Cortico-genetic mapping links individual brain maturity in youths to cognitive and psychiatric traits

Abstract

Neurodevelopmental trajectories are shaped by interactions between coordinated biological 14processes and individual experiences throughout ontogeny, yet the specific genetic and 15 environmental impact on brain development is enigmatic. Here, we map the genetic 16 architectures of cognitive traits and psychiatric disorders onto the brain, show that such 17 canonical genetic maps are associated with individual normative patterns in youths, and 18 provide evidence that trauma exposure and parental education may alter this relationship. 19Psychiatric, cognitive and brain imaging traits are highly heritable and polygenic [1][2][3][4][5][6][7][8][9][10][11][12][13] . Individual 20 genetic architecture contributes to individual differences in neurodevelopmental trajectories and 21 subsequent scaffolding and maintenance of brain structure and function throughout ontogeny. 22However, the links between the genetic and neural configurations and how their interplay shapes 23 individual differences in cognitive function and mental health remain poorly understood. This 24 knowledge gap has nurtured a debate on the extent of environmental influence and genetic 25 constraints on brain development. Here, we provide a comprehensive neuroanatomical mapping 26 of the genetic architecture of various cognitive traits and psychiatric disorders using brain 27 imaging and genetic data in a large population based sample (UK Biobank 14 ), and link the 28 resulting canonical genetic maps to individual patterns of brain maturity in the Philadelphia 29 Neurodevelopmental Cohort 15 . 30 2 We accessed data from the UK Biobank 14 and used Freesurfer 16 for cortical reconstruction based 31 on T1-weighted magnetic resonance images obtained from 16,612 healthy individuals with 32 European ancestry aged 40 to 70 years (mean: 55.8 years, sd: 7.5 years, 52.1% females). We 33 computed surface maps for cortical thickness and area, registered to fsaverage4 space (2,562 34 vertices), smoothed using a kernel with full width at half maximum of 15 mm. Next, we 35 performed a genome-wide association study (GWAS) for every vertex using PLINK 17 , linking 36 single-nucleotide-polymorphism (SNP) data with a given vertex's thickness and area, 37 respectively. Each GWAS accounted for effects of age, age², sex, scanning site and the first four 38 genetic principal components 18 to account for population stratification. 39We first estimated SNP-heritability of cortical morphology using LD Score regression 19 for each 40 vertex (Suppl. Fig 1). The spatial correlation between thickness and area heritability maps was 41 moderate (r=0.28, pperm=0.0004; Suppl. Fig 1b), and surface area was significantly more 42 heritable than thickness (Suppl. Fig 1a). Both measures showed regional differences, with high 43 heritability of thickness in the postcentral gyrus and Heschl's gyrus (Suppl. Fig 1c), and of 44 surface area in the lingual gyrus and the temporal lobe (Suppl. Fig 1d). These results from 45 vertex-wise SNP-based analysis largely confirmed earlier reports from twin...

Pleiotropy of polygenic factors associated with focal and generalized epilepsy in the general population

“…Vertex-wise genetic correlation maps with brain morphology were built following the methods presented by Kaufmann et al (2018) [25]. Cortical reconstruction was performed in a UKB [12] subset of 16,612 predominantly healthy [26] individuals of European ancestry with T1-weighted MRI data, using Freesurfer [27].…”

“…Cortical reconstruction was performed in a UKB [12] subset of 16,612 predominantly healthy [26] individuals of European ancestry with T1-weighted MRI data, using Freesurfer [27]. Of note, the impact of confounding factors such as comorbid disorders or medication on cortical morphology was reduced by the MRI data acquisition from predominantly healthy participants in the UKB [25]. We computed surface maps for cortical thickness and cortical surface area, registered to fsaverage4 space (2,562 vertices), smoothed using a kernel with full width at half maximum of 15 mm.…”

“…GE and FE show similar patterns of genetic correlation at vertex-level with cortical thickness and cortical surface area. We used a recent cortico-genetic mapping approach [25] to explore whether GE and FE are genetically correlated at the level of brain morphology. We performed vertex-level genetic correlation analyses between epilepsy (GE and FE) and cortical surface area and thickness.…”