Noumera Labs · AI-native neuroscience
A new era of understanding the mind.
More than 7 million Americans are living with Alzheimer's, and nearly 13 million people care for them without pay. Most of that care happens at home, where medicine rarely sees it.
Noumera Labs is an AI-native neuroscience lab. We are building systems that learn how brain disease unfolds in everyday life, starting with Alzheimer’s, so that families and clinicians can act earlier and better.
The problem
Dementia is a chronic disease that families manage alone, at home, with almost no data.
Alzheimer's is now the sixth leading cause of death in the United States, and the number of older Americans living with it could nearly double, from 7.4 million today to 13.8 million by 2060. Health and long-term care for people with dementia will cost an estimated $409 billion in 2026. On top of that, family and friends gave 19.6 billion hours of unpaid care last year, worth $446.3 billion.
The hardest part is often not memory. Behavioral and psychological symptoms, like agitation, wandering and disrupted sleep, affect up to 90% of people with dementia over the course of the disease, and they are one of the main reasons families move a loved one into a nursing home. These moments happen at 2 a.m. in a kitchen, not in a clinic, so they are barely measured.
-
7.4M
Americans 65+ living with Alzheimer's, 2026
-
19.6B
hours of unpaid dementia care, 2025
-
$409B
health and long-term care cost, 2026
-
59%
of dementia caregivers report high or very high emotional stress
Sources: Alzheimer's Association, 2026 Alzheimer's Disease Facts and Figures.
Our approach
Every hard day holds information. We are building the loop that learns from it.
We start where the disease is actually lived. Memora, our first product, turns what caregivers see each day into structured, consented records, offers vetted non-drug strategies, and checks back to see what worked. Each turn of that loop makes the next suggestion better and the picture of the disease clearer.
Over time this becomes something research has rarely had: a longitudinal, real-world record of how dementia progresses between clinic visits. We pair it with clinical and biological data to model the course of the disease for one person, not an average patient. That is what we mean by AI-native: the learning system is the method, not an add-on.
- Capture. Caregivers describe what happened in their own words, by voice.
- Structure. Each moment becomes a clean, comparable record, with consent and privacy built in.
- Learn. Outcomes flow back into the models, and the models flow back into care.
Our vision
From managing decline to extending the healthy years of the mind.
Medicine has learned to keep hearts beating and blood sugar steady for decades. The brain is the organ it still mostly watches fail. We think that is a solvable gap, and that the way through it is data from real lives, at a scale no clinic can collect.
The evidence already points this way. The 2024 Lancet Commission found that 14 modifiable risk factors account for nearly half of dementia cases worldwide. Much of brain aging is not fixed fate. Alzheimer's is our first problem. Parkinson's and ALS come next, and the long bet is longevity research for the brain itself: understanding why minds age, and keeping them working as long as the body does.
Memora
A copilot for the hardest hours of caregiving.
Memora helps family caregivers handle agitation, wandering and sundowning as they happen, then learns from what works. It is built for the CMS GUIDE dementia care model, which since 2024 has paid care programs to support caregivers, not just patients.
One turn of the loop
Fictional example
-
01Voice note
Tuesday, 6:32 pm 0:14He started pacing around six again and kept asking to go home. I put the old radio station on and he settled after about twenty minutes.
-
02Structured log
From the voice note, in seconds
- When
- Tuesday, around 6:00 pm
- What happened
- Pacing, asking to go home
- Around it
- Late afternoon, light fading
- What was tried
- Familiar radio station
- How it went
- Settled in about 20 minutes
-
03Strategies to try
-
01 Non-drug
Lamps on before dusk
Turn the lights on early, so the room does not dim as evening comes.
-
02 Non-drug
Music a little sooner
Start the familiar station at half past five, before the pacing begins.
-
03 Non-drug
A short walk after lunch
Some movement earlier in the day, together and unhurried.
-
-
04Next-day check-in
Wednesday, 9:00 am
You tried music a little sooner yesterday. Did it help?
- It helped
- A little
- Not this time
Noted. Music moves up for this household.
-
05Weekly report
Hard evenings Week of example data
- Helped most
- Familiar music, started early
- Still hard
- The hour before dinner
What helped goes back into the loop, so the next suggestion fits this family better.
News
Reading that shapes our work.
Research and policy we keep coming back to. Links open the original source.
- 2026 Alzheimer's Disease Facts and Figures
- Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission
- The GUIDE Model: Guiding an Improved Dementia Experience
- Dementia: key facts
- Behavioural and psychological symptoms in dementia and the challenges for family carers: systematic review
Careers
Contact
For research partnerships, care programs and press, write to hello@noumeralabs.com.