The questions they never asked.
Built for this AI moment in neurological research, when patient experience data can finally be fully collected, connected, and put to use.
The data exists. It lives in you.
- Incomplete data
- Unconnected data
- Widespread misdiagnosis
- Time-limited physician visits
- Antiquated systems
- Missing areas of inquiry
Lived experience. Modern platform. Everyone's future.
If you have MS, pieces of your experience have been captured across many different systems. Your neurologist has documented your relapses. A radiologist has noted your lesions. Your insurance company has logged every prescription filled. Apps on your phone have tracked your steps and your sleep. Researchers may have enrolled you in registries and asked you to complete surveys twice a year. You may have had your ancestry checked by a service.
Every one of those records sits in a different system, owned by a different institution, in a different format. None of them can see the others. Much of it is inaccessible for research. None of it has ever been connected.
What has not been consistently asked of everyone, and recorded alongside the rest, is where you grew up and the places you have spent meaningful time throughout your life. Whether you had mononucleosis, and when. What your life looked like in the years before anyone gave your symptoms a name.
This is not an oversight. It is a structural limitation that no existing MS data source is designed to overcome. WeCureUs is. It exists to collect, organize, and make available the richest possible dataset of lived neurological disease experience, beginning with multiple sclerosis and extending to ALS and Parkinson's disease.
The clinical records. The lived experience. The ancestral context. Connecting these dimensions is what this platform was built to do. Together they form something no existing MS dataset has ever held. And that is our path to the cure.
The AI moment in neurological research.
For decades, the most important data in neurological research has been the hardest to collect. Not because it does not exist. It exists in every appointment, every symptom, every experience that never made it into a clinical record. The ability to collect it systematically simply did not exist.
That has changed. The AI models that all researchers now use can find patterns across datasets of patient experience at a scale and depth that was not possible before. They can surface connections between symptoms, ancestry, environment, and treatment response that no individual study could reach. But only if the data exists in a clean, structured, queryable form. That is the gap this platform was built to close.
This is not a future possibility. The tools are here. The moment is now. What has been missing is the data. WeCureUs builds it.
The questions they never asked are the answers research has been waiting for.
Designed to grow richer over time.
Five named capabilities that no existing MS data source can match.
Built for this moment, not retrofitted
Existing MS datasets were designed for the statistical methods of the decade in which they were built. Structured data with consistent variable definitions, typed response fields, and a stable queryable schema is what AI models are actually designed to train on. Every data point collected here is structured at the point of collection. No preprocessing required from researchers. No normalization. Research-ready from the first data point. That is not a retrofit. It is a design decision made for this moment.
K-anonymity at the disclosure layer
No query returns results for fewer than five participants. The threshold is enforced before any aggregate output reaches a researcher. Privacy is not a posture. It is a property of the platform that makes more precise and intersectional reporting possible because the boundary is mathematically clear.
Data enrichment and added granularity
A response recorded today gains dimensions next year. A participant who reports cognitive fog in 2026 can add onset date in 2027, pattern in 2028, and treatment-change correlation in 2029. Targeted enrichment reaches exactly the participants whose existing records make the new question meaningful. The dataset appreciates in value over time in a way no static registry can match. A radiology report contributed today gains new meaning when ancestry data arrives next year. Cross-type enrichment is something no existing registry was designed to enable.
One Shots and Twofers
When research surfaces a question that did not exist when a module was authored, that question can be asked directly. Thirty seconds, full context preserved, immediately added to the dataset. No new study. No new recruitment. The platform stays nimble in a field where science evolves faster than registry update cycles.
Structured for machines from day one
Most health data requires significant preprocessing before an AI model can use it. Clinical notes, inconsistent registry exports, scanned documents. These are engineering problems before they are research opportunities. Questionnaire responses, radiology reports, ancestry data, and clinical test results all carry a defined type and a consistent schema. Where processing is required, it runs through the same pipeline every time, producing identical, structured output regardless of input source. The dataset is queryable, trainable, and research-ready without an intermediate processing step. This is what AI-centric researchers are looking for and what no existing MS data source was designed to provide.
Your pace. Your terms.
WeCureUs fits around your life, not the other way around. You set your own schedule and your own pace. If you want to be reminded by email or text when it is time to contribute, that option is there. If you prefer to open the platform whenever you feel like it with no reminders at all, that is the default. There is no right way to participate. Every contribution counts. Every answer moves us forward.
Choose a daily or weekly question target. The platform tracks your progress and picks up exactly where you left off.
No schedule needed. Open the platform whenever you feel like it. Every answer is saved and waiting for you.
Optional email or text reminders keep you connected on your own terms. Turn them on, off, or adjust any time.
Two systems. No connection between them.
Individual data and research output are structurally separated by design. Not by policy.
No passwords. Ever.
Access is granted through a verified email link and a confirmed phone number. Two factors that confirm a real person without the vulnerability passwords introduce. This authentication architecture became available at consumer scale only recently. It was not an option when most existing health data platforms were built. Returning participants sign in with a single click. Nothing to remember. Nothing to reset. Nothing to steal.
Two systems. No shared query path.
The database holding individual responses and the system producing research output are structurally separate. No query crosses from one to the other. K-anonymity is enforced mathematically at the output layer before any result reaches a researcher. This is a design constraint, not a policy. It cannot be accidentally bypassed.
Structured data. No preprocessing required.
Every response is typed and stored against a stable question ID in a consistent schema. The data is research-ready and AI-trainable the moment it is collected. No normalization step. No cleaning step. The infrastructure investment that makes AI-scale research possible starts here.
Have you ever been tested for Epstein-Barr virus?
A landmark study confirmed that MS almost never develops without prior EBV infection, yet EBV serology is not part of the standard diagnostic workup and most diagnosed patients have never been tested. Knowing who has and who has not been tested is itself a research question worth answering.
Add your voice to questions like this one.
Have the actual results? You can bring those too.
Answering questions is one way to contribute. But if you have records from your care team, you can contribute those directly. A radiology report from your last MRI. Lab results from an outside lab or your MyChart. Epstein-Barr serology results. A blood panel. Results from an ancestry service. Whatever you have in digital form, this platform is designed to receive it.
You do not need to format it. You do not need to clean it up. Paste the text as it appears in your records. A structured extraction process reads it, classifies it, and matches the findings to a controlled vocabulary, converting it into research-ready data before it enters the dataset.
For most records the entire process takes under a minute. The value it adds to the dataset is lasting.
Every record contribution area is clearly labeled. All of it is entirely optional.
The data exists. Now there is somewhere to put it.
One answer. A thousand patterns.
Every response added becomes part of a larger picture. A single answer about how a mobility device affects daily fatigue joins thousands of similar answers from people with similar profiles. Patterns emerge that no clinic, no registry, and no research study has ever captured, because no one has ever asked these questions at scale before.
The picture grows with every new contribution. When more people answer, every existing answer becomes more meaningful. Patterns sharpen. Comparisons become more precise. The insights reach more people.
Individual specifics are never exposed. What the platform surfaces is always aggregate. How groups of people with similar profiles respond. No individual answer is ever identifiable.
Devices and daily function
Someone with SPMS in a cold climate answers three questions about a cooling vest. That answer joins a dataset that can show how people with secondary progressive MS in colder regions report responding to temperature management, broken down by fatigue level, time since diagnosis, and whether a DMT is active.
Diet and symptom patterns
Someone answers a question about diet and reports reduced fatigue on a Mediterranean-pattern approach. That answer joins a growing record of how people of a specific age and MS subtype report responding to dietary changes. Information no pharmaceutical study has ever been designed to collect.
Geography and exposure history
Someone answers a question about where they grew up and environmental exposures before diagnosis. That answer contributes to a geographic dataset that researchers are beginning to understand may hold keys to MS onset patterns.
Because the data is structured and queryable, the questions it can answer are not fixed in advance. Researchers, participants, and product developers can combine any filters. Subtype, age range, geography, treatment history, symptom pattern, device use. In any combination. As the dataset grows the number of questions it can answer grows with it. A dataset that today shows how people with RRMS in warm climates report managing fatigue will tomorrow be able to show how that picture changes by time since diagnosis, by DMT type, and by reported sleep quality. The questions evolve as the science evolves. The data is ready.
For participants: a personal record that grows richer over time, showing your own patterns before any clinician sees them.
For researchers: a structured, AI-ready dataset of lived experience that no existing registry was designed to provide.
For companies building products for the MS community: real evidence of what people actually use, what helps, and what does not, from the people who know best.
No answer is too small. Every contribution makes the next one more meaningful.
See how it works.
Here is a preview of the portal you will be using once you register. This is a sample question from our Sensation and Neuropathy module, which has not launched yet. The demo is automated. You can watch how the process works without doing anything.
“Despite recent advances in MS diagnosis, the disease is misdiagnosed up to 40% of the time, even at MS centers.”
Dr. Jeffrey Cohen, Director, Experimental Therapeutics Program, Cleveland Clinic Mellen Center for Multiple Sclerosis Treatment and Research. Lancet Neurology, 2023.
This is not a fringe finding. It is a statement made by one of the most senior MS specialists in the United States, published in one of the most respected neurology journals in the world.
Dr. Alise Carlson of the Cleveland Clinic Mellen Center, presenting at the ACTRIMS 2021 Forum, confirmed that approximately 20 to 30 percent of initial MS diagnoses ultimately prove to be misdiagnoses. Dr. Daniel Ontaneda, also of the Mellen Center, calculates that roughly one in five patients referred to an MS specialist with a prior diagnosis does not actually have the disease.
The consequences are not abstract. A landmark multicenter study published in Neurology found that more than half of misdiagnosed patients carried the incorrect diagnosis for at least three years. One-third carried it for ten years or longer. More than five percent carried it for over twenty years. Seventy percent had already received disease-modifying therapy. Thirty-one percent suffered unnecessary harm as a direct result.
Participating MS specialists identified a clear missed opportunity to make an earlier correct diagnosis in 72 percent of patients.
The problem extends beyond individual patient harm into the research itself.
The same multicenter study found that among the 110 confirmed misdiagnosed patients, 14 had received natalizumab, a drug associated with a potentially fatal brain infection. One patient died having been misdiagnosed with MS while the appropriate therapy for the actual condition was withheld.
And four of the 110 had been enrolled as research subjects in clinical trials for experimental MS therapies, testing unproven drugs for a disease they did not have.
A 2024 systematic review and meta-analysis examining 3,910 studies found that MS misdiagnosis frequency across the research literature ranges from 5 to 41 percent depending on the setting. Gaitán and Correale, writing in Frontiers in Neurology in 2019, independently confirmed the pattern first documented in the multicenter study: misdiagnosed patients enrolled in clinical trials compromise the accuracy of results while simultaneously being exposed to experimental treatments for a disease they do not have.
This is the version of the problem that the research community rarely states directly but that the data makes unavoidable: clinical trials designed to study MS are enrolling patients who do not have MS.
The trials are designed to understand a disease. When the study population includes people who do not have that disease, the findings describe something that does not exist. Decades of research effort and billions of research dollars depend on the accuracy of the diagnostic foundation underneath them. The foundation is not solid.
This is what happens when the foundation under decades of research is data that was never designed to be connected.
This is a data problem. It is solvable. And it is exactly the kind of problem this platform was built to address.
5,000 Years of Unanswered Questions
The findings are striking.
In January 2024, researchers from the University of Cambridge, the University of Copenhagen, and the University of Oxford published a study in Nature that traced the genetic origins of MS risk across 5,000 years of human history.
By analyzing DNA from ancient human bones and teeth recovered across Eurasia, the team tracked a major migration of livestock herders known as the Yamnaya people from the Pontic Steppe, a region spanning present-day Ukraine, Russia, and Kazakhstan, into northwestern Europe approximately 5,000 years ago.
The genetic variants they carried, now known to significantly increase MS risk, appear to have provided a survival advantage at the time, most likely by strengthening immune response to infections from domesticated animals. In the modern environment, those same variants increase autoimmune risk.
“What we found surprised everyone.”
The distribution of Yamnaya ancestry in modern populations tracks the north-south MS prevalence gradient across Europe with striking precision. The researchers wrote that the steppe ancestry gradient in modern populations across the continent “may cause this phenomenon.”
The signal crosses continental boundaries. African American individuals with MS carry significantly more European steppe ancestry in the relevant HLA region than African Americans without MS. Asian American individuals with MS carry less European ancestry in that region than Asian American controls. The ancestry pattern predicts MS risk across racial lines.
The absence is equally revealing. MS is virtually absent in populations with no Yamnaya genetic heritage: Inuit, Australian Aboriginals, South African Bantu, indigenous North and South Americans, Samis of northern Scandinavia, Yakuts and other Siberian tribes, and Canadian Hutterites. These findings span multiple continents and represent independent confirmation of the same pattern.
Source: Barrie W. et al., Nature, January 2024. DOI: 10.1038/s41586-023-06618-z
Have you ever been asked?
Do you know where your ancestors came from? Do you have any sense of your genetic heritage? Have you ever been tested for any of the genetic markers now known to be associated with MS risk?
These questions are not part of a standard neurologist visit. They are not recorded consistently. They cannot be reconstructed from any existing medical record. They are not collected at scale anywhere.
And yet the research producing some of the most significant findings in MS history is built entirely on exactly this kind of information: who people are, where they come from, and what their ancestry carries.
That infrastructure does not need to be invented. It needs to be built.
This platform is built for exactly this moment. It is not limited to pre-configured answer choices or open text fields. It is designed to accept complex inputs, including genealogical records, ancestry reports, and genetic heritage information, and process them through an AI layer into structured, queryable data. That structured output sits alongside every other contribution: symptom history, treatment record, diagnosis timeline, geographic background.
If you have used an ancestry service, you can paste your results here directly. No formatting required. The platform converts the text into structured data ready for research.
This means a researcher can ask, for the first time at scale: among participants who report Scandinavian or Eastern European ancestry, how does fatigue presentation differ from those with Mediterranean or East Asian heritage? Among those who have had genetic ancestry testing, what patterns emerge across the HLA variants the Yamnaya carried?
These are not hypothetical future questions. They are questions this platform is designed to answer as the data grows.
The Yamnaya carried these genes across a continent 5,000 years ago. The science that traced them has existed for barely a year. The infrastructure to ask what comes next is being built now.
New meets old. And the answers have never been closer.
From Anecdote to Evidence.
A single data point is an observation. A thousand unconnected data points are a thousand separate observations. They describe individual moments but cannot reveal patterns, confirm hypotheses, or answer the questions that matter most to people living with MS and the researchers trying to understand it.
The history of MS research is partly a history of data that could not speak to other data. MRI findings that could not be connected to symptom histories. Genealogical information that was never collected. Radiological reports that existed in one system while treatment records lived in another. Patient accounts that captured what no structured form was designed to hold, and were lost entirely because no system existed to receive them.
This platform is that system. Radiology reports, ancestry data, lab results, and the lived accounts of daily experience with MS all have somewhere to go now.
Individual information that cannot be connected remains anecdote. An MRI report that cannot be linked to ancestry data cannot help researchers understand why people with certain genetic backgrounds develop different lesion patterns. A family history that cannot be cross-referenced with treatment response data cannot help clinicians understand why one person responds to a therapy that fails another. The information exists. The connection does not.
What this platform is designed to do is different.
It is not a survey. It is not a registry. It is not a symptom tracker. It is a platform designed to accept the full range of what it means to have MS, and to connect it.
Structured module responses across symptom domains, treatment history, and diagnostic experience. Open text accounts that capture what no checkbox can contain. Radiological reports processed through an AI layer into structured, queryable data. Genealogical and ancestry information, including DNA ancestry reports, converted into data that sits alongside everything else and can be queried in combination with it.
This is what the AI moment makes possible. Not someday. Now. The infrastructure to ingest complex, heterogeneous data (genealogical, radiological, qualitative, structured), extract meaning from it, and make it queryable at population scale exists today. It has never been applied to patient-reported MS data before.
The Yamnaya genes that traveled across a continent 5,000 years ago can now be connected to the MRI findings, the fatigue patterns, the treatment histories, and the lived accounts of the people who carry them today. Not as isolated observations. As one dataset.
New meets old. And for the first time, the full picture is within reach.
Or explore what the research shows before you decide.
Your participation shapes what comes next.
The platform does not set its own agenda. The people who use it do. Every experience shared, every gap named, every suggestion submitted moves the platform toward territory it could not reach without the community behind it. The questions evolve because the people who know what is missing are the ones who shape what gets asked next.
This happens in two ways. The first is through the platform itself. When patterns emerge across contributions, when many people with similar experiences point to the same gap, that gap becomes a candidate for a new question or a new module. The second is direct. The portal includes a space where participants can name what is missing, suggest what should be asked, and flag areas that need more attention. Both paths lead to the same place: a platform that gets better at asking the right questions the more people use it.
The next question this platform asks may exist because of what you share today.