Precision brain health

Know your brain's normal.
Learn what keeps it healthy.

Sanovu joins continuous at-home measurement, blood biomarkers and AI pathway models in one loop that learns what works for each person.

The problem

Brain care is the least precise part of medicine.

Once

Measured late, and once

Memory is usually first checked with a single clinic test, years after the biology began to change. One test is a snapshot, and snapshots mislead.

Same

The same plan for everyone

Advice and treatment are one size fits all, and almost nobody checks whether they are working for that person.

Half

Much of the risk is changeable

Nearly half of serious memory disorders are linked to 14 changeable risks, such as hearing, blood pressure, inactivity and isolation (Lancet Commission, 2024).

Our approach

One loop for each person.

The tools to act early now exist: blood tests for brain proteins, phones and wearables that measure every week, and AI that can model biology. We join them.

1

Measure

Short, enjoyable games and a validated memory test at home, with sleep and activity from wearables. We learn each person's normal and see real change.

2

Read the biology

Blood markers of brain health and inflammation, aiming for simple at-home finger-prick collection.

3

Model

AI-built pathway models link dose, blood level, brain level and target activity, simulated across thousands of virtual people.

4

Act

Personal lifestyle plans tested as small experiments now; medicines only inside physician-supervised research studies.

5

Learn

Every result updates the person's model and the shared one, so the next choice is better.

Our measurement and coaching tools support general wellness and do not diagnose. Any medicine is studied only in physician-supervised, ethics-board-approved research.

Science · Pathway engine

Simulate before you treat.

Our first engine models GSK-3β, a brain enzyme linked to tau, the protein that tangles in aging brains. Lithium, a decades-old generic medicine, acts on it.

  • A chain from dose to blood level, brain level and enzyme activity, with uncertain parameters scanned rather than assumed.
  • Calibrated against blood levels measured in human studies; tens of thousands of virtual participants to check safety and rank dosing strategies.
  • Chemistry and molecular-dynamics checks of how lithium behaves in the body.

The engine suggests that a recent low-dose study reached blood levels that only weakly engaged the target, so they could not fairly test the idea. Choosing the right dose for the right person is a modeling problem, and that is what we build.

Model estimate of GSK-3β activity blocked as serum lithium level rises, with uncertainty band
Model estimate of GSK-3β activity blocked versus blood lithium level. Line: median; band: 10th–90th percentile across plausible biology. Research model; not medical advice.
Built

GSK-3β

Dose-to-target model for lithium carbonate, shaping a planned clinical study with academic partners.

Next

NF-κB inflammation

A shared inflammation hub that genes, statins and omega-3 status all touch, and that can be read in blood.

Then

Parkinson's

The same models and the same loop, applied to alpha-synuclein biology with continuous measures of movement, sleep and thinking.

Evidence and products

Built on a decade of measurement.

1M+

MemTrax memory tests taken, with 23 peer-reviewed publications. Licensed to Sanovu.

~50k

Brain Health Registry participants tested repeatedly over four years, now being analyzed with UCSF data.

0.59→0.81

Test reliability when three at-home sessions are averaged instead of one. That is why we measure continuously.

2

Free games live in the App Store, with no ads: Charmed Lives and Treasure in the Fog, a spatial-navigation game.

Work with us

Clinics, researchers, health plans and investors.

If you run a prevention practice, study brain health, or want to bring precision to how we care for the brain, we would like to hear from you.

hello@sanovu.ai