The One-Sentence Version

A blood testosterone level tells you what is circulating in your bloodstream. It does not tell you what your brain is actually exposed to, because the brain runs its own hormone chemistry: it makes its own DHT and its own estradiol locally, and the amount that reaches brain tissue barely tracks the number on your lab report.

1. The Blood-to-Brain Gap, In Numbers

When researchers measure hormones in blood and in cerebrospinal fluid (CSF, the fluid around the brain and spinal cord) from the same people using gold-standard mass spectrometry, the two compartments barely line up.

So the disconnect is real, but it is steroid-specific. Blood is a decent stand-in for DHEA, and a genuinely poor one for testosterone, estradiol and progesterone. The honest takeaway is not "blood is meaningless," it is "blood does not equal brain, and how well it tracks depends on which hormone you are asking about."

2. Where Androgen Receptors Sit, and What They Do

Androgen receptors (AR) are the docking sites testosterone and DHT act through. In the brain they are spread widely, densest in the hypothalamus (the diagonal band of Broca, the mamillary nuclei, the preoptic area, the paraventricular, suprachiasmatic and ventromedial nuclei) and clearly present across the cortex, hippocampus, amygdala and brainstem.

What they govern reads like a menopause symptom list, which is not a coincidence. Documented AR-linked functions in the brain span sexual desire, mood, cognition, thermoregulation (temperature and hot flashes), sleep, visual-spatial skill and language.

3. Why Blood Cannot Equal Brain: The Brain's Own Steroid Factory

The reason the two compartments come apart is that the brain does not passively receive finished hormones from the blood. It manufactures and transforms them on-site. Three mechanisms drive the decoupling.

1. It makes DHT locally (and DHT is the stronger signal)

The brain expresses its own 5α-reductase enzymes (types 1 and 3), which convert testosterone into DHT right in the tissue. DHT binds the androgen receptor with higher affinity than testosterone and lets go more slowly, so local conversion amplifies the androgen signal beyond what testosterone alone would produce. The enzymes sit in region-specific patterns (type 1 concentrated in white-matter and myelin-rich areas, type 3 in the hippocampus and cerebellum).

2. It makes estradiol locally (and lots of it)

The brain also expresses aromatase, the enzyme that converts testosterone into estradiol. Strikingly, the aromatase-carrying neurons sit in the same cells that carry the androgen receptor, so testosterone delivered to those cells becomes both a stronger androgen and an estrogen, locally. Locally synthesized estradiol in the hippocampus runs roughly six times higher than the circulating pool, which is exactly why blood estradiol is a poor proxy for hippocampal estradiol.

3. Protein binding filters what gets in

Only the free (unbound) fraction of a hormone crosses into the low-protein CSF. Since testosterone is about 98% protein-bound in blood, most of the circulating pool never reaches the central compartment as free hormone.

4. What This Means for Women (Menopause, Brain Fog, Mood, Libido)

Testosterone is partly a woman's estrogen supply line

After menopause, when the ovaries stop making estrogen directly, peripheral and local aromatization of androgens becomes the primary source of a woman's estrogen. So testosterone is not just "the libido hormone" for women, post-menopause it is partly the raw material the body and brain use to keep making estrogen. That reframes the whole conversation about testosterone for women.

The APOE4 twist

This is directly relevant to APOE4 carriers. In an ADNI-based study, the relationship between testosterone and cognition in women was APOE-e4-dependent: in female e4 carriers, higher free testosterone tracked better verbal memory (and trended toward better global cognition on the MMSE), while in non-carriers the relationship reversed toward poorer cognition. There was a statistically significant testosterone-by-APOE-e4 interaction.

Brain fog, mood and libido, reframed

If central hormone exposure is governed by local brain metabolism rather than blood levels, then brain fog, mood and libido symptoms are partly downstream of what the brain makes, not only what the blood delivers. That is the mechanistic reason symptoms and lab numbers so often disagree, and a strong argument for treating symptoms rather than titrating to a target blood number.

5. Two Claims We Are NOT Making

Adversarial verification killed two tempting claims. Flagging them so they never end up in a published piece:

6. The Open Questions (Your Best Article Angles)

• What are the actual serum-to-CSF and serum-to-brain-tissue testosterone, DHT and estradiol ratios in women, across the menopause transition and on testosterone therapy? Nearly all direct human quantification to date is male-only.
• Does testosterone therapy in women meaningfully raise brain androgen and estradiol exposure, or does local brain steroid production buffer the CNS so that titrating blood levels poorly predicts the effect on mood, libido and cognition?
• Is the cognitive benefit in APOE-e4 carrier women driven by direct androgen action or by conversion to estrogen? No formal mediation analysis has separated the two, and it matters for APOE4 women weighing testosterone versus estrogen therapy.
• Given that there is "no absolute level that defines deficiency," what markers (if any) actually predict central androgen sufficiency in symptomatic menopausal women, and should treatment be titrated to symptoms rather than blood numbers?

Sources

All peer-reviewed primary studies and clinical reviews. Every claim in this reference was verified by a 3-vote adversarial fact-check (a claim needed to survive skeptics to be included). No blog-grade sources.

Compiled July 2026 from 24 peer-reviewed sources via a fan-out / fact-check / synthesis research pass (23 claims confirmed, 2 refuted and excluded). Research synthesis for personal reference and content development, not a substitute for physician judgment. Full source set and DOIs logged to the central ResearchLibrary.