Approved Research

Effects of moderating or desisting from drinking on brain structural and functional characteristics

Lay summary

There is scientific evidence suggesting that drinking alcohol during adolescence and early adulthood may have harmful effects on brain development. This is a crucially important period in brain development, when the capacity to exert control over one's behavior, make decisions flexibly, anticipate and plan for the future, and behave in an adaptive manner develop more fully, all of which help lay the foundation for being a competent adult. It is important therefore to determine the exact nature of drinking effects on the brain. Understanding such effects is clearly a public health issue of substantial public interest. Because most people moderate their drinking, however, understanding the extent to which any effects of alcohol use lessen when individuals moderate their drinking or stop altogether is arguably equally important, with equally important public health implications.

We will address these two broad aims in a sample of young adult twins from the US and in the UK Biobank sample. It is impossible to conduct controlled studies with people, assigning some to continue drinking heavily and others not to. Twins represent a "natural experiment," however. We will use an innovative twin-differences design to determine whether differences between twins in their drinking patterns are associated with differences between them in brain structural and functional measures. In effect, the brain of the twin who drinks less or who reduces drinking more approximates what the brain of the twin who drinks more or who continues to drink heavily would have looked like had they not continued to drink heavily. This is our best approximation to an actual scientific experiment. We will examine several MRI measures of brain structure and function to determine which of these are associated with drinking during adolescence and which are associated with reducing the amount of drinking one does in one's 20s and 30s. The twin-difference design will inform us about whether any of these associations that are statistically significant indicate that drinking, or reducing one's drinking, truly causes the association, and which are part of a pre-existing liability for drinking heavily and continuing to drink heavily in the first place.

We will determine whether any findings consistent with a causal influence of drinking patterns also holds true in the youngest participants in the UK Biobank sample. This would indicate that they are not specific to one particular sample of participants, thereby increasing our confidence that they are real.