They mentioned in this paper that magnesium gluconate in milk was slightly effective at raising CSF Mg levels over 3 weeks. And my statistics suck, but it seems that normal forms of Mg were at least partially effective at increasing short-term object recognition (though not long-term).
In a separate study, the bioavailability (evaluated by absorption, excretion, and retention rate of magnesium) of four commercially available Mg2+ compounds (magnesium-chloride, -citrate, -glycinate, and -gluconate) and two Mg2+ preparations we developed (magnesium-L-threonate, MgT, and magnesium-gluconate in milk) was compared in rats. We found that both MgT and magnesium-gluconate in milk have higher bioavailability (X.Z., F. Mao, Y. Shang, N.A., and G.L., unpublished data).
Other magnesium compounds did not elevate [Mg2+]CSF significantly when compared to control (Figure 1A).
(A) Elevation of magnesium concentration in the cerebrospinal fluid ([Mg2+]CSF) following treatment with different magnesium compounds. Total Mg2+ in CSF was measured before magnesium treatment (day 0), 12, and 24 days after magnesium treatment. Two-way ANOVA analysis revealed significant effect of treatment (F3,69 = 4.76, p = 0.0045, n = 6–8). Data were calculated and presented as a percentage of baseline level.
Rats treated with MgT showed significant enhancement of short-term memory (10 min retention interval, one-way ANOVA analysis, p < 0.05) using a modified NORT (see Figure S4 and Supplemental Experimental Procedures). Rats treated with magnesium-chloride or -citrate displayed enhanced short-term memory as well, but this enhancement was not statistically significant (Figure 1B).
(B) Rat performance during a short-term memory test (10 min retention interval) evaluated by novel object recognition test.
Surprisingly, although magnesium-gluconate in milk has a comparable bioavailability to MgT (X.Z., F. Mao, Y. Shang, N.A., and G.L., unpublished data), it failed to enhance memory (Figure 1B). For the long-term memory test (12 hr retention interval), only MgT-treated rats exhibited enhanced performance (p < 0.05, Figure 1C).
[C] Long-term memory test (12 hr) using novel object recognition test. One-way ANOVA analysis revealed significant effect of treatment on shortterm memory (F4,34 = 2.89, p = 0.037, n = 7–9) and long-term memory (F4,31 = 4.50, p = 0.005, n = 5–10). Post hoc test revealed significant effect of magnesium-L-threonate (MgT) on short term and long-term memory.
Perhaps sustained treatment with Mg gluconate might be partially effective at raising brain levels? Though according to the short- and long-term NORTs, it's probably not nearly as good as MgT. (It's worth noting that gluconic acid
increases aluminum retention in many organs, including the brain)
The question of whether all this would be beneficial in humans is a good one. The authors propose (and confirm by several analyses) that increased synaptic plasticity and activity is due to upregulation of NMDA receptors (via homeostatic response to Mg NMDAR channel blockage, and leading to higher NMDA current during burst activity). Higher brain Mg increased the ratio of phosphorylated CamKII and CREB (but not their total expression), as well as the ratio of NR2B to NR2A/NR1 subunits (these have a positive effect on short- and long-term potentiation, and plasticity). Downstream of this, BDNF was increased by 36% (BDNF is regulated by level of CREB activation). Tests also found increased levels of synaptophysin and synaptobrevin, indicating increased density of presynaptic boutons, and an increase in the number of functional presynaptic release sites. (As Lufega said, this paper is a great read...they really followed the question through pretty deeply)
So, does an upregulation of NMDAR and increased burst current pose any potential problem to humans, in addition to enhancement of cognition? What about a "reconfiguration of synaptic networks from a small number of synapses with high release probability to a larger number of synapses with low release probability?"
