Magnesium L-threonate is a real, named, low-elemental salt with a rat-CSF story and a thin, entirely industry-tied human file. What the file actually supports: two sponsor-linked RCTs (Liu 2016 n = 44 completers, composite d = 0.91 at 12 weeks; Lopresti 2026 n = 100, NIH composite +2.8 points over placebo at 6 weeks, one-tailed p = 0.043) report better *composite cognition* than placebo at ~108–145 mg elemental/day. Individual tests, Raven’s, and “N-year younger brain” arithmetic do not carry those headlines. Sleep is not established for this salt: Liu’s PSQI missed, Hausenblas’s ISI missed, Lopresti’s Oura sleep metrics missed; leftover wins are subscales and proprietary ring scores. Anxiety is a between-group null in the only RCT, published twice. ADHD and dementia remain open-label n = 15.
The distinctive claim — that Magtein, unlike glycinate or oxide, raises **human** brain or CSF magnesium — has **no human measurement**. Slutsky’s +7–15% CSF is rats at 50 mg/kg/day elemental, roughly an order of magnitude above the human elemental dose. Until someone taps CSF or runs ³¹P-MRS after oral Magtein, “brain magnesium” is a marketing sentence.