Table 3.
Results of tests of associations between standardized total gray matter volume and dimensions of psychological problems defined alternatively by correlated factors and bifactor modeling of parent ratings of European ancestry children in the baseline assessment, and first and second annual follow-up assessments, controlling age, sex, scanner, and the first 10 principle components for ancestry among European ancestry children with non-missing externalizing polygenic scores.
| Correlated Factors Models of Psychological Problem Ratings | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Internalizing | Conduct Problems | ADHD | |||||||||
| β (SE) |
R2 | p | β (SE) |
R2 | p | β (SE) |
R2 | p | |||
| Baseline Assessment (N = 4063) | |||||||||||
| −0.089 (0.024) |
1.0% | <0.0001 | −0.216 (0.023) |
4.7% | <0.0001 | −0.175 (0.023) |
3.0% | <0.0001 | |||
| Follow-up Wave 1 Assessment (N = 3940) | |||||||||||
| −0.051 (0.024) |
<1.0% | 0.036 | −0.176 (0.025) |
3.1% | <0.0001 | −0.131 (0.024) |
1.7% | <0.0001 | |||
| Follow-up Wave 2 Assessment (N = 3705) | |||||||||||
| −0.056 (0.025) |
<1.0% | 0.025 | −0.196 (0.024) |
3.8% | <0.0001 | −0.136 (0.024) |
1.8% | <0.0001 | |||
| Correlated Factors Models of Psychological Problem Ratings | |||||||||||
| General Psychopathology | Specific Internalizing | Specific Conduct Problems | Specific ADHD | ||||||||
| β (SE) |
R2 | p | β (SE) |
R2 | p | β (SE) |
R2 | p | β (SE) |
R2 | p |
| Baseline Assessment (N = 4063) | |||||||||||
| −0.177 (0.025) |
3.1% | <0.0001 |
0.062
(0.027) |
<1.0% | 0.021 | −0.135 (0.030) |
1.8% | <0.0001 | −0.041 (0.032) |
0.210 | |
| Follow-up Wave 1 Assessment (N = 3940) | |||||||||||
| −0.128 (0.025) |
1.6% | <0.0001 |
0.058
(0.024) |
<1.0% | 0.016 | −0.145 (0.032) |
2.1% | <0.0001 | −0.026 (0.033) |
0.431 | |
| Follow-up Wave 2 Assessment (N = 3705) | |||||||||||
| −0.154 (0.027) |
2.4% | <0.0001 |
0.078
(0.28) |
1.0% | 0.005 | −0.129 (0.032) |
1.7% | <0.0001 | 0.003 (0.035) |
0.942 | |
Note. Factor loadings in the measurement models used in each predictive regression were fixed based on unstandardized factor loadings in a bifactor model without covariates. Tabled p’s are unadjusted; coefficients in bold are significant at p < 0.05 after false discovery rate correction for 21 tests of significance.