Neuroscience

What Twin Studies Can Tell Us About Cannabis and IQ

·

Two research participants arrange colored shapes at neighboring desks.

A teenager takes an intelligence test, starts using cannabis and later scores lower on part of a second test. The sequence is important. An exposure has to precede a change if it is to cause it. Yet the sequence alone leaves open a familiar possibility. Things that shaped the teenager’s test scores may also have shaped the decision to use cannabis. A twin who grew up in the same household offers an unusually informative comparison.

First compare families, then look inside them

In two longitudinal studies of twins, researchers examined changes in intellectual performance and asked whether the twin who used cannabis showed a greater decline than the co-twin who did not. Looking across families, young people who used cannabis differed from nonusers on some measures. The comparison within pairs did not provide the pattern one would expect from a straightforward account in which exposure itself explained the differences.

Why can the answer change? Imagine that twins in one family grow up with plentiful books and a stable school routine, while another pair experiences repeated school changes. If cannabis use also varies between those families, an ordinary comparison could attribute some difference in scores to the drug that belongs to schooling. Comparing siblings within each family removes much of that shared history from the contrast. The actual factors may be far harder to name than the hypothetical books.

What twins do and do not share

Identical twins share nearly all their inherited DNA. Fraternal twins share less, though they are typically the same age and have much family experience in common. Neither kind of twin lives an identical life. Different friends, illnesses, school experiences and other substance use can distinguish one sibling from the other. A co-twin design tightens the comparison. It does not randomly assign cannabis exposure.

The comparison also becomes less informative if siblings have similar use histories or if those histories are measured imprecisely. The investigator needs enough pairs with a meaningful difference in exposure. A large total sample can contain a much smaller number of pairs that qualify. Readers trying to judge precision should look for the size of that informative subgroup and the uncertainty around the estimate.

What exactly changed on the test?

An intelligence test includes tasks with different demands. A measure drawing heavily on vocabulary reflects years of exposure to language and education. Another task may ask someone to manipulate unfamiliar shapes under time pressure. A difference on one component does not automatically imply an equivalent difference everywhere. Repeated testing also gives investigators more to work with than a single score obtained after use has begun.

A separate co-twin investigation of cannabis and cognition illustrates why timing and the particular cognitive measure recur in this literature. Researchers can study an immediate effect during intoxication, a difference shortly after abstinence or a developmental change across years. The word cognition can apply to all three. Their causal questions are quite different. A study that followed young people through abstinence, for example, could test for changes in learning over the course of a month. It could not put the clock back to their first exposure.

An informative result with a narrow reach

A meta-analysis of cognitive tests in young people approached the evidence from another direction, pooling results across samples and examining time since use. That design gives breadth. Twin studies give leverage over family-level confounding. Agreement between designs is particularly interesting precisely because their weaknesses differ.

The cautious conclusion from the twin comparison is specific. A difference seen between unrelated users and nonusers was less convincing as an effect of cannabis when siblings with shared backgrounds were compared. That finding cannot establish the absence of an effect at every level of exposure or for every individual. It establishes something useful about the original story. The teenager’s later score needs a comparison that takes the rest of the teenager’s life seriously. A twin makes part of that life unusually visible to science.