Cannabis, Honestly

The Dunedin study asked a question most drug research cannot: what happens to IQ when you measure the same people before they ever used cannabis — and again at 38? The answer reshaped the debate. Then twin studies pushed back, and the honest position landed somewhere neither side's marketing admits: age of onset matters most, the effect sizes are contested, and today's products are not the ones the old studies tested.

The Dunedin Findings

The Dunedin study has followed everyone born in one New Zealand city in 1972–73 their entire lives. That design gave researchers something no cross-sectional survey can: IQ tests from childhood, before anyone used cannabis, and the same tests again at age 38 (Meier et al., PNAS, 2012). The finding that made headlines: participants who met cannabis-dependence criteria at three or more assessments and who started as teenagers lost about 8 IQ points on average relative to the rest of the cohort — and quitting in adulthood did not restore the loss. Those who started using as adults showed smaller declines.

• 🧮 The strength: pre-use baseline measurements mean the decline cannot be explained by users starting out lower — a confound that ruins most drug studies.
• 🔬 The weakness: the dependent group was small — 38 people met dependence criteria repeatedly — so the 8-point figure comes with wide uncertainty bands.
• 📐 What it is not: not evidence that occasional adult use costs IQ points. The signal belonged to early onset plus heavy, persistent use.
• 🧠 The mechanism story: adolescence is a late-window of frontal development, which is the brain-aging primer's framework for why timing would matter.

The Twin-Study Pushback

Four years later, a twin analysis reached for the same question and got a different answer. Looking at two longitudinal twin cohorts, researchers compared cannabis-using twins with their non-using siblings — same genes, same household, same neighborhood — and found no significant IQ decline (Jackson et al., PNAS, 2016). If the Dunedin decline disappears when family background is held constant, the argument goes, then what Dunedin measured was partly the family environment cannabis users disproportionately come from, not the cannabis.

A large meta-analysis sits between the two poles: small but significant overall cognitive effects in adolescent and young-adult users, concentrated in verbal learning and memory, with most deficits attenuating after at least 72 hours of abstinence (Scott et al., JAMA Psychiatry, 2018). The honest synthesis: effects exist, they are modest for most users, they are largest for early-onset heavy use — and the strongest single claim in the literature (8 permanent points) rests on a small group and has a credible confound rebuttal.

Age of Onset Is the Variable

Every major dataset points the same direction when split by age: the adolescent-onset story and the adult-onset story are different stories. The brain's frontal systems keep maturing into the mid-20s, and that is where most of the developmental concern lives. The pattern in the data:

• 🧒 Onset before ~18: the strongest associations with lasting deficits, especially with daily or dependent use — the Dunedin pattern.
• 🧑 Onset in the 20s and later: weaker and often transient associations; acute effects while using, limited evidence of permanent change.
• 🗣️ The consistent domain: verbal learning and memory — the same domain that shows up in the aging-brain timeline as an early, sensitive measure.
• 📅 Dose and frequency track risk more than "ever used" does: the studies that find effects are studies of heavy, regular use.

Potency Has Changed Under You

One of the quietest confounds in this literature is that "cannabis" today is not the substance the older studies measured. Seized-cannabis testing shows average THC content climbing from roughly 4% in 1995 to about 12% by 2014 (ElSohly et al., Biological Psychiatry, 2016) — and modern flower, extracts, and vaporizer oils routinely reach far higher. A 2020s "single use" can deliver several times the THC of a 1990s joint.

Why this matters: dose-response logic. If the harms concentrate at high doses and high frequency, then potency inflation quietly moves today's casual user toward the exposure levels the old studies called heavy. Concentrates and edible products also make dose control harder — the difference between 5 mg and 50 mg of THC is not obvious from one brownie.

Acute Effects: The Part Nobody Debates

Buried under the long-term arguments is a set of findings everyone agrees on, because they show up in every careful study: while intoxicated, attention, working memory, and reaction time are measurably impaired, and the impairment outlasts the subjective high. Regular users develop partial tolerance and tend to underestimate their own impairment — which is exactly what makes the driving data consistent.

• 🚗 Driving risk rises with recent use — roughly doubling crash involvement in several epidemiological analyses, with the strongest signal in the hours after use.
• ⏳ Edibles delay the dose. Onset runs one to two hours and lasts far longer than smoked flower, which is why inexperienced users overshoot — impairment arrives after they assumed they were fine.
• 🧠 Attention and memory take the hit even at single doses — the same domains the long-term studies flag, which is a hint that the chronic findings are a cumulative version of the acute ones.

The Older-Adult Question

Cannabis use is rising fastest among people over 55 — often for sleep, pain, or anxiety. The honest problem: almost none of the cognitive literature studied aging brains, and the pharmacology is not neutral. THC interacts with medications that already burden cognition in this age group — sedatives, anticholinergics, and alcohol — and adds fall risk on top of them. The defensible position for older adults is caution plus a clinician conversation, not the marketing claim that cannabis is the "natural" alternative to everything. Medical use for specific conditions (chemotherapy nausea, some seizure disorders, certain pain syndromes) is a different question from recreational cognitive effects, and one for a prescribing clinician. One product-adjacent note: CBD — the non-intoxicating cannabinoid — is approved as a medication for specific seizure disorders, but its popular claims for cognition, anxiety, and sleep rest on thin evidence. CBD is not the "safe half" of the plant in any sense that the trials have established.

Questions, Answered Briefly

• 🧠 Did cannabis permanently lower my IQ? Probably not, unless use started in adolescence and stayed heavy — the lasting-deficit signal belongs to that specific pattern, and even then the magnitude is debated.
• 🛑 If I stop, does my brain recover? Most acute deficits clear within days to weeks of abstinence; the Dunedin cohort is the cautionary exception for early-onset dependent use.
• 🌿 Is occasional adult use risky? The evidence for lasting cognitive harm from occasional adult use is weak; the settled costs are acute impairment, driving risk, and the possibility of escalating frequency.
• 👵 Should I use it for sleep or pain in my 60s? That is a clinician question, not a counter question — the interaction and fall risks change the calculus in ways this page cannot.
• ⚖️ What should I believe online? Neither the "melt your brain" nor the "totally harmless" camps survive contact with the data; both are marketing.

The Moderation Question, Honestly

If you use, the levers the evidence actually supports are unglamorous: start late, keep frequency low, choose lower-potency forms, measure edibles in small units, and never drive within hours of use. None of those rules are about identity or politics — they are just the dose and timing structure the findings point at.

Related Topics