What it's actually good for
Inositol is a sugar alcohol your body makes from glucose and also gets from food (beans, grains, citrus fruit). Myo-inositol, the form that matters here, is a precursor in the phosphatidylinositol signaling pathway — the same second-messenger system that serotonin and insulin receptors use to relay signals inside cells. That dual role is why the research splits cleanly into two unrelated stories: psychiatric symptom relief at very high doses, and metabolic/hormonal improvement at moderate doses in PCOS.
The psychiatric case rests on a small number of well-designed but old trials. Palatnik and colleagues (2001) ran a crossover comparing up to 18 g/day of inositol against 150 mg/day of fluvoxamine in panic disorder patients — both groups improved comparably on anxiety measures, inositol edged out fluvoxamine on panic attack frequency (a drop of 4.0 vs. 2.4 attacks per week), and fluvoxamine caused more nausea and tiredness. Fux and colleagues (1996) found the same 18 g/day dose significantly lowered Y-BOCS scores versus placebo in a 13-patient OCD crossover trial. These are real, randomized, controlled results — not anecdote — but the samples are tiny by modern standards (20 and 13 people), the trials are decades old, and neither has been meaningfully replicated at scale. Broader anxiety evidence is thinner still: some observational work links lower inositol levels to generalized anxiety disorder, but there's no large dedicated GAD trial to confirm supplementation fixes it.
What the research says
Panic disorder and OCD (Grade B). The Palatnik and Fux trials are the backbone of inositol's psychiatric reputation, and they're genuinely good data as far as they go — double-blind, placebo- or active-comparator-controlled, with clinically meaningful effect sizes. What keeps this at a B rather than an A is scale: two small, old trials aren't enough to call this settled, and nothing since has come close to replicating them in a larger cohort.
PCOS (Grade B, probably the sturdier claim). A 2017 meta-analysis by Unfer and colleagues pooled 9 RCTs and 496 women with PCOS and found myo-inositol significantly reduced fasting insulin and the HOMA insulin-resistance index. SHBG (a marker of improved androgen handling) rose meaningfully, but only in trials running 24 weeks or longer — this is not a fast fix. Testosterone trended down but didn't reach statistical significance. More trials and more consistent dosing protocols (myo- and D-chiro-inositol are often combined in a 40:1 ratio) would strengthen this further, but the trial count here is larger than the psychiatric literature.
How much, and which form
For mood and OCD, the studied dose is 12-18 g/day, split into 2-3 doses — a meaningful amount of powder (roughly a tablespoon or more per dose), which is the main reason this isn't a mainstream first-line approach. For PCOS, 2-4 g/day of myo-inositol, often paired with D-chiro-inositol at a 40:1 ratio, matches the trials Unfer's meta-analysis pooled. Myo-inositol is the studied form; D-chiro-inositol alone has less independent support.
Safety & interactions
Inositol is naturally occurring and well-tolerated even at gram-level doses. The main downside is gastrointestinal: nausea, gas, and loose stools become more common above roughly 12 g/day, and are usually manageable by titrating up slowly. Because inositol feeds into the same signaling pathway serotonin receptors use, there's a theoretical case for additive effects when combined with SSRIs or other serotonergic drugs — worth a conversation with a prescriber rather than a reason to avoid combining outright. This is informational, not medical advice — check with a clinician before starting, especially if pregnant, nursing, or on psychiatric medication.
How we picked the brand
An inositol product earns a spot when it's pure myo-inositol (not a proprietary blend padded with fillers), independently tested, and sold at a price that makes daily gram-level dosing affordable rather than a splurge — powder form, since capsules can't realistically deliver 12-18 g/day.