@davidjonesbrain.bsky.social
Behavioral neurologist and AI director working to end degeneration of mental function and use technology to transform the practice of neurology. Neuro AI Program (NAIP) https://naip.mayo.edu https://www.youtube.com/channel/UCj9q9jnuNZNNZnx48-P1vkg
Focusing on AD only, out of context of all degenerative disease of intelligence is one of the major errors IMO. Intelligence is energetically expensive and FDG-PET provides a landscape for all dementias, and connects to the clinic to computational neuroscience: bsky.app/profile/davi...
03.07.2025 19:20 — 👍 0 🔁 0 💬 0 📌 0newsnetwork.mayoclinic.org/discussion/h...
01.07.2025 02:26 — 👍 0 🔁 0 💬 0 📌 018/
If we want to catch reversible cases, personalize care, and speed up new treatments — we need smarter tools.
Tools that understand dementia with nuance.
We built one. And it’s working.
🧠🔍
#Endalz #NeuroAI #HealthTech #FutureOfCare
@mayoclinic.org
17/
It is grounded in the Global Functional State Space (GFSS) — a neuroscience-informed model of brain function and dysfunction.That is why we call it StateViewer.It connects clinical care to systems neuroscience, paving the way for new science and new cures. #NeuroAI www.nature.com/articles/s41...
16/
StateViewer isn’t just a model.
It’s part of a shift:
🧠 From opaque to interpretable
🧠 From limited access to scalable tools
🧠 From "maybe it’s dementia" to "here’s what kind and what’s next"
Clinical AI built by neurologists, for patients.
#NeuroTech
15/
From our
@GreenJournal
paper (July 2025):
📈 Sensitivity = 0.89
📊 AUROC = 0.93
✅ 9 syndromes detected
📉 Reader errors reduced 3.3×
🔬 External validation in ADNI
🧠 Interpretable and scalable
#Neurology #AI
14/
This also matters for clinical trials.
🧬 Disease mechanisms don’t map 1:1 to syndromes
🎯 Precision targeting requires deeper models
⚡️ Cleaner cohorts = faster progress
StateViewer captures this complexity and makes it actionable.
#Alzheimers #ClinicalTrials
13/
It’s been trained on 3,600+ real patient cases.
Externally validated in the ADNI dataset.
And it doesn’t just give answers — it explains them.
It shows how a new case relates to past ones.
Transparent, interpretable, clinical-grade AI.
#ExplainableAI
12/
In a blinded study, clinicians using StateViewer were 3.3× more likely to make the correct diagnosis than those using today’s standard tools.
Even the most experienced doctors improved with it.
#DigitalNeurology #AIinHealthcare
11/
Here’s what makes StateViewer different:
✅ Needs only 1 input: a brain FDG-PET scan
✅Returns multiple diagnoses with statistics
✅Visual maps to explain “why”
✅Supports (not replaces) the physician
It’s made for real-world use and it’s already used in the real-world.
10/
It’s called StateViewer.
A clinical AI tool that reads brain FDG-PET scans and recognizes patterns of 14 different dementia syndromes — like a specialist would. (only nine were published)
But it’s built to augment, not replace, the clinician.
#CDSS #NeuroAI #DementiaTech
9/
That’s why we’re building scalable solutions.
So that even where no specialist is present, expertise is.
Inside Mayo Clinic, we built a platform to bring precision diagnosis anywhere.
👉 siliconangle.com/2025/04/16/i...
#AI
8/
But here’s the problem:
There are too many people with dementia symptoms...
And not enough specialists.
🧑⚕️ Cognitive neurology is scarce
🌎 Rural access is limited
📈 Dementia is rising fast
We need to scale expert-level decision-making.
#healthequity
7/
There’s also a human cost.
🧭 There is peace in knowing.
Once a diagnosis is clear, people can focus on what comes next — not more tests, second opinions, or doubt.
That peace matters.
#NeuroEthics #PatientCare
6/
Misdiagnosis doesn’t just cause harm.
It wastes years.
⏳ Endless referrals
💰Financial cost
😔Uncertainty and grief
Families deserve clarity.
Trials need precision.
Without a clear diagnosis, progress stalls.
#Alzheimers #DementiaDiagnosis
5/
Not all “Alzheimer’s” is the same.
Some patients have language-predominant, visual, or dysexecutive forms — each with different symptoms, prognosis, and response to therapy.
Others seem like AD, but aren’t. (We call these cases LANS.)
Precision matters.
#AD
@alzassociation.bsky.social
4/
Take DLB (Dementia with Lewy Bodies). #DLB
DLB ≠ AD ≠ Parkinson’s ≠ PSP ≠ NPH ≠ MSA. But it’s easy to mistake.
A proper diagnosis changes everything. Don’t know how to do that?
Read this:
link.springer.com/article/10.1...
3/
This isn’t rare. I see cases like this every week.
🧠 Dementia symptoms are complex
🔍Many causes overlap
💊Some are treatable
📋All require different approaches
If you don’t know the cause, you can’t manage it properly.
#Neurology #DementiaCare
2/
One of the most striking examples?
🧠 Normal Pressure Hydrocephalus (NPH)
– Gait trouble
– Cognitive slowing
– Incontinence
It mimics other dementias. But it’s reversible.
If you miss the diagnosis, you miss the chance to help. #NPH
continuum.aan.com/doi/10.1212/...
1/
Every week in my neurology clinic, I see people lost in the search for a diagnosis.
Sometimes, what’s causing their dementia symptoms is treatable — but was missed for years.
We just published a new AI tool to change that. Here's why it matters🧵
@mayoclinic.org
#ENDAlz
A new AI tool can identify nine types of dementia from a single FDG-PET brain scan, supporting faster and more accurate diagnosis, even in clinics without neurology specialists. doi.org/g9rnhp
28.06.2025 04:32 — 👍 4 🔁 1 💬 1 📌 0Me too! First glimpse of the DMN came from glucose before you gave us fMRI for that :)
28.06.2025 01:43 — 👍 2 🔁 0 💬 0 📌 0
"The work we reviewed here supports an alternative view that the manifold and circuit approaches to cognition are inseparable." Can I get an amen? 😊
www.nature.com/articles/s41...
🚨 Video alert: I was able to share our @mayoclinic.org work on #NeuroAI at #GoogleCloudNext, including how we hope to scale expert interpretation of brain scans to improve dementia care.
🧠 Watch here: youtu.be/lJ1SVeeFWZQ?...
#DigitalHealth #AI #endalz
Discovery funded by 3 NIH grants:
22.03.2025 20:34 — 👍 47 🔁 5 💬 0 📌 0
Great working with Sujala Ghatamaneni, Val Lowe, and colleagues on this high resolution autoradiography study of MK-6240 and flortaucipir🧠🖥️
link.springer.com/article/10.1...
@drchristinamoloney.bsky.social @davidjonesbrain.bsky.social
