Assessing the impact of environmental factors on the adolescent brain: the importance of regional analyses and genetic controls

There is substantial brain development during adolescence, which continues up to the early 20s. One of the earliest questions in neuroscience has been the role of experiences, or environmental factors, in that development. As pointed out by Turkheimer ¹ , “development is fundamentally nonlinear, interactive, and difficult to control experimentally”. But, in the last two decades, there has been an enormous progress in brain measurements, cognitive testing, and sample sizes.

Perhaps the most well‐studied environmental factor in cognitive development is socioeconomic status (SES). This index is a combination of multiple factors that can impact the cognition of a child, such as the influence of parental education (e.g., the types of books in the household and the intellectual stimulation at the dining table); the influence of income in the quality of the school and the number of extra‐curricular activities; the influence of the neighborhood in the type of peers and services available. Children born and raised to parents with low SES have on average a worse development in a wide range of areas: they tend to have lower cognitive abilities and worse academic performance, and to suffer more frequently from mental disorders2, 3.

It is not surprising then that functional magnetic resonance imaging (fMRI) studies show that the brain systems whose activity is affected by SES are frontal and parietal regions related to reasoning and executive functions, temporal language areas, as well as the hippocampus and the medial temporal lobe, which are related to long‐term memory ⁴ .

The neurological impact of SES goes even beyond task‐specific brain activity at the moment of scanning (as measured by fMRI). SES is one of the few environmental variables that we know can impact the very macrostructure of the brain (as measured by structural MRI), such as cortical surface area, which is typically stable over months and even years of life. A large study imaging the brains of 1,099 individuals between the ages of 3 and 20 years found that the total cortical surface area was related to both parental income and parental educational level ⁵ . There were regional associations in most parts of the cortex, but in particular in regions supporting language, reading, executive functions, and spatial skills. Other studies have supported these findings, but some show that the frontal cortex is especially targeted, while others show no single region that is specifically connected to SES ⁴ .

There are, however, some caveats that should be kept in mind when interpreting MRI results in this field. First, studies typically suffer from a methodological “blind spot”, because global differences in the structural measures, for example in cortical surface area, are often not taken into account. It is therefore unclear if regional findings mean that low SES selectively impacts only specific brain areas or if the impact is better described as broad and global, with minor local variability. Second, SES is likely not to be an entirely environmental factor, but to have a substantial genetic component. For example, it has been reported ⁶ that a set of genetic markers explained as much as half of SES contributions to school achievement in 16‐year‐olds.

In a recent study ⁷ , we used a sample of 551 typically developing adolescents, studied at ages 14 and 19, to try to tackle the problem of the entanglement between genetic and environmental effects in the developing brain. In order to estimate genetic effects, we used a combined measure, called polygenic score (PGS), from several thousand DNA markers that were selected and given a weight to optimize prediction of educational attainment (henceforth called EduYears‐PGS). As expected, EduYears‐PGS and SES were moderately correlated. But, even when controlling for this overlap, SES still had independent effects on cognitive ability at age 14. Interestingly, the SES effect was about twice as strong as that from EduYears‐PGS. When analyzing the change in cortical surface area from age 14 to 19, there was an effect of SES, but not EduYears‐PGS. This indicated that SES continued to affect brain maturation throughout adolescence.

One limitation of the study was that, although the EduYears‐PGS measure is the most powerful genetic predictor available for educational attainment, it does not capture all the genetic variance associated with SES, as suggested by twin studies. Another limitation is that the EduYears‐PGS was optimized to predict educational outcome, rather than SES. However, our post‐hoc analysis suggested that the SES associations that we found were driven almost exclusively by differences in parental education and, as a control for that, the EduYears‐PGS we used is optimal.

Regarding the methodological “blind spot” problem mentioned above, we also obtained structural MRI from the adolescents. Initially, we found that both EduYears‐PGS and SES were positively correlated with total cortical surface area. However, after controlling for the global effects, there were no additional regional associations of SES to cortical surface area. This means that there were no signs of any particular structure or neural system being selectively affected above and beyond the broad effects of SES. The EduYears‐PGS, on the other hand, had an additional regional association with cortical surface area in the right parietal lobe.

The association of SES to global cortical surface area means that the behavioral and psychological consequences of low SES are likely wide‐ranging. What could be the environmental factors behind such a broad effect in the developing adolescent brain? Low SES is associated with a range of environmental factors that could impact cognition and brain development. These include toxins, infections and stress during gestation, inferior nutrition, chronic stress, and lack of cognitive stimulation during childhood and adolescence8, 9.

Because research typically shows that the impact of SES continues throughout adolescence, one could expect that the environmental factors during this period play an especially important role, such as chronic stress or lack of intellectual stimulation, rather than gestational factors. Furthermore, if these broad brain impacts (as suggested by regional analyses controlling for global measure) are indeed true, this has negative implications for societies. It makes it less likely that any particular intervention, such as language training, could compensate for the cognitive and behavioral problems. An unfortunate implication of poverty.

However, it is possible that the global neural effect of low SES is the result of a combination of a multitude of environmental effects, and that each of these can be identified and targeted. Future research might thus highlight the role of specific environmental factors in affecting cognitive development, which could help inform policy decisions.