Exploring the influence of the DRD2 gene on mathematical ability: perspectives of gene association and gene-environment interaction

The study investigates whether variants in the dopamine D2 receptor gene (DRD2)—a candidate gene implicated in working memory and language—are associated with children's mathematical ability and whether parental education (PE), as a proxy for home environment and socioeconomic status, moderates these genetic effects. Grounded in the Generalist Gene hypothesis, the authors propose: H1: DRD2 variants linked to working memory and language also affect mathematical ability; H2: DRD2 and PE interactively influence children's mathematical ability; H3: The DRD2×PE interaction pattern conforms to either the diathesis–stress or differential-susceptibility model, assessed via simple slope analyses. The study aims to clarify genetic association at the SNP and haplotype levels, evaluate aggregate genetic effects via haplotypes on complex math phenotypes, and test gene–environment interaction patterns.

Prior work indicates mathematical ability is polygenic (GWAS findings) and shares genetic influences with reading and general cognitive abilities, consistent with the Generalist Gene hypothesis. DRD2, expressed across cortical regions, has been linked to cognitive processes and working memory, with specific SNPs (e.g., rs1076560, rs2283265) affecting receptor splicing and neural networks relevant to executive function. DRD2 variants have been associated with language-related abilities (e.g., vocabulary, verbal performance). Environmental influences on the dopaminergic system (stress, enrichment) can alter dopamine function and cognitive outcomes. Parental education correlates with children's academic, intelligence, and language performance and serves as a proxy for home environment quality and SES; prior studies report GxE effects involving PE for reading-related genes. These lines of evidence motivate examining DRD2's role in math ability and its interaction with PE.

Participants: 1,097 primary school students (ages 7–13; mean 9.59; 48.4% girls) from urban and rural schools in Shaanxi and Gansu provinces, China, recruited via convenience sampling; exclusions included severe mental disorders and math test scores below three SDs (n=97). Nonverbal IQ was screened using Raven’s Standard Progressive Matrices; all retained had normal IQ. Ethics approval from Shaanxi Normal University; parental consent obtained. Parental Education (PE): Available for 798 participants; coded 1–8 (1=primary school to 8=post-doctoral). If both parents' data were present, their average was used; if only one parent’s data, that value was used. SNP selection and genotyping: 28 DRD2-related SNPs tied to working memory/language were genotyped and imputed (Michigan Imputation Server, Minimac4; Genome Asia Pilot-GasP GRCh37/hg19). QC thresholds: MAF>1%, call rate≥0.95, mind≥0.90, HWE p≥1e-5. Phenotype: Chinese version of the Heidelberg Mathematics Test assessing 11 subtests across arithmetic operations, mathematical reasoning, and visuospatial skills; reliability α=0.88; validity=0.91. Data analysis: Linear regressions (PLINK) tested SNP main effects; SNP×PE interactions via PLINK and GLM in R. Haplotype blocks defined by LD (r^2); haplotype linear association tests via PLINK; haplotype×PE interactions via GLM in R. Rare haplotypes (freq<0.02) excluded. Multiple-testing correction: Bonferroni thresholds—single SNP and SNP×PE p≤0.0018 (0.05/28); haplotype and haplotype×PE thresholds varied by haplotype set size (e.g., three-SNP sets p≤0.0019; four-SNP p≤0.0020; five-SNP p≤0.0021). Distributional checks indicated near-normal phenotypes; SNP–PE correlations were nonsignificant, supporting use of PE as an environmental moderator.

- Single SNP main effects: No individual DRD2 SNP showed a significant association with math phenotypes after correction. - SNP×PE interactions: Significant interactions for division and equation tasks: rs4648317×PE (A allele; Division: BETA −0.9128, p=0.0016); rs4350392×PE (A; Division: −0.931, p=0.0012; Equation: −0.7301, p=0.0020); rs4938019×PE (C; Division: −1.043, p=0.0003; Equation: −0.8132, p=0.0005); rs10891556×PE (T; Division: −1.051, p=0.0003; Equation: −0.7661, p=0.0010). Simple slope analyses indicated that homozygotes for the first allele of these SNPs showed worse performance under certain PE levels, consistent with differential susceptibility. - Haplotype associations: Robust effects on basic math abilities. Significant 3-SNP haplotypes: rs2734831–rs1125394–rs7103679 TCC (Magnitude perception: BETA 1.64, p=0.001; Quantity counting: BETA 1.15, p=0.001); rs1125394–rs7103679–rs7125415 CCC (Magnitude perception: 1.64, p=0.001; Quantity counting: 1.16, p=0.001). Significant 4-SNP haplotypes: rs2734836–rs2734831–rs1125394–rs7103679 TTCC (Magnitude perception: 1.744, p=0.0005; Quantity counting: 1.193, p=0.0008); rs2734831–rs1125394–rs7103679–rs7125415 TCCC (Magnitude perception: 1.625, p=0.0011; Quantity counting: 1.117, p=0.0017); rs1125394–rs7103679–rs7125415–rs4648318 CCCC (Magnitude perception: 1.698, p=0.0007; Quantity counting: 1.153, p=0.0013). Significant 5-SNP haplotypes included rs1125394–rs7103679–rs7125415–rs4648318–rs12574471 CCCCT (Subtraction: 1.68, p=0.0005; Magnitude perception: 1.807, p=0.0004) and rs7103679–rs7125415–rs4648318–rs12574471–rs4274224 CCCTA (Subtraction: 1.354, p=0.0012). - Haplotype×PE interactions (Bonferroni-significant): Magnitude perception: rs4648318–rs12574471–rs4274224 CTAXPE (BETA 1.54532, p=0.0009); Subtraction: rs12574471–rs4274224–rs4581480–rs7131056 CGTCXPE (BETA −5.561375, p=0.0011); Mathematical reasoning: rs6275–rs1076560–rs2511521 ACGXPE (BETA −24.02675, p=0.0010) and GAAXPE (BETA −23.9267, p=0.0010); Quantity counting: rs7103679–rs7125415–rs4648318 TCCXPE (BETA 1.5394, p=0.0006) and TCTXPE (BETA 1.24921, p=0.0005); Spatial conception: rs4350392–rs4938019–rs1799978–rs10891556 CTTGXPE (BETA 0.94203, p=0.0017). Overall, haplotypes showed stronger associations than single SNPs, and PE moderated several haplotype effects, especially for quantity counting and other complex tasks.

Findings support DRD2 as a generalist gene influencing mathematical ability, consistent with its established roles in working memory and language. While single-SNP main effects were not detected, haplotypes comprising multiple DRD2 variants showed robust associations with basic math components (magnitude perception, quantity counting) and some arithmetic skills (subtraction). Gene–environment analyses revealed that PE moderates DRD2 effects predominantly for more complex mathematical tasks (division, equations, mathematical reasoning), with simple slope analyses indicating a disordinal interaction pattern aligning with the differential-susceptibility model: individuals carrying risk alleles performed worse in low-PE environments but better in high-PE contexts. Certain intronic variants (e.g., rs1076560, rs2283265) affected math outcomes only within haplotype×PE interactions, underscoring environmental modulation of dopaminergic genetic influences. These results highlight the importance of considering aggregated genetic effects and environmental quality when studying complex cognitive traits like mathematical ability.

The DRD2 gene contributes to children's mathematical ability, primarily via haplotype-level effects. Parental education moderates DRD2’s influence, with interactions for both basic and advanced math skills and evidence favoring the differential-susceptibility model for individual SNP×PE interactions. The study underscores the universal importance of positive educational environments and suggests that identifying generalist genes can inform personalized educational interventions. Future research should replicate in larger cohorts, examine multi-gene interactions, and include working memory and neural mechanisms to clarify pathways linking DRD2 to math abilities.

- Exploratory cohort size; findings require replication in larger, independent samples. - PE, while treated as environmental, may carry genetic correlates; disentangling intergenerational vs environmental influences is needed. - DRD2 cannot explain all facets of mathematical ability; polygenic influences should be examined. - Gaussian model fitting for haplotype and interaction analyses may introduce numerical instability, risking false positives; more robust statistical models are warranted. - Working memory and other neural mechanisms were not directly measured, limiting mechanistic interpretation of DRD2 effects on math.