "Developmental Trajectories and Psychiatry", Université Paris-Saclay, Ecole Normale supérieure Paris-Saclay, CNRS, Centre Borelli, France
  16. AP-HP. Sorbonne Université, Department of Child and Adolescent Psychiatry, Pitié-Salpêtrière Hospital, France
  17. Psychiatry Department, EPS Barthélémy Durand, France
  18. Institute of Medical Psychology and Medical Sociology, University Medical Center Schleswig-Holstein, Kiel University, Germany
  19. Department of Child and Adolescent Psychiatry and Psychotherapy, University Medical Centre, Germany
  20. Department of Psychiatry and Neuroimaging Center, Technische Universität Dresden, Germany
  21. Department of Psychiatry and Neurosciences, Charité–Universitätsmedizin Berlin, corporate member of Freie Universität BerlinHumboldt-Universität zu Berlin, and Berlin Institute of Health, Germany
  22. School of Psychology and Global Brain Health Institute, Trinity College Dublin, Ireland
  23. Centre for Population Neuroscience and Stratified Medicine (PONS Centre), ISTBI, Fudan University, China
  24. Centre for Population Neuroscience and Stratified Medicine (PONS), Department of Psychiatry and Psychotherapy, Charité Universitätsmedizin, Germany
  25. Huashan Institute of Medicine, Huashan Hospital affiliated to Fudan University, China
  26. MOE Frontiers Center for Brain Science, Fudan University, China
  27. Zhangjiang Fudan International Innovation Center, China
  28. Department of Computer Science, University of Warwick, United Kingdom
https://doi.org/10.7554/eLife.94970.3
Share this article
Cite this article
  1. Haojing Duan
  2. Runye Shi
  3. Jujiao Kang
  4. Tobias Banaschewski
  5. Arun LW Bokde
  6. Christian Büchel
  7. Sylvane Desrivières
  8. Herta Flor
  9. Antoine Grigis
  10. Hugh Garavan
  11. Penny A Gowland
  12. Andreas Heinz
  13. Rüdiger Brühl
  14. Jean-Luc Martinot
  15. Marie-Laure Paillère Martinot
  16. Eric Artiges
  17. Frauke Nees
  18. Dimitri Papadopoulos Orfanos
  19. Luise Poustka
  20. Sarah Hohmann
  21. Nathalie Nathalie Holz
  22. Juliane Fröhner
  23. Michael N Smolka
  24. Nilakshi Vaidya
  25. Henrik Walter
  26. Robert Whelan
  27. Gunter Schumann
  28. Xiaolei Lin
  29. Jianfeng Feng
(2024)
Population clustering of structural brain aging and its association with brain development
eLife 13:RP94970.
https://doi.org/10.7554/eLife.94970.3
  2. Download BibTeX
  3. Download .RIS
18 figures, 15 tables and 1 additional file

Figures

Figure 1
Download asset Open asset
Overview of the study workflow.

(a) Population cohorts (UK Biobank and IMAGEN) and data sources (brain imaging, biological aging biomarkers, cognitive functions, genomic data) involved in this study. (b) Brain aging patterns were identified using longitudinal trajectories of the whole brain GMV, which enabled the capturing of long-term and individualized variations compared to only use cross-sectional data, and associations between brain aging patterns and other measurements (biological aging, cognitive functions and PRS of major neuropsychiatric disorders) were investigated. (c) Mirroring patterns between brain aging and brain development was investigated using z-transformed brain volumetric change map and gene expression analysis.

Figure 2
Download asset Open asset
Global (a) and selected regional (b, c) cortical gray matter volume rate of change among participants with brain aging patterns 1 (red) and 2 (blue).

Rates of volumetric change for total gray matter and each ROI were estimated using GAMM, which incorporates both cross-sectional between-subject variation and longitudinal within-subject variation from 40,921 observations and 37,013 participants. Covariates include sex, assessment center, handedness, ethnic, and ICV. Shaded areas around the fit line denotes 95% CI.

Figure 2—source data 1

Related to Figure 2.

Global (a) and selected regional (b, c) cortical gray matter volume rate of change among participants with brain aging patterns 1 (red) and 2 (blue).

https://cdn.elifesciences.org/articles/94970/elife-94970-fig2-data1-v1.xlsx
Figure 3
Download asset Open asset
Distributions of biological aging biomarkers (leucocyte telomere length (LTL) and PhenoAge) among participants with brain aging patterns 1 and 2.

Boxes represent the interquartile range (IQR), lines within the boxes indicate the median. Two-sided p values were obtained by comparing LTL or PhenoAge Levine et al., 2018 between brain aging patterns using unadjusted multivariate linear regression models. Results remained significant when adjusting for sex, age, ethnic, BMI, smoking status and alcohol intake frequency in the LTL model Demanelis et al., 2020 and sex, age, ethnic, BMI, smoking status, alcohol frequency and education years in PhenoAge model. Stars indicate statistical significance after Bonferroni correction. ****: p ≤ 0.0001, *: p ≤ 0.05.

Figure 3—source data 1

Related to Figure 3.

Distributions of biological aging biomarkers (leucocyte telomere length (LTL) and PhenoAge) among participants with brain aging patterns 1 and 2.

https://cdn.elifesciences.org/articles/94970/elife-94970-fig3-data1-v1.xlsx
Figure 4
Download asset Open asset
Effect size (Cohen’s D or odds ratio) for comparing the cognitive functions between participants with brain aging patterns 1 and 2.

Results were adjusted such that negative Cohen’s D and Odds Ratio less than 1 indicate worse cognitive performances in brain aging pattern 2 compared to pattern 1. Width of the lines extending from the center point represent 95% confidence interval. Two-sided p values were obtained using both unadjusted and adjusted (for sex, age, and TDI, education and income) multivariate regression models. Stars indicate statistical significance after FDR correction for 11 comparisons. ****: p ≤ 0.0001, ***: p ≤ 0.001, **: p ≤ 0.01, ns: p>0.05.

Figure 4—source data 1

Related to Figure 4.

Effect size (Cohen’s D or odds ratio) for comparing the cognitive functions between participants with brain aging patterns 1 and 2.

https://cdn.elifesciences.org/articles/94970/elife-94970-fig4-data1-v1.xlsx
Figure 5
Download asset Open asset
Participants with accelerated brain aging (brain aging pattern 2) had significantly increased genetic liability to ADHD and delayed brain development.

Polygenic risk score (PRS) for ADHD, ASD, AD, PD, BIP, MDD, SCZ and delayed brain development (unpublished GWAS) were calculated at different p-value thresholds from 0.005 to 0.5 at an interval of 0.005. Vertical axis represents negative logarithm of P values comparing PRS in brain aging pattern 2 relative to pattern 1. Red horizontal dashed line indicates FDR corrected p value of 0.05. Colors represent traits and dots within the same color represent different p value thresholds. The trigonometric symbol indicates the average PRS across all p-value thresholds for the same trait.

Figure 5—source data 1

Related to Figure 5.

Participants with accelerated brain aging (brain aging pattern 2) had significantly increased genetic liability to ADHD and delayed brain development.

https://cdn.elifesciences.org/articles/94970/elife-94970-fig5-data1-v1.xlsx
Figure 6
Download asset Open asset
Genome-wide association study (GWAS) identified 6 independent SNPs associated with accelerated brain aging.

Total GMV at 60 years old was estimated for each participant using mixed effect models allowing for individualized baseline GMV and GMV change rate, and was used as the phenotype in the GWAS. (a) At genome-wide significance level (p=5×10–8, red dashed line), rs10835187 and rs7776725 loci were identified to be associated with accelerated brain aging. (b) Quantile–quantile plot showed that the most significant p values deviate from the null, suggesting that results are not unduly inflated.

Figure 7
Download asset Open asset
The ‘last in, first out’ mirroring patterns between brain development and brain aging.

(a) The annual percentage volume change (APC) was calculated for each ROI and standardized across the whole brain in adolescents (IMAGEN, left) and mid-to-late aged adults (UK Biobank, right), respectively. For adolescents, ROIs of in red indicate delayed structural brain development, while for mid-to-late aged adults, ROIs in blue indicate accelerated structural brain aging. (b) Estimated APC in brain development versus early aging (55 years old, left), and versus late aging (75 years old, right). ROIs in red indicate faster GMV decrease during brain aging and slower GMV decrease during brain development, that is stronger mirroring effects between brain development and brain aging. (c) Mirroring patterns between brain development and brain aging were more manifested in participants with accelerated aging (brain aging pattern 2). The arrows point to ROIs with more pronounced mirroring patterns in each subfigure.

Figure 7—source data 1

Related to Figure 7.

The ‘last in, first out’ mirroring patterns between brain development and brain aging.

https://cdn.elifesciences.org/articles/94970/elife-94970-fig7-data1-v1.xlsx
Figure 8
Download asset Open asset
Functional enrichment of gene transcripts significantly associated with delayed brain development and accelerated brain aging.

(a) 990 genes were spatially correlated with the first PLS component of delayed structural brain development, and were enriched for trans-synaptic signal regulation, forebrain development, signal release and cAMP signaling pathway. (b) 2293 genes were spatially correlated the first PLS component of accelerated structural brain aging, and were enriched for macroautophagy, pathways of neurodegeneration, establishment of protein localization to organelle and histone modification. Size of the circle represents number of genes in each term and P values were corrected using FDR for multiple comparisons.

Figure 8—source data 1

Related to Figure 8.

Functional enrichment of gene transcripts significantly associated with delayed brain development and accelerated brain aging. Each gene’s contribution to the first PLS component for IMAGEN and UKB.

https://cdn.elifesciences.org/articles/94970/elife-94970-fig8-data1-v1.xlsx
Figure 8—source data 2

Related to Figure 8.

Functional enrichment of gene transcripts significantly associated with delayed brain development and accelerated brain aging. Gene set enrichment for genes spatially positively correlated with the first PLS component in IMAGEN and negatively correlated with the first PLS component in UKB.

https://cdn.elifesciences.org/articles/94970/elife-94970-fig8-data2-v1.xlsx
Appendix 1—figure 1
Download asset Open asset
The sample selection workflow.
Appendix 1—figure 2
Download asset Open asset
Estimated rates of change in regional volumes for 33 bilateral brain regions.
Appendix 1—figure 3
Download asset Open asset
Stratification of the identified brain aging patterns using linear and non-linear dimensionality reduction methods.
Appendix 1—figure 4
Download asset Open asset
Effect size for comparing each individual blood biochemical metric (used to calculate the PhenoAge) between participants with brain aging patterns 1 and 2.
Appendix 1—figure 5
Download asset Open asset
Gene set enrichment of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and gene ontology (GO) of biological processes.
Appendix 1—figure 6
Download asset Open asset
Optimal number of clusters was chosen using elbow method (a) and silhouette method (b).
Author response image 1
Download asset Open asset
Stratification of the identified brain aging patterns using linear and non-linear dimensionality reduction methods.

(a) The principal component space of PC1 and PC2, and (b) two-dimensional projected locally linear embedding space derived from brain volumetric measures. Points have been colored and shaped according to grouping labels of the brain aging patterns.

Author response image 2
Download asset Open asset
Total gray matter volume (TGMV) (a) and the estimated rate of change (b) for females (red) and males (blue).

Rates of volumetric change for total gray matter and each ROI were estimated using GAMM, which incorporates both cross-sectional between-subject variation and longitudinal withinsubject variation from 22,067 observations for 19,958 females, and 18,854 observations for 17,055 males. Covariates include assessment center, handedness, ethnic, and ICV. Shaded areas around the fit line denotes 95% CI.

Author response image 3
Download asset Open asset
Distributions of biological aging biomarkers (leucocyte telomere length (LTL) and PhenoAge) among participants with brain aging patterns 1 and 2.
Author response image 4
Download asset Open asset
The “last in, first out” mirroring patterns between brain development and brain aging.

Tables

Appendix 1—table 1
ICD-10 primary and secondary diagnostic codes for exclusion criteria.
ConditionCode
Malignant neoplasmC70, C71
DementiaF00, F01, F02, F03, F04
Mental and behavioural disorders due to psychoactive substance useF10-F19
Schizophrenia, schizotypal and delusional disordersF20-F29
Mood [affective] disordersF30-F39
Mental retardationF70-F79
Disorders of psychological developmentF80-F89
Hyperkinetic disordersF90
Inflammatory diseases of the central nervous systemG00-G09
Systemic atrophies primarily affecting the central nervous systemG10, G11, G122, G13
Extrapyramidal and movement disordersG20, G21, G22, G23
Other degenerative diseases of the nervous systemG30-G32
Demyelinating diseases of the central nervous systemG35-G37
Episodic and paroxysmal disordersG40, G41, G45, G46
Infantile cerebral palsyG80
Cerebrovascular diseasesI60-I69
Down’s syndromeQ90
Intracranial injuryS06
Appendix 1—table 2
Self-reported illness codes for exclusion criteria.
Field IDConditionCode
20001Brain cancer/primary malignant brain tumour1032
Meningeal cancer/malignant meningioma1031
20002Benign neuroma1683
Brain abscess/intracranial abscess1245
Brain haemorrhage1491
Cerebral aneurysm1425
Cerebral palsy1433
Chronic/degenerative neurological problem1258
dementia/alzheimers/cognitive impairment1263
Encephalitis1246
Epilepsy1264
Fracture skull/head1626
Head injury1266
Ischaemic stroke1583
Meningioma/benign meningeal tumour1659
Meningitis1247
Motor Neurone Disease1259
Multiple Sclerosis1261
Nervous system infection1244
Neurological injury/trauma1240
Other demyelinating disease (not Multiple Sclerosis)1397
Other neurological problem1434
Parkinson’s Disease1262
Spina Bifida1524
Stroke1081
Subarachnoid haemorrhage1086
Subdural haemorrhage/haematoma1083
Transient ischaemic attack1082
Appendix 1—table 3
Cortical and subcortical brain regions.
Desikan–Killiany Atlas
bankssts
caudal anterior cingulate
caudal middle frontal
cuneus
entorhinal
fusiform
inferior parietal
inferior temporal
isthmus cingulate
lateral occipital
lateral orbitofrontal
lingual
medial orbitofrontal
middle temporal
parahippocampal
paracentral
pars opercularis
pars orbitalis
pars triangularis
pericalcarine
postcentral
posterior cingulate
precentral
precuneus
rostral anterior cingulate
rostral middle frontal
superior frontal
superior parietal
superior temporal
supramarginal
frontal pole
transverse temporal
insula
ASEG Atlas
thalamus proper
caudate
putamen
pallidum
hippocampus
amygdala
accumbens area
Appendix 1—table 4
Loadings matrix for the first 15 principal components.
PC1PC2PC3PC4PC5PC6PC7PC8PC9PC10PC11PC12PC13PC14PC15
Cortical
bankssts0.12–0.250.010.30–0.130.010.090.340.040.05–0.100.16–0.09–0.200.02
caudal.anterior.cingulate0.12–0.06–0.020.130.390.280.13–0.09–0.370.240.050.03–0.04–0.130.07
caudal.middle.frontal0.150.01–0.10–0.110.34–0.210.100.060.03–0.28–0.23–0.200.240.000.25
cuneus0.130.47–0.010.06–0.010.000.070.120.010.070.12–0.05–0.090.08–0.07
entorhinal0.090.070.130.080.100.17–0.53–0.180.060.04–0.19–0.100.130.01–0.04
fusiform0.18–0.030.010.13–0.020.02–0.20–0.020.150.30–0.01–0.28–0.230.080.21
inferior.parietal0.15–0.180.010.39–0.090.010.14–0.040.160.05–0.20–0.02–0.15–0.020.02
inferior.temporal0.16–0.130.010.26–0.010.15–0.04–0.100.14–0.08–0.08–0.250.200.29–0.01
isthmus.cingulate0.150.230.020.11–0.14–0.030.08–0.03–0.23–0.07–0.220.160.430.040.11
lateral.occipital0.160.300.000.12–0.020.020.050.100.160.180.06–0.23–0.230.120.17
lateral.orbitofrontal0.22–0.04–0.08–0.17–0.050.15–0.02–0.150.09–0.02–0.010.40–0.010.220.17
lingual0.120.430.030.14–0.01–0.020.070.11–0.01–0.03–0.100.160.02–0.09–0.08
medial.orbitofrontal0.21–0.06–0.06–0.06–0.050.10–0.01–0.130.18–0.060.130.170.050.01–0.25
middle.temporal0.17–0.17–0.010.27–0.140.100.210.250.09–0.15–0.090.040.070.05–0.11
parahippocampal0.120.090.130.050.060.14–0.460.02–0.020.09–0.220.090.03–0.320.02
paracentral0.19–0.05–0.10–0.070.18–0.23–0.11–0.060.010.03–0.020.06–0.35–0.01–0.35
pars.opercularis0.16–0.02–0.06–0.23–0.190.100.140.01–0.16–0.13–0.43–0.29–0.16–0.060.00
pars.orbitalis0.180.02–0.05–0.17–0.110.240.03–0.170.230.080.010.41–0.140.050.24
pars.triangularis0.160.02–0.03–0.35–0.260.230.13–0.05–0.06–0.06–0.25–0.17–0.21–0.08–0.05
pericalcarine0.100.47–0.010.08–0.010.010.090.13–0.02–0.010.020.08–0.030.03–0.17
postcentral0.20–0.04–0.110.060.16–0.29–0.090.120.080.03–0.090.100.040.010.04
posterior.cingulate0.18–0.08–0.040.020.240.090.08–0.02–0.300.01–0.070.10–0.05–0.12–0.31
precentral0.21–0.01–0.10–0.080.28–0.29–0.090.080.15–0.05–0.150.13–0.060.050.12
precuneus0.200.02–0.020.13–0.24–0.260.01–0.35–0.20–0.040.06–0.040.000.030.07
rostral.anterior.cingulate0.15–0.07–0.070.050.310.210.15–0.05–0.300.130.11–0.060.000.180.15
rostral.middle.frontal0.200.02–0.05–0.050.100.140.22–0.110.17–0.110.24–0.140.19–0.220.14
superior.frontal0.22–0.03–0.09–0.210.14–0.13–0.010.060.20–0.180.06–0.110.020.00–0.16
superior.parietal0.170.02–0.050.14–0.18–0.290.01–0.45–0.160.000.14–0.08–0.08–0.060.04
superior.temporal0.21–0.13–0.02–0.06–0.100.01–0.070.40–0.070.130.24–0.050.02–0.060.08
supramarginal0.17–0.15–0.080.06–0.24–0.23–0.130.03–0.210.040.110.100.14–0.08–0.06
frontal.pole0.160.03–0.100.010.000.160.03–0.130.280.050.28–0.160.26–0.38–0.22
transverse.temporal0.15–0.01–0.10–0.27–0.17–0.10–0.160.25–0.220.190.24–0.100.090.040.16
insula0.19–0.080.01–0.22–0.120.10–0.070.11–0.060.15–0.080.000.190.16–0.23
Subcortical
thalamus.proper0.09–0.020.31–0.080.03–0.080.080.000.02–0.210.110.10–0.19–0.370.33
caudate0.06–0.030.32–0.120.02–0.100.15–0.020.110.31–0.110.090.270.140.12
putamen0.08–0.030.42–0.120.03–0.130.13–0.030.050.22–0.05–0.140.07–0.02–0.17
pallidum0.03–0.030.42–0.070.00–0.130.16–0.060.040.20–0.070.000.04–0.20–0.06
hippocampus0.120.010.330.04–0.020.11–0.200.08–0.16–0.390.090.03–0.11–0.030.05
amygdala0.13–0.030.310.05–0.010.12–0.130.09–0.07–0.350.25–0.09–0.020.24–0.01
accumbens.area0.11–0.050.31–0.020.12–0.070.11–0.040.00–0.030.100.09–0.140.34–0.18
Appendix 1—table 5
Baseline and demographic characteristics for participants in the total population and stratified by brain aging patterns.
Total (n=37,013)Pattern 1 (n=18,929)Pattern 2 (n=18,084)
Age (years), mean (SD)63.9 (7.63)63.9 (7.64)63.8 (7.63)
Female, n (%)19,958 (53.9)10,117 (53.4)9,841 (54.4)
Ethnicity, n (%)   
 White34,219 (92.5)17,509 (92.5)16,710 (92.4)
 Mixed1,137 (3.1)573 (3.0)564 (3.1)
 Asian or Asian British1,210 (3.3)612 (3.2)598 (3.3)
 Other447 (1.2)235 (1.2)212 (1.2)
Smoking status, n (%)*   
 Never smoker23,633 (64.4)12,269 (65.4)11,364 (63.4)
 Previous smoker12,213 (33.3)6,085 (32.4)6,128 (34.2)
 Current smoker833 (2.3)414 (2.2)419 (2.3)
TDI, mean (SD)†–1.94 (2.69)–1.97 (2.66)–1.90 (2.71)
BMI (kg/m2), mean (SD)‡26.4 (4.32)26.3 (4.17)26.5 (4.46)
Years of Schooling, mean (SD)§16.8 (4.32)16.9 (4.29)16.8 (4,35)

  1. TDI = Townsend Deprivation Index, BMI = Body Mass Index.

  2. *

    Missing 334

  3. †

    Missing 36

  4. ‡

    Missing 1937

  5. §

    Missing 337

Appendix 1—table 6
Associations between results of biological aging biomarkers and subgroups stratified by whole-brain TGMV trajectories.

Cohen’s d measures the standardized difference of means between brain aging pattern 2 and brain aging pattern 1.

Biological aging biomarkersBrain aging pattern 1Brain aging pattern 2
NMean(SD)NMean(SD)
LTL17,6910.083 (0.98)16,8760.055 (0.97)
PhenoAge15,22841.35 (8.17)14,32341.58 (8.32)
Biological aging biomarkersUnadjustedAdjusted
Cohen’s d (95% CI)pP.BonferroniCohen’s d (95% CI)pP.Bonferroni
LTL–0.028 (-0.049,–0.007)0.0090–0.030 (-0.051,–0.009)0.0060.011
PhenoAge0.027 (0.004, 0.050)0.01900.092 (0.070, 0.116)3.05E-156.11E-15
Appendix 1—table 7
Associations between results of cognitive function tests and subgroups stratified by whole-brain TGMV trajectories.

Cohen’s d measures the standardized difference of means between brain aging pattern 2 and brain aging pattern 1.

Cognitive functionsBrain aging pattern 1Brain aging pattern 2
NMean(SD)NMean(SD)
Reaction time17,749–594.70 (108.15)16,831–594.68 (110.50)
Numeric memory13,3506.82 (1.26)12,3466.72 (1.27)
Fluid intelligence17,5806.73 (2.06)16,6126.53 (2.04)
Trail making A13,052–224.50 (84.82)12,036–227.56 (84.85)
Trail making B12,743–563.06 (246.74)11,753–576.55 (260.75)
Matrix pattern completion13,0648.07 (2.11)12,0647.91 (2.14)
Symbol digit substitution13,07719.08 (5.15)12,05618.87 (5.35)
Tower rearranging12,9529.96 (3.20)11,9589.83 (3.23)
Paired associate learning13,1847.01 (2.59)12,1986.88 (2.65)
Prospective memory17,831N/A16,949N/A
Pairs matching17,840–3.58 (2.90)16,956–3.67 (2.94)
Cognitive functionsUnadjustedAdjusted
Cohen’s d (95% CI)PP.FDRCohen’s d (95% CI)PP.FDR
Reaction time0.000 (-0.021, 0.021)0.990.990.006 (-0.016, 0.028)0.610.61
Numeric memory–0.082 (-0.106,–0.057)5.97E-113.28E-10–0.080 (-0.106,–0.055)8.99E-104.95E-09
Fluid intelligence–0.102 (-0.123,–0.080)5.94E-216.54E-20–0.99 (-0.121,–0.077)3.30E-183.63E-17
Trail making A–0.036 (-0.061,–0.011)0.0040.006–0.050 (-0.074,–0.024)1.46E-042.29E-04
Trail making B–0.053 (-0.078,–0.028)3.16E-058.68E-05–0.067 (-0.093,–0.041)6.16E-071.69E-06
Matrix pattern completion–0.076 (-0.101,–0.051)1.84E-096.74E-09–0.078 (-0.104,–0.052)3.67E-091.35E-08
Symbol digit substitution–0.040 (-0.065,–0.015)0.0020.002–0.053 (-0.079,–0.027)6.15E-051.13E-04
Tower rearranging–0.041 (-0.066,–0.016)0.0010.002–0.049 (-0.075,–0.023)2.18E-043.00E-04
Paired associate learning–0.051 (-0.076,–0.027)4.64E-051.02E-04–0.054 (-0.079,–0.028)4.89E-051.08E-04
Prospective memoryOR: 0.943 (0.891, 0.999)0.0470.052OR: 0.940 (0.883, 1.000)0.0520.057
Pairs matching–0.029 (-0.050,–0.008)0.0060.008–0.033 (-0.055,–0.011)0.0030.004
Appendix 1—table 8
Genome-wide association study details.

Loci associated with risk were thresholded at p<5×10–8, then distance-based clumping was used to define independently significant loci.

StudyNumber casesNumber controlsNumber of genome-wide independently significant lociDownload link
Attention deficit hyperactivity disorder Demontis et al., 201920,18335,19112https://figshare.com/ndownloader/files/28169253
Autism spectrum disorder Grove et al., 201918,38127,9695https://figshare.com/ndownloader/files/28169292
Alzheimer’s disease Jansen et al., 201971,880383,37825https://ctg.cncr.nl/software/summary_statistics
Parkinson’s disease Nalls et al., 201937,688, and 18,618 (proxy-cases)1,417,79190https://drive.google.com/file/d/1FZ9UL99LAqyWnyNBxxlx6qOUlfAnublN/view?usp=sharing
Bipolar disorder Mullins et al., 202141,917371,54964https://figshare.com/ndownloader/files/40036705
Major depressive disorder Wray et al., 2018135,458344,90144https://figshare.com/ndownloader/files/39504667
Schizophrenia Trubetskoy et al., 202276,755243,649287https://figshare.com/ndownloader/files/34517828
Delayed Brain Development Shi et al., 20237662 (proxy phenotype, continuous)1https://delayedneurodevelopment.page.link/amTC
Appendix 1—table 9
Polygenic Risk Scores comparisons between two subgroups.

Data supporting these scores were obtained either entirely from external GWAS data (the Standard PRS set). The bold P values reflect significance after FDR correction.

Traitn1n2statisticpp.adjust
AAM18,42917,586–2.2180.0270.080
AMD18,42917,5861.7530.0800.169
AD18,42917,5860.7350.4620.616
AST18,42917,586–0.8610.3890.543
AF18,42917,586–0.1000.9200.945
BD18,42917,5863.5573.75E-040.002
BMI18,42917,586–3.3090.0010.005
CRC18,42917,586–0.5440.5860.703
BC18,42917,586–3.1400.0020.008
CVD18,42917,586–2.1040.0350.091
CED18,42917,5861.0460.2960.484
CAD18,42917,586–1.5880.1120.202
CD18,42917,586–0.0940.9250.945
EOC18,42917,586–2.1830.0290.080
EBMDT18,42917,586–11.343<1.00E-20<1.00E-20
HBA1C_DF18,42917,586–2.9480.0030.013
HEIGHT18,42917,5866.6582.81E-113.37E-10
HDL18,42917,5860.8840.3770.543
HT18,42917,586–3.5394.02E-040.002
IOP18,42917,586–1.6050.1090.202
ISS18,42917,586–2.3830.0170.056
LDL_SF18,42917,586–0.6860.4920.627
MEL18,42917,5862.0250.0430.103
MS18,42917,586–0.0690.9450.945
OP18,42917,58612.029<1.00E-20<1.00E-20
PD18,42917,5861.4560.1450.249
POAG18,42917,586–0.8560.3920.543
PC18,42917,586–0.2400.8100.941
PSO18,42917,586–1.7810.0750.169
RA18,42917,586–2.4370.0150.053
SCZ18,42917,5860.1580.8740.945
SLE18,42917,5861.6950.0900.180
T1D18,42917,5860.6660.5050.627
T2D18,42917,586–5.5233.35E-083.02E-07
UC18,42917,5860.8830.3770.543
VTE18,42917,586–0.1700.8650.945
Appendix 1—table 10
Polygenic Risk Scores comparisons between two subgroups.

Data supporting these scores were obtained external and internal UK Biobank data (the Enhanced PRS set). The bold p values reflect significance after FDR correction.

Traitn1n2statisticpp.adjust
AAM3,4073,409–1.7080.0880.344
AMD3,4073,4090.5470.5840.931
AD3,4073,4090.7560.4500.820
APOEA3,4073,4090.0230.9820.993
APOEB3,4073,4090.1190.9050.968
AST3,4073,4090.1120.9110.968
AF3,4073,4091.3060.1920.600
BD3,4073,4090.5610.5750.931
BMI3,4073,409–0.9760.3290.730
CRC3,4073,4090.9840.3250.730
BC3,4073,409–0.9950.3200.730
CAL3,4073,409–1.7860.0740.326
CVD3,4073,409–0.0090.9930.993
CED3,4073,4091.2800.2000.600
CAD3,4073,4090.2310.8180.961
DOA3,4073,409–0.3260.7450.961
EOC3,4073,409–2.1670.0300.155
EBMDT3,4073,409–6.1111.04E-092.65E-08
EGCR3,4073,4090.4130.6800.961
EGCY3,4073,409–0.2100.8340.961
HBA1C_DF3,4073,4090.1300.8960.968
HEIGHT3,4073,4094.3511.38E-052.35E-04
HDL3,4073,4090.2940.7690.961
HT3,4073,409–0.8840.3770.743
IOP3,4073,409–2.3660.0180.151
ISS3,4073,4090.0660.9470.986
LDL_SF3,4073,409–0.1930.8470.961
MEL3,4073,4092.6590.0080.080
MS3,4073,409–2.2930.0220.151
OTFA3,4073,4090.3180.7500.961
OSFA3,4073,4090.7700.4410.820
OP3,4073,4096.4849.54E-114.87E-09
PD3,4073,4091.0410.2980.730
PDCL3,4073,4090.6630.5070.862
PHG3,4073,4090.3920.6950.961
PFA3,4073,4090.4950.6210.932
POAG3,4073,409–1.0830.2790.730
PC3,4073,4090.6750.5000.862
PSO3,4073,409–2.2340.0260.151
RMNC3,4073,4090.5010.6170.932
RHR3,4073,409–2.8650.0040.053
RA3,4073,4090.2970.7660.961
SCZ3,4073,409–0.8800.3790.743
SGM3,4073,409–0.2240.8230.961
SLE3,4073,4091.4580.1450.493
TCH3,4073,4090.1910.8480.961
TFA3,4073,4090.8920.3720.743
TTG3,4073,4091.2120.2260.640
T1D3,4073,4091.7710.0770.326
T2D3,4073,409–2.2180.0270.151
VTE3,4073,409–1.6130.1070.390
Appendix 1—table 11
Most significant single-variant associations (p < 510-8) detected in the GWAS analyses.

Six independent SNPs at genome-wide significance level were identified by linkage disequilibrium (LD) clumping (r2 < 0.1 within a 250 kb window). The location (chromosome [chr] and base position [bp]), alleles (A1 = effect allele and A2 = other allele), effect (β) and its standard error (β SE) with respect to A1, and association p-values from regression model of the variants are given, along with functional consequences of SNPs on gene by performing ANNOVAR.

SNPA1A2p-valueββ SELocation (chr:bp)Gene symbolPosition relative to gene
rs10835187CT1.70e-14–0.025580.00333311:27505677LGR4, LIN7Cintergenic
rs7776725CT4.47e-13–0.026400.0036447:121033121FAM3Cintronic
rs779233904AACA1.57e-090.020830.0034496:151910404CCDC170intronic
rs2504071TC3.34e-090.019590.0033116:152084862ESR1intronic
17:43553496:A:AATAAAT6.65e-09–0.024720.00426117:43553496PLEKHM1intronic
10:104227791:G:GAGAG1.48e-08–0.018890.00333410:104227791TMEM180intronic
Appendix 1—table 12
Association between gene expression profiles of mapped genes and estimated APC during brain development.

The bold p values reflect significance after the spatial permutation test.

Gene SymbolSpearman's ρp valueP.permutation
ACTR1A0.0960.4400.328
ARHGAP270.0510.6840.445
ARL17B0.0470.7050.452
BDNF-AS0.2560.0380.038
CCDC1700.2680.0300.069
ESR10.0210.8700.483
FAM3C–0.0960.4440.356
KANSL1–0.2620.0340.073
KANSL1-AS10.0670.5940.313
LGR40.5581.78E-062.50E-04
LIN7C0.0360.7750.464
LRRC37A4P–0.2720.0270.148
MAPT0.0240.8460.405
PLEKHM1–0.2760.0250.109
SPPL2C0.1160.3510.189
STH–0.1470.2380.277
SUFU0.4070.0010.028
Appendix 1—table 13
Association between gene expression profiles of mapped genes and estimated APC during brain aging.

The bold p values reflect significance after the spatial permutation test.

Gene SymbolSpearman's ρPvalueP.permutation
ACTR1A–0.2350.0580.052
ARHGAP270.4864.48E-055.50E-04
ARL17B0.0900.4730.240
BDNF-AS0.4903.68E-051.50E-04
CCDC1700.2060.0980.075
ESR10.5326.02E-061.50E-04
FAM3C–0.3660.0030.005
KANSL10.2130.0860.078
KANSL1-AS1–0.2620.0340.059
LGR40.0700.5760.348
LIN7C0.1770.1540.120
LRRC37A4P0.1430.2500.165
MAPT–0.2870.0200.022
PLEKHM10.2020.1040.080
SPPL2C0.2110.0890.059
STH–0.0010.9970.490
SUFU–0.0360.7730.373
Appendix 1—table 14
Model evaluation results using relative measures: AIC, BIC, likelihood ratio test and intra-class correlation (ICC).
modelAICBIClrtestICC_adjustedICC_unadjusted
lmer_intr_thalamus.proper79444.2479591.296.45E-150.88750.3958
lmer_slope_thalamus.proper79382.8979547.246.45E-150.88890.3986
lmer_intr_caudate95845.4895992.542.68E-710.95430.6824
lmer_slope_caudate95524.4995688.842.68E-710.95640.6817
lmer_intr_putamen92106.0692253.122.40E-420.93980.5946
lmer_slope_putamen91918.492082.752.40E-420.94240.5933
lmer_intr_pallidum90500.390647.364.21E-420.88590.5116
lmer_slope_pallidum90313.7690478.124.21E-420.88620.5093
lmer_intr_hippocampus89833.9289980.971.37E-080.93090.5505
lmer_slope_hippocampus89801.7189966.061.37E-080.93290.5505
lmer_intr_amygdala89976.9890124.034.01E-060.86970.4898
lmer_slope_amygdala89956.1390120.484.01E-060.87130.4897
lmer_intr_accumbens.area96936.1997083.241.69E-140.82280.5295
lmer_slope_accumbens.area96876.7797041.121.69E-140.82330.5310
lmer_intr_bankssts97981.9698129.010.0018180.93390.6798
lmer_slope_bankssts97973.3498137.690.0018180.93540.6814
lmer_intr_caudal.anterior.cingulate109133.2109280.30.1315070.86380.7653
lmer_slope_caudal.anterior.cingulate109133.2109297.50.1315070.86440.7659
lmer_intr_caudal.middle.frontal93118.1393265.198.31E-060.92270.5929
lmer_slope_caudal.middle.frontal93098.7493263.098.31E-060.92460.5949
lmer_intr_cuneus101801.4101948.50.0144290.92420.7256
lmer_slope_cuneus101796.9101961.30.0144290.92450.7262
lmer_intr_entorhinal109404.9109551.91.63E-140.80130.6908
lmer_slope_entorhinal109345.4109509.71.63E-140.80370.6928
lmer_intr_fusiform87545.8887692.939.17E-050.91390.5062
lmer_slope_fusiform87531.2887695.639.17E-050.91630.5074
lmer_intr_inferior.parietal89044.4389191.485.89E-140.93740.5564
lmer_slope_inferior.parietal88987.5189151.865.89E-140.94190.5603
lmer_intr_inferior.temporal84066.4784213.525.73E-070.93840.4956
lmer_slope_inferior.temporal84041.7384206.085.73E-070.94050.4977
lmer_intr_isthmus.cingulate92442.1292589.170.1271910.92750.5862
lmer_slope_isthmus.cingulate9244292606.350.1271910.92840.5869
lmer_intr_lateral.occipital89550.489697.450.0039430.91210.5273
lmer_slope_lateral.occipital89543.3389707.680.0039430.91290.5282
lmer_intr_lateral.orbitofrontal86224.1386371.181.29E-080.84660.4323
lmer_slope_lateral.orbitofrontal86191.7986356.141.29E-080.84970.4345
lmer_intr_lingual102605.2102752.20.0412420.91810.7268
lmer_slope_lingual102602.8102767.20.0412420.91810.7272
lmer_intr_medial.orbitofrontal89954.3690101.412.25E-060.78320.4219
lmer_slope_medial.orbitofrontal89932.3590096.72.25E-060.78720.4241
lmer_intr_middle.temporal83331.0183478.060.00190.91770.4615
lmer_slope_middle.temporal83322.4883486.830.00190.91950.4629
lmer_intr_parahippocampal108997.1109144.24.99E-050.86860.7639
lmer_slope_parahippocampal108981.3109145.64.99E-050.86900.7639
lmer_intr_paracentral98058.6398205.680.0156860.86950.5958
lmer_slope_paracentral98054.3298218.670.0156860.87050.5960
lmer_intr_pars.opercularis96829.0596976.11.20E-120.93540.6658
lmer_slope_pars.opercularis96778.1596942.51.20E-120.93960.6698
lmer_intr_pars.orbitalis96989.6197136.671.39E-050.87850.5892
lmer_slope_pars.orbitalis96971.2597135.61.39E-050.88050.5912
lmer_intr_pars.triangularis96637.5896784.634.55E-180.94020.6710
lmer_slope_pars.triangularis96561.7296726.074.55E-180.94390.6748
lmer_intr_pericalcarine105115.91052630.0228410.94290.8190
lmer_slope_pericalcarine105112.4105276.70.0228410.94410.8200
lmer_intr_postcentral91605.691752.650.0005490.87640.5189
lmer_slope_postcentral91594.5991758.940.0005490.87890.5203
lmer_intr_posterior.cingulate96853.7397000.782.21E-190.89130.6014
lmer_slope_posterior.cingulate96771.8296936.172.21E-190.89490.6022
lmer_intr_precentral91186.5991333.642.61E-120.84780.4864
lmer_slope_precentral91137.2591301.62.61E-120.85040.4864
lmer_intr_precuneus84734.5884881.631.23E-080.90880.4708
lmer_slope_precuneus84702.1684866.511.23E-080.91260.4730
lmer_intr_rostral.anterior.cingulate95688.6895835.732.64E-120.90930.6097
lmer_slope_rostral.anterior.cingulate95639.3695803.722.64E-120.91220.6129
lmer_intr_rostral.middle.frontal80873.8481020.892.57E-170.91370.4354
lmer_slope_rostral.middle.frontal80801.4480965.792.57E-170.91910.4395
lmer_intr_superior.frontal79730.679877.651.25E-110.89210.4038
lmer_slope_superior.frontal79684.3879848.731.25E-110.89720.4060
lmer_intr_superior.parietal92895.95930438.55E-070.88990.5487
lmer_slope_superior.parietal92872.0193036.368.55E-070.89340.5512
lmer_intr_superior.temporal86495.2486642.290.0030210.91480.4959
lmer_slope_superior.temporal86487.6486651.990.0030210.91680.4971
lmer_intr_supramarginal87390.3987537.443.67E-130.92630.5221
lmer_slope_supramarginal87337.1287501.473.67E-130.93200.5258
lmer_intr_frontal.pole110426.3110573.41.95E-650.69070.5868
lmer_slope_frontal.pole110132.3110296.71.95E-650.69570.5882
lmer_intr_transverse.temporal102753.2102900.20.0324130.91300.7247
lmer_slope_transverse.temporal102750.3102914.70.0324130.91440.7258
lmer_intr_insula88693.288840.267.61E-140.82630.4416
lmer_slope_insula88636.7988801.147.61E-140.82900.4420
Author response table 1
Contributions and effect directions of the first PLS components in brain development and brain aging of genes that mapped to GWAS significant SNP.

The bold P values reflect significance (P < 0.005, inclusion in the functional enrichment analysis) after FDR correction.

IMAGEN
Genebootstrap.weightspvaluep.adjust
LGR43.702.17E-042.00E-03
CCDC1701.591.11E-012.41E-01
SUFU1.351.77E-013.31E-01
SPPL2C0.973.32E-015.04E-01
ACTRIA0.943.49E-015.21E-01
MAPT0.893.75E-015.46E-01
KANSL1-AS10.764.50E-016.12E-01
ARHGAP270.675.03E-016.58E-01
LIN7C0.526.03E-017.39E-01
ARL17B0.456.53E-017.76E-01
ESRI0.168.75E-019.29E-01
BDNF-AS0.009.99E-019.99E-01
FAM3C-0.109.20E-019.57E-01
STH-1.272.06E-013.68E-01
PLEKHM1-2.391.70E-026.13E-02
KANSLI-2.401.65E-025.99E-02
LRRC37A4P-2.697.09E-033.13E-02
UKBiobank
Genebootstrap.weightspvaluep.adjust
BDNF-AS4.732.25E-065.11E-05
ARHGAP274.025.92E-054.38E-04
ESRI3.919.31E-056.13E-04
SPPL2C3.309.56E-043.69E-03
SUFU2.202.79E-025.72E-02
LIN7C2.153.14E-026.33E-02
CCDC1702.083.78E-027.34E-02
LGR41.767.88E-021.35E-01
ARL17B1.481.39E-012.13E-01
PLEKHM10.694.93E-015.89E-01
LRRC37A4P0.277.84E-018.39E-01
KANSL1-AS10.178.63E-019.00E-01
KANSLI-0.565.75E-016.64E-01
STH-1.033.04E-014.01E-01
MAPT-1.381.67E-012.47E-01
ACTRIA-2.133.35E-026.67E-02
FAM3C-3.682.38E-041.23E-03

Additional files

MDAR checklist
https://cdn.elifesciences.org/articles/94970/elife-94970-mdarchecklist1-v1.pdf

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Haojing Duan
  2. Runye Shi
  3. Jujiao Kang
  4. Tobias Banaschewski
  5. Arun LW Bokde
  6. Christian Büchel
  7. Sylvane Desrivières
  8. Herta Flor
  9. Antoine Grigis
  10. Hugh Garavan
  11. Penny A Gowland
  12. Andreas Heinz
  13. Rüdiger Brühl
  14. Jean-Luc Martinot
  15. Marie-Laure Paillère Martinot
  16. Eric Artiges
  17. Frauke Nees
  18. Dimitri Papadopoulos Orfanos
  19. Luise Poustka
  20. Sarah Hohmann
  21. Nathalie Nathalie Holz
  22. Juliane Fröhner
  23. Michael N Smolka
  24. Nilakshi Vaidya
  25. Henrik Walter
  26. Robert Whelan
  27. Gunter Schumann
  28. Xiaolei Lin
  29. Jianfeng Feng
(2024)
Population clustering of structural brain aging and its association with brain development
eLife 13:RP94970.
https://doi.org/10.7554/eLife.94970.3