# Magnesium L-Threonate: Brain Access, Synaptic Function, and the Limits of Clinical Evidence

- Canonical: https://nutrifit.health/knowledge/magnesium-l-threonate-blood-brain-barrier-synapses
- Language: en
- Published: 2026-10-08T04:39:18.553Z
- Updated: 2026-10-08T04:39:18.553Z
- Author: NutriFit Editorial

> How magnesium L-threonate relates to brain magnesium, NMDA receptors, and synaptic plasticity. This review covers the compound’s chemistry, transport across barriers, animal and human memory studies, dosing, and safety without overstating the evidence.

Magnesium has two roles in the nervous system that are easy to conflate. Cells need it to work with ATP, nucleic acids, and enzymes, while it also acts as a physiological regulator of excitability: at a negative membrane potential, Mg²⁺ blocks the pore of the NMDA receptor. A [blood magnesium measurement](/knowledge/magnesium-functions-food-tests-safe-supplementation) alone therefore cannot answer whether a magnesium supplement might affect memory. The relevant questions are whether bioavailable magnesium in the brain changes, how that relates to synapses, and whether experimental findings translate to people.

Magnesium L-threonate drew attention after experiments in which sustained elevation of brain magnesium in animals accompanied changes in synaptic plasticity and performance on learning tasks. Magtein® is the trade name for this salt's commercial ingredient; it is neither the name of a neurotransmitter nor evidence of a special delivery route. This article examines the complex's chemistry, transport across barriers, synaptic function, and the limits of small clinical studies. A crucial distinction separates Mg²⁺ biophysics, findings in rodents, and the still-limited outcomes measured in people.

## From the molecule in a capsule to magnesium in the brain

### L-threonate chemistry and the ligand's origin

Magnesium L-threonate is commonly described as one cation complexed with two L-threonate anions: $\mathrm{Mg(C_4H_7O_5)_2}$. At a molar mass of about 294.5 g/mol, elemental magnesium accounts for $24.305/294.5\approx8.25\%$ of the idealized anhydrous compound. A real product may contain water and excipients, so dosing should be based on the amount of **elemental magnesium** on the label, not just the powder's mass. L-threonic acid has two chiral centers; the source gives the $(2R,3S)$ configuration for 2,3,4-trihydroxybutanoic acid. The carboxyl and neighboring hydroxyl groups may coordinate Mg²⁺, changing solubility and the dissociation equilibrium. In water, free Mg²⁺ is hydrated; its first shell is commonly described as containing six water molecules. These chemical features help frame questions about absorption, but do not show that the intact complex crosses the BBB.

L-threonate is associated with the breakdown of ascorbic acid. Ascorbate is oxidized to dehydroascorbate; opening of the lactone ring yields 2,3-diketogulonate, from which threonate may form among other products. Schematically, this route is written as $\text{L-ascorbate}\rightarrow\text{dehydroascorbate}\rightarrow\text{2,3-diketo-L-gulonate}\rightarrow\text{L-threonate}$. The intermediate 2,3-diketogulonate can break down along several routes, including oxalate formation; the equation above shows only the branch relevant to threonate and is not a complete carbon balance. This origin explains why the ligand occurs in biological systems, but proves neither unconditional safety at every dose nor the existence of a dedicated transporter for the intact complex. Ascorbate breakdown has other branches; it should not be presented as a single stoichiometric route to threonate.

### Two barriers and an open mechanistic question

Brain-capillary endothelium is joined by tight junctions involving claudin-5, occludin, and ZO proteins; together with pericytes and astrocyte processes, it forms the neurovascular unit. Tight junctions sharply restrict the passage of hydrated ions between cells. The choroid plexus is a second interface and regulates cerebrospinal-fluid composition. Plasma, cerebrospinal fluid, and brain-tissue Mg²⁺ concentrations are therefore three distinct quantities. Even a measured change in cerebrospinal fluid is not a measurement of free magnesium at every synapse.

Cellular magnesium regulation involves the TRPM7 channel-kinase, MagT1, the SLC41A1 exchanger, and CNNM-family proteins, including the discussed CNNM2 and CNNM4. Feedback regulation of ion entry and ion export helps explain why a large rise in plasma Mg²⁺ need not produce a proportional rise in cerebrospinal fluid. The original diagram also names CaSR and ARL15 as regulatory candidates. Their functions have been studied in different tissues and experimental systems, but those findings do not establish a ready-made sequence in which “MCT1 carries Mg(L-threonate)₂ across the BBB and MagT1 then unloads it into the brain.” The supplied evidence does not establish such a route for the intact complex; MCT1 and SMCT1 should not be described as confirmed Mg-threonate transporters. Nor do CaSR, ARL15, or CNNM2 form an established sequence of transport steps here. The working question is which concentrations of free and bound magnesium arise after absorption and how their exchange with the central nervous system changes. Supporting the proposed “Trojan horse” mechanism would require measuring chemical species in blood and on both sides of the endothelium, identifying a transporter, blocking transport competitively, and measuring flux. A change in total cerebrospinal-fluid Mg alone would not be enough. The choroid plexus deserves separate attention: its epithelium interfaces with cerebrospinal fluid differently from capillary endothelium's interface with the interstitial space, so findings from the two interfaces cannot be combined without additional measurements.

[![Diagram distinguishing blood magnesium, BBB endothelium, and brain magnesium, and marking the unknown route of the intact Mg(L-threonate)₂ complex.](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-mg-barrier.en.png)](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-mg-barrier.en.png)

Experiments by Slutsky and colleagues found that raising brain magnesium with L-threonate in animals accompanied improvements in several memory and synaptic-plasticity measures. This is a reason to study the form further, but not a basis for assigning precisely zero brain penetration to every other compound. The original comparison table presented bioavailability percentages, day-by-day cerebrospinal-fluid Mg²⁺ values, and morphometric effects for oxide, chloride, citrate, and bisglycinate as if they came from one direct experiment. The supplied studies do not establish such a complete, comparable matrix; using it to guide clinical choices with precision to tenths of a percent would be an error.

| Measured level | What it can establish | What it does not establish |
| --- | --- | --- |
| Elemental magnesium in a dose | The amount of Mg²⁺ supplied by the product | How much reaches the brain |
| Blood magnesium | Systemic status and some safety questions | The concentration in the synaptic cleft |
| Cerebrospinal-fluid magnesium | A change in one central compartment | A uniform change across all brain regions |
| Synaptic markers and rodent behavior | An association between exposure and a biological response in a model | Clinical memory improvement in people |

## When a synapse changes its structure

### The hippocampus and prefrontal cortex

The hippocampus is not a uniform “memory center.” Incoming signals pass through the dentate gyrus and the CA3 and CA1 fields; these networks contribute to separating similar events, linking elements of experience, and subsequent consolidation. The medial prefrontal cortex supports the active maintenance of information, action selection, and control of emotional responses. In magnesium experiments, these regions are important sites for assessing plasticity, not established sites of selective supplement accumulation in people. The absence of a measured effect in another region does not prove that the effect is entirely absent there.

### Presynaptic stores and the postsynaptic scaffold

Synaptophysin is a protein in synaptic-vesicle membranes, so its staining can help assess presynaptic terminals. PSD-95, encoded by *DLG4*, organizes the protein scaffold of the postsynaptic density and holds signaling complexes near AMPA and NMDA receptors. Animal studies have compared higher brain magnesium with changes in these markers and in the number of functional contacts. However, greater staining intensity does not equal a defined percentage of new mature synapses, and the shape of a mushroom-shaped spine does not guarantee better behavior. The universal increases of 36–44% in synaptophysin and 32–41% in PSD-95 claimed in the original text cannot be generalized across regions, time points, and doses.

The original experiments used morphometry, immunolabeling, and electrophysiological responses; these methods address different questions. A synaptophysin-positive punctum may indicate the presence of vesicular membrane, PSD-95 may indicate organization of the postsynaptic density, while a paired functional synapse requires coordinated localization of pre- and postsynaptic structures and a reproducible current. Thus, even increases in two markers do not replace counting functioning contacts while controlling for neuronal population size and tissue volume.

Presynaptic vesicles occupy different functional states. The reserve pool is associated with the cytoskeleton, the recycling pool participates in repeated exocytosis cycles, and the readily releasable pool (RRP) comprises vesicles that are docked and primed for rapid release. When Ca²⁺ enters the terminal, the SNARE complex, including vesicular VAMP2/synaptobrevin, supports membrane fusion and glutamate release. The approximate division given in the original source—80–90% reserve, 10–15% recycling, and 1–2% immediately releasable—depends on synapse type and measurement method; it illustrates the order of magnitude rather than fixed proportions in the human brain. In model studies, a more readily available pool together with changes in release probability is discussed as one way to strengthen responses to a series of impulses while keeping background noise low. RRP size is typically estimated with stimulus trains or hypertonic sucrose and inferred from total release charge, adjusted for release probability and recycling. Stimulation at about 40 Hz, mentioned in the original source, can test transmission at high frequency, but the result depends on the protocol and is not a measurement of vesicle number in the living human brain. This hypothesis should not be turned into a claim that taking a particular capsule increases a person's RRP by a fixed percentage.

The presynapse-to-spine connection is best viewed as a feedback sequence: a vesicle releases glutamate, an AMPA receptor changes membrane potential, an NMDA receptor passes Ca²⁺ when the required conditions coincide, and intracellular signaling changes PSD-95 organization and receptor availability. This sequence explains why morphology, electrophysiology, and behavior must be measured separately. It also avoids reducing memory to one protein or one spine type.

## The magnesium “plug” and the window for plasticity

### How the NMDA receptor detects coincident signals

The NMDA receptor is a heterotetrameric ion channel. Binding of glutamate to GluN2 and of [glycine or D-serine](/knowledge/glycine-nmda-sleep-core-temperature) to GluN1 is required for activation, but at negative membrane potentials, Mg²⁺ limits current through the channel. Conserved asparagine residues in the M2 pore region, often discussed as the N-site (identified as Asn598 in the source), contribute to voltage-dependent block; the classic experiments by Nowak and Mayer demonstrated this dependence long before L-threonate appeared as a supplement. “Magnesium plug” is a useful metaphor, but it does not mean that the receptor is sealed at every resting potential or that calcium current is always zero.

When neighboring AMPA receptors conduct Na⁺, the membrane depolarizes, the likelihood of Mg²⁺ block falls, and Ca²⁺ may enter through an activated NMDA channel. This ion becomes an intracellular signal that can alter synaptic strength. The “coincidence detector” concept refers to the joint presence of presynaptic release and postsynaptic depolarization. In electrophysiology, block is characterized by a current–voltage curve at specified extracellular Mg²⁺ and agonist concentrations. If an experiment changes Mg²⁺, it must also account for membrane state, glycine or D-serine availability, and the proportions of different subunits; otherwise, a change in current cannot be attributed to the “plug” alone. The mechanism is not about maximizing calcium influx: a prolonged or spatially misplaced calcium signal can damage a cell, and the physiological response depends on the timing, location, and activity context.

[![Diagram showing glutamate release, magnesium block of NMDA at rest, calcium entry during depolarization, and two plasticity pathways](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-mg-synapse.en.png)](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-mg-synapse.en.png)

This has led to the hypothesis that changing available Mg²⁺ might tune the signal-to-background ratio. It should not be framed as guaranteed, complete suppression of spontaneous glutamate release or protection from all forms of excitotoxicity. If disrupted calcium regulation leads to mitochondrial stress, opening of the permeability transition pore, and caspase activation, that is a separate pathological pathway, not a demonstrated clinical effect of L-threonate. Direct receptor agonists, ampakines, and racetams also do not form one comparison group with a known, uniform risk: the original article contrasted them with magnesium too categorically.

### GluN2A, GluN2B, and the LTP cascade

The GluN2A and GluN2B subunits differ in kinetics, regulation, and distribution in the developing and mature brain. Receptors containing GluN2B often deactivate more slowly, so their signal-integration window may be longer. But they cannot simply be labeled the “juvenile type,” with GluN2A as the sole “adult” type: both subunits coexist in complex combinations. The 30–50 and 200–400 ms values shown in the original source are useful only as examples of experimentally dependent kinetics, not as universal channel-opening times in vivo.

In animal models, higher brain magnesium has been studied alongside GluN2B expression and function. The study by Tang and colleagues on genetic overexpression of *Grin2b* demonstrates the subunit's importance for learning in mice, but a genetic model is not equivalent to a dietary supplement. The claim that L-threonate reproduces “Doogie mice” or raises GluN2B by 60–80% in people is therefore unsupported. It conflates different interventions, tissues, and endpoints.

After Ca²⁺ enters, calmodulin activates CaMKII. Autophosphorylation of CaMKIIα at Thr286 sustains enzyme activity after a brief signal; phosphorylation of GluA1 at Ser831 and changes in AMPA-receptor trafficking may strengthen transmission. In parallel, the calcium-sensitive adenylyl cyclases AC1/AC8 increase cAMP, while PKA and CREB participate in longer-term changes in gene expression, including processes involving BDNF and the immediate-early proteins c-Fos and Egr1/Zif268. CREB phosphorylation at Ser133 is a commonly measured node in this pathway, but it is not a unique signature of magnesium action. The sequence $\mathrm{Ca^{2+}/CaM\rightarrow CaMKII\rightarrow GluA1}$ and the branch $\mathrm{Ca^{2+}\rightarrow AC1/8\rightarrow cAMP\rightarrow PKA\rightarrow CREB}$ describe general plasticity physiology; no one has shown that taking L-threonate activates every link at once in people.

Long-term potentiation (LTP) is a lasting increase in response after a particular stimulation pattern; long-term depression (LTD) is a sustained weakening of a connection. LTD is not simply “erasing old memories”: it contributes to fine-tuning networks and depends, among other things, on phosphatases such as calcineurin. Hippocampal slices make it possible to compare these processes under controlled conditions, but even a larger LTP amplitude does not translate directly into years of “rejuvenation” of human memory. The late phase of LTP depends on transcription, protein synthesis, and spine stabilization, while induction depends on a specific timing pattern, such as theta-burst stimulation. Calcium signals with different amplitudes and durations can shift the balance between kinases and phosphatases, so higher magnesium need not enhance every form of plasticity. The 40–60% LTP increase cited in the original source, without a precise protocol, is not a general clinical fact.

## What human trials actually allow us to say

### MMFS-01: formulation, duration, and measured outcomes

The double-blind randomized study by Liu and colleagues enrolled 44 middle-aged and older adults who reported cognitive decline. MMFS-01 was compared with placebo over 12 weeks using a battery of cognitive tasks. The original protocol description gives a daily dose of 1.5–2.0 g, with a body-weight cutoff of 70 kg, and assessments at weeks six and twelve. MMFS-01 was described as a magnesium L-threonate-based product; without further analysis, the result for the combination cannot be attributed to a single ingredient. The small sample, study duration, and participant selection limit extrapolation to dementia prevention, Alzheimer's disease, or healthy young adults.

The authors reported improvement in some composite measures relative to placebo. The original text separately claims a 22.4% reduction in TMT-B time, a composite Z-score increase to +0.66, and “rejuvenation” by nine years. The last figure is a conversion of test results against an age-normative curve, not reversal of biological aging. The same interpretation mentions a difference from placebo of about 7.3 normative years: this is a derived test scale, not a clinical measure of neuronal recovery. Exact magnitudes depend on the analysis chosen; they cannot be presented as a guaranteed effect of 2 g of the salt or as a change in brain structure. Individual tests of executive function and working memory should be considered alongside the full set of outcomes and confidence intervals, rather than extracting the most striking result. For example, the Trail Making Test Part B measures speed in connecting alternating sequences, but completion time is also influenced by vision, motor function, and familiarity with the procedure. Repeated testing requires control for practice effects; a composite score depends on which tests are included and how they are weighted. The original interpretation “22.4% faster” does not independently establish a change in synaptic density.

A later study of a Magtein®-based formula in healthy adults also expands the human evidence base, but it used a specific multicomponent formula and its own tests. Such studies support further testing of the hypothesis; they do not establish a clinical indication for treating mild cognitive impairment. For scientific interpretation, it is especially important to distinguish the dose of the **compound** from the dose of Mg²⁺, note any other ingredients, and determine whether an outcome was prespecified as primary.

| Evidence level | What was observed | Main limitation |
| --- | --- | --- |
| Cells and tissue slices | Receptor currents, markers, LTP/LTD | Do not reproduce the whole organism or everyday memory |
| Rodents | Behavior, brain magnesium, synaptic structure | Dose, species, and tasks differ from clinical settings |
| Small human trials | Individual cognitive tests over several weeks | No evidence of dementia prevention or sustained synaptic change |

### Fear extinction and sleep: two distinct lines of evidence

Fear extinction is learning that a previous danger cue no longer predicts a threat; the original memory does not necessarily disappear. In animal work by Abumaria and colleagues, higher brain magnesium was studied alongside plasticity in the infralimbic prefrontal cortex and lateral amygdala. Connections from the prefrontal cortex to inhibitory intercalated amygdala cells provide a plausible neural-network framework. Extinction depends on interactions among infralimbic and prelimbic mPFC regions, the lateral amygdala, and inhibitory intercalated neurons; recording plasticity at one site does not capture the whole network. Conditioned-fear behavior in animals is not equivalent to the complex clinical picture of PTSD, which can include re-experiencing, avoidance, and changes in sleep. These findings therefore are not a human trial of treatment for PTSD or generalized anxiety, and a supplement cannot be promised to “switch off” pathological fear.

Sleep matters for memory consolidation: stage N3 is characterized by slow cortical oscillations of roughly 0.5–4 Hz and coordination with other rhythms. Magnesium contributes to neuronal excitability, but the supplied clinical data do not establish a direct allosteric action of L-threonate at a specific GABAₐ-receptor site, obligatory thalamic hyperpolarization, suppression of ACTH and cortisol, or a 15–20% increase in N3 after an evening dose. A study of ordinary magnesium for primary insomnia in older adults involved a different salt and does not prove the same effect for Magtein®. Suppression of nighttime hypothalamic–pituitary–adrenal activity, ACTH, and cortisol remains a hypothesis for this specific intervention. Polysomnography distinguishes N1, N2, N3, and REM using EEG and other signals; the subjective impression of having “slept more deeply” is not a substitute for measuring the proportion of N3. Even a measured change in the percentage of one stage must be considered alongside total sleep duration, awakenings, and daytime function. If a person notices a change in sleep, it should be assessed separately from the marketing promise of “delta sleep without morning grogginess.”


## Dose, combinations, and safety boundaries

### How much magnesium is in a gram of the salt?

For the anhydrous formula, the calculation gives about 82.5 mg of elemental Mg²⁺ per 1,000 mg of L-threonate. This is useful for checking the order of magnitude, but manufacturers should state the actual elemental-magnesium content per serving. The salt's mass and the cation's mass differ by roughly a factor of twelve. The figures below are arithmetic reference points, not treatment steps based on body weight or degree of cognitive complaint.

| Magnesium L-threonate mass | Mg²⁺ at the theoretical 8.25% fraction | How to read the figure |
| --- | ---: | --- |
| 1,000 mg | about 83 mg | Check the product's actual label |
| 1,500 mg | about 124 mg | Do not count 1,500 mg as elemental magnesium |
| 2,000 mg | about 165 mg | Compare with total daily exposure |
| 2,500 mg | about 206 mg | Do not prescribe as a “loading” protocol from one table |

The original text calls 1,500–2,000 mg a standard for different weight categories, and 2,500 mg a short-term regimen after traumatic brain injury. These are not established universal treatment regimens. Even if a similar dose was used in a study, it should not automatically be applied to someone with kidney disease, another source of magnesium in the diet, or concomitant medications. When correcting a confirmed deficiency, assessing its cause, total intake, and safety matters more than choosing a “brain” salt based on a promised advantage.

### Timing and “neuronutraceutical stacks”

A split morning-and-evening dose is sometimes used; it may be convenient for tolerability and adherence. The available evidence is insufficient to assign GluN2B upregulation to the morning dose and a guaranteed increase in N3 to the evening dose. Taking it 60–90 minutes before bed, or splitting 500–1,000 mg during the day and 1,000 mg in the evening, should be understood as a practice described for a particular product, not an established chronopharmacological rule. Sleep and attention problems should first be assessed for their causes and for concomitant medications.

The first combination proposed in the original text brought together L-threonate, phosphatidylserine, DHA, and uridine-5′-monophosphate. These substances have different biological roles: phosphatidylserine is a membrane phospholipid, DHA is a polyunsaturated fatty acid, and uridine contributes to pyrimidine-nucleotide and phospholipid synthesis. This is not enough to claim that taking them together at doses of 200–300 mg phosphatidylserine, 1,000–1,500 mg DHA, and 250–500 mg uridine restores synapses after a neuroinfection. Each ingredient, product quality, and potential interaction requires separate assessment; the mixture cannot be evaluated from the mechanism of one ingredient.

The second combination paired L-threonate with apigenin and L-theanine for sleep. Apigenin is studied in receptor and enzyme systems, including GABAₐ and CD38, while theanine is studied in relation to subjective stress and electrophysiological measures. These separate lines of research do not demonstrate that the combination affects the thalamocortical network, the 8–12 Hz alpha rhythm, or a particular patient's insomnia. The suggested doses of 50 mg apigenin and 200–400 mg theanine are not an evidence-based protocol here. Adding several substances on top of sleeping pills or psychotropic medication without reviewing the full regimen is especially unwise.

### Gastrointestinal effects and special situations

Unabsorbed magnesium retains water in the intestinal lumen, so high doses of some salts can loosen stools, cause cramps, and lead to diarrhea. Risk depends on solubility, the amount of free Mg²⁺, dose, and individual factors; a fixed 400 mL of water should not be attributed to every gram of unabsorbed magnesium. The original text proposed passive diffusion of an intact chelate and transport by organic-acid transport proteins in the proximal intestine. The evidence is insufficient to treat this exact sequence as the established main mechanism of human absorption. L-threonate may also cause gastrointestinal symptoms. A small randomized trial does not establish a universal dyspepsia rate of “under 2%” or justify promising that diarrhea will not occur.

With reduced kidney function, magnesium excretion is impaired, and even a supplement may increase the risk of hypermagnesemia. Intravenous MgSO₄ or MgCl₂ given for other clinical purposes cannot be used to support oral elevation of brain magnesium: the route, rate of increase in plasma concentration, and risks differ. High doses may cause vasodilation, weakness, and reduced tendon reflexes without the desired central effect. Weakness, reduced reflexes, and cardiac rhythm disturbances require medical assessment, especially when multiple magnesium-containing medicines or antacids are combined. Magnesium can reduce absorption of some medicines when taken at the same time, including certain antibiotics and thyroid medicines; timing and whether to use a supplement should be determined for the specific medicine. Evidence is insufficient to recommend L-threonate for cognitive indications during pregnancy or breastfeeding, in children, or after brain injury.

## Comparing forms without a false hierarchy

Comparing magnesium salts is useful only when tied to a question: is the goal to correct a deficiency, change gastrointestinal tolerability, or test a specific research hypothesis about a cognitive outcome? Solubility and elemental-magnesium content are not equivalent to efficacy against insomnia or dementia. The original matrix assigned each compound an exclusive tissue target—malate to muscle, taurate to the heart, citrate to the gut, and L-threonate to the hippocampus. This exclusivity has not been established: after dissociation, the shared ion participates in many systems, and differences among ligands and products require direct comparison.

| Form | Practically relevant feature | What cannot yet be inferred |
| --- | --- | --- |
| L-threonate | Low fraction of elemental Mg²⁺; a line of animal plasticity research and small cognitive trials exists | Proven treatment for PTSD, insomnia, or dementia; exclusive transport across the BBB |
| Bisglycinate | A chelate that some people find easier to tolerate | Guaranteed “peripheral” selectivity or a special sedative effect |
| Malate | Magnesium source with an organic anion | Proven increase in endurance from targeted delivery into the Krebs cycle |
| Taurate | Magnesium source combined with taurate | A confirmed antiarrhythmic indication as a property of the salt as a whole |
| Citrate | Soluble form; larger doses may loosen stools | Complete absence of central effects |
| Oxide | High magnesium mass fraction with different solubility; also used as an antacid | Zero absorption or inevitable “extreme” diarrhea in every person |

For clinicians and researchers, the takeaway is measurement discipline: distinguish the salt from elemental magnesium, cerebrospinal fluid from synapses, and molecular mechanisms from clinical endpoints. For consumers, the simpler rule is that the form of a supplement does not replace diagnosis of the cause of memory, anxiety, or sleep problems. Magnesium L-threonate remains an interesting subject in neurobiology and a possible focus of further clinical trials. So far, its scientific story is much more convincing as a testable hypothesis about synaptic plasticity than as a ready-made universal protocol for “rejuvenating the brain.”

## How this review was prepared

This is a narrative review of magnesium L-threonate chemistry, magnesium transport, and synaptic physiology. It draws on a literature collection assembled earlier that contains 20 publications. Editorial work was completed on October 4, 2026; no systematic search under a prespecified protocol, meta-analysis, or formal risk-of-bias assessment for each study was conducted. The reference list should therefore not be considered exhaustive.

Findings from chemical, cellular, animal, and clinical studies are considered separately. Conclusions about efficacy are limited to the populations, intervention forms, and endpoints studied. A change in magnesium in cerebrospinal fluid or brain tissue is not equivalent to a change in serum, and an observation in animals does not confirm a clinical effect. Formulas and diagrams explain mechanisms; their inclusion does not establish supplement efficacy or transport of the intact complex across the blood–brain barrier.

## FAQ

### Does Magtein cross the blood–brain barrier?

L-threonate was used to raise brain magnesium in animal experiments. However, an established route for transport of the intact complex through a specific BBB transporter has not been identified; blood, cerebrospinal-fluid, and synaptic magnesium levels cannot be treated as interchangeable.

### Has improved memory been proven in people?

Small trials have reported favorable results on some cognitive measures. They do not prove treatment for dementia, reversal of brain aging, or the same effect for every product containing L-threonate.

### What does magnesium do at the NMDA receptor?

At negative membrane potentials, Mg²⁺ limits current through an activated channel. Depolarization weakens the block, allowing a calcium signal that may contribute to plasticity. This describes receptor physiology; it does not establish that a capsule is effective.

### How much elemental magnesium is in 2,000 mg of L-threonate?

The theoretical amount is about 165 mg for the anhydrous formula with an 8.25% magnesium fraction. The actual amount depends on the product and is stated on its label.

### Can it be taken for sleep or PTSD?

The evidence presented does not support L-threonate as a treatment for insomnia or PTSD. Fear-extinction experiments were conducted in animals, and evidence on sleep architecture for this form is limited.

### When is it especially important to discuss a supplement with a clinician?

Individual assessment is important with kidney disease, use of multiple magnesium-containing products, pregnancy or breastfeeding, medicines that may interact, or persistent memory and sleep concerns.

## Sources

- [Slutsky I., Abumaria N., Wu L.J., Huang C., Zhang L., Li B., Zhao X., Govindarajan A., Zhao M.G., Zhuo M., Tonegawa S., Liu G. Enhancement of learning and memory by elevating brain magnesium. Neuron. 2010 Jan 28;65(2):165-177](https://doi.org/10.1016/j.neuron.2009.12.026)
- [Liu G., Weinger J.G., Lu Z.L., Xue F., Sadeghpour S. Efficacy and Safety of MMFS-01, a Synapse Density Enhancer, for Treating Cognitive Impairment in Older Adults: A Randomized, Double-Blind, Placebo-Controlled Trial. J Alzheimers Dis. 2016;49(4):971-990](https://doi.org/10.3233/JAD-150538)
- [Abumaria N et al. Effects of elevation of brain magnesium on fear conditioning, fear extinction, and synaptic plasticity in the infralimbic prefrontal cortex and lateral amygdala. J Neurosci. 2011;31(42):14871–14881](https://doi.org/10.1523/JNEUROSCI.3782-11.2011)
- [Li W., Yu J., Liu Y., Huang X., Abumaria N., Zhu Y., Huang X., Xiong W., Ren C., Liu M.A., Liu G. Elevation of brain magnesium prevents and reverses cognitive deficits and synaptic loss in Alzheimer's disease mouse model. Mol Brain. 2014 Sep 13;7:65](https://doi.org/10.1186/s13041-014-0065-y)
- [Tang Y.P., Shimizu E., Dube G.R., Rampon C., Kerchner G.A., Zhuo M., Liu G., Tsien J.Z. Genetic enhancement of learning and memory in mice. Nature. 1999 Sep 2;401(6748):63-69](https://doi.org/10.1038/43432)
- [Nowak L., Bregestovski P., Ascher P., Herbet A., Prochiantz A. Magnesium gates glutamate-activated channels in mouse central neurones. Nature. 1984 Feb 2;307(5950):462-465](https://doi.org/10.1038/307462a0)
- [Mayer M.L., Westbrook G.L., Guthrie P.B. Voltage-dependent block by Mg2+ of NMDA responses in spinal cord neurones. Nature. 1984 May 17;309(5965):261-263](https://doi.org/10.1038/309261a0)
- [Paoletti P., Bellone C., Zhou Q. NMDA receptor subunit diversity: impact on receptor properties, synaptic plasticity and cognition. Nat Rev Neurosci. 2013 Jun;14(6):383-400](https://doi.org/10.1038/nrn3504)
- [Cull-Candy S., Brickley S., Farrant M. NMDA receptor subunits: diversity, development and disease. Curr Opin Neurobiol. 2001 Jun;11(3):327-335](https://doi.org/10.1016/s0959-4388(00)00215-4)
- [Zhang C., Hu Q., Li S., Dai F., Qian W., Hewlings S., Yan T., Wang Y. A Magtein®, Magnesium L-Threonate, -Based Formula Improves Brain Cognitive Functions in Healthy Chinese Adults. Nutrients. 2022 Dec 8;14(24):5235](https://doi.org/10.3390/nu14245235)
- [Wang J., Liu Y., Zhou L.J., Wu Y., Long C., Liu X.G. Magnesium L-threonate prevents and restores memory deficits associated with neuropathic pain by inhibition of TNF-α. Pain Physician. 2013 Sep-Oct;16(5):E563-E575](https://pubmed.ncbi.nlm.nih.gov/24077195/)
- [Xiong Y., Ruan Y.T., Zhao J., Yang Y.W., Chen L.P., Mai Y.R., Yu Q., Cao Z.P., Shen Y.Q. Magnesium-L-threonate exhibited a neuroprotective effect against oxidative stress damage in HT22 cells and improved cognitive deficiency in aging mice. Food Funct. 2022 Dec 13;13(24):12727-12738](https://doi.org/10.1039/d2fo02543d)
- [Romani A.M. Cellular magnesium homeostasis. Arch Biochem Biophys. 2011 Aug 1;512(1):1-23](https://doi.org/10.1016/j.abb.2011.05.010)
- [Goytain A., Quamme G.A. Identification and characterization of a novel mammalian Mg2+ transporter with channel-like properties, MagT1. BMC Genomics. 2005 May 11;6:48](https://doi.org/10.1186/1471-2164-6-48)
- [Kolisek M., Nestler A., Vormann J., Schweigel-Röntgen M. SLC41A1 is a novel mammalian Mg2+/Na+ exchanger. Am J Physiol Cell Physiol. 2012 Jun 15;302(12):C1779-C1792](https://doi.org/10.1152/ajpcell.00045.2012)
- [Boyle N.B., Lawton C., Dye L. The Effects of Magnesium Supplementation on Subjective Anxiety and Stress—A Systematic Review. Nutrients. 2017 Apr 26;9(5):429](https://doi.org/10.3390/nu9050429)
- [Abbasi B., Kimiagar M., Sadeghniiat K., Shirazi M.M., Hedayati M., Rashidkhani B. The effect of magnesium supplementation on primary insomnia in elderly: A double-blind placebo-controlled clinical trial. J Res Med Sci. 2012 Dec;17(12):1161-1169](https://pubmed.ncbi.nlm.nih.gov/23853635/)
- [Schweigel M., Martens H. Magnesium transport in the gastrointestinal tract. Front Biosci. 2000 Nov 1;5:D666-D677](https://doi.org/10.2741/schweigel)
- [Sun Q., Weinger J.G., Mao F., Liu G. Regulation of structural and functional synapse density by L-threonate through modulation of intraneuronal magnesium concentration. Neuropharmacology. 2016 Sep;108:426-439](https://doi.org/10.1016/j.neuropharm.2016.05.006)
- [Dewhirst RA, Fry SC. The oxidation of dehydroascorbic acid and 2,3-diketogulonate by distinct reactive oxygen species. Biochem J. 2018;475(21):3451–3470](https://doi.org/10.1042/BCJ20180688)