✦ AI-Powered Early Detection

Catch it early.
Just by listening.

NeuroSpeak turns a voice recording into an early warning system for ALS, Alzheimer's, and Parkinson's: built to flag risk years before it's usually caught.

See how it works Partner with us

Non-invasive

Just your voice, nothing else

Accessible

Screen from anywhere

3 diseases

ALS, Alzheimer's & Parkinson's screened together

A peek inside the app's additional pages

Why NeuroSpeak

A warning sign hiding in your voice

Every year, millions of families wait months, and sometimes years, for a diagnosis that could have come sooner. Neurodegenerative diseases like ALS, Alzheimer's, and Parkinson's affect 58 million people worldwide, and that number is climbing fast as the population ages. The earlier these conditions are caught, the more options stay on the table and the better the outcome.

The real problem is a lack of access. Today's diagnostic tools require expensive scans, invasive procedures, and specialist appointments that can take years to schedule. Meanwhile, one of the earliest warning signs is hiding in plain sight: the voice. NeuroSpeak listens for it.

✓ No needles - just a short voice recording
✓ Trained on thousands of real patient recordings
✓ Delivers a result from any device, no lab required
✓ Built to support doctors, not replace them

The Problem

Diagnosis today is slow, costly, and often too late

Neurodegenerative disease is one of the most underdiagnosed categories in medicine - dismissed as normal aging, gated by cost, or delayed by long waits for a specialist. By the time a diagnosis is confirmed, meaningful damage has often already been done.

58M+

people living with a neurodegenerative disease worldwide - and rising

$3,000–$5,000

average cost of a full diagnostic workup (MRI, PET, or spinal fluid analysis)

One disease at a time

most screening tools test for a single condition, even though symptoms often overlap

Alzheimer's

3.5 years to diagnose

Parkinson's

2.75 years to diagnose

ALS

1.4 years to diagnose

Diagnosis is hardest early on, and often only comes after significant damage has already occurred.
Delays can mean missed treatment windows, when timing matters most for outcomes.
Gold-standard tests require specialized equipment many clinics simply don't have.
Symptoms are frequently written off as "normal aging" until it's too late to act early.

How It Works

Three simple steps

NeuroSpeak was designed to feel as simple as recording a voice memo.

1. Listen

You speak into the app. That's it - no special equipment.

2. Analyze

The AI, trained on thousands of real patient recordings, looks for the subtle vocal patterns linked to ALS, Alzheimer's, and Parkinson's.

3. Flag Early

You get a clear, easy-to-read result - built to flag risk early enough that you can act on it.

Built and tested on tens of thousands of real speech samples from three respected medical research datasets covering Alzheimer's, Parkinson's, and ALS.

Proven Approach

Grounded in real research, not guesswork

NeuroSpeak's model was benchmarked against rigorous statistical testing to make sure the results hold up.

Multi-disease

One screening, several conditions

Data-driven

Built on real patient recordings

Statistically tested

Validated through real testing

Clinician-minded

Designed to support real care

How we compare

Most existing research tools are built to test for a single disease at a time. NeuroSpeak was designed to screen for multiple conditions in one pass, and it specifically tackles one of the hardest problems in the field: telling early ALS apart from healthy speech, a distinction that trips up most other systems.

What's next

More data: expanding to more languages, accents, and communities so the tool works for everyone.
A fuller picture: pairing voice analysis with facial expression and movement tracking.
Long-term tracking: a mobile app that watches for small changes over months and years.
Built for doctors: designed to support clinical decisions, not replace them.

Like any early-stage medical AI, NeuroSpeak isn't perfect yet. The team is actively working to expand the training data, improve consistency across recording conditions, and make results easier for clinicians to trust and act on. Additionally, it is key to note that this is meant to be an accessible screening tool, not a replacement for a professional diagnosis.

Founder

Built by a student researcher

Ilina Rao

Founder, NeuroSpeak · 12th Grade, Interlake High School

Ilina is the founder of NeuroSpeak and a senior at Interlake High School, recognized nationally for applying AI to healthcare access. Her research placed first at the Washington State Science and Engineering Fair, and she's also a Technovation Global semifinalist, a Toshiba ExploraVision top-10% finisher, and an IEEE presenter. She started NeuroSpeak because she believes early detection should be available to everyone, everywhere.

WSSEF - First Place Technovation Global Semifinalist Toshiba ExploraVision - Top 10% IEEE Presenter

Catch it early.Just by listening.