Embedded Files
A Four-Layer Computational Framework for Migraine Stratification, Safety Signal Detection, and Longitudinal Outcome Modeling
December 2025- Current
Introduction
Introduction
Some headaches are not merely headaches. They are an incessant pounding behind the eyes, a nausea that bends the body inward, a sensitivity to light so acute that even a crack beneath the door is bright. They last hours or days and reshape people’s lives.
A close friend of mine has experienced this reality for many years now. She’s cycled through medication after medication, and each with its own cruel signature of side effects. Sometimes the medications don’t even work.
It is a deeply wretched experience witnessing someone you care about endure something you cannot immediately fix. I have wished many times, in vain, that I could take the pain for her. But I'll never claim to know her experience: no one who has not lived it can. Even still, I've seen how much it takes.
This is a profoundly understudied story.
Adolescents with severe migraine fall through the gaps of a research enterprise built largely for adults. Clinical trials recruit older populations. Pharmacovigilance databases smooth over age specific signals. Treatment response becomes a snapshot rather than the rolling film it really is. No existing framework connects the suspicion of a genetic channelopathy to the particularity of a clinical phenotype to the violence of medication side effects to the slow arc of long term outcomes.
I built the Migraine Stratification Outcomes Framework (MiSOF) to address these gaps. It's a unified computational structure for stratifying migraine patients, anticipating treatment responses, and tracing outcomes across time. Further, MiSOF prioritizes a population that existing tools overlook: adolescent females, whose migraines remain poorly characterized and poorly treated.
I dedicate this invention to my friend. Even if what I’ve built comes too late to change her story, I sincerely hope it might change another’s.
Cheers,
Angie X.
MiSOF - Project Series Explanation
MiSOF - Project Series Explanation
The Migraine Stratification Outcomes Framework (MiSOF) operates as a 4-layer pipeline. Each project performs a distinct analytical function and, together, they turn raw data into clinically meaningful insights.
The 4 layers follow an order: ChanVar feeds into TraitStrata, TraitStrata into SigVigil, and then NeuroTrack. Each layer builds on the previous.
ChanVar (Layer 1)
ChanVar handles variable encoding and feature harmonization. It takes heterogeneous clinical data and structures it into analyzable channels. The name reflects a dual focus: clinical variables and the genetic channelopathies suspected in early onset treatment resistant migraine. ChanVar is the foundation since without clean variable representation, downstream stratification cannot succeed.
TraitStrata (Layer 2)
TraitStrata performs phenotypic subgroup identification. It applies clustering methods to clinical and behavioral traits. The goal is to separate patients who look similar on the surface but differ in treatment response, side effect burden, or long term trajectory. This layer addresses a real clinical puzzle: 2 siblings can have the same genetic condition yet experience different outcomes. TraitStrata aims to capture this kind of within-family heterogeneity.
SigVigil (Layer 3)
SigVigil detects adverse effect signals in pharmacovigilance data. It analyzes datasets modeled after FAERS. The focus is not only on whether a side effect occurs but on temporal patterns. When does hair loss appear after starting valproate? How quickly does weight loss escalate on topiramate? Adolescent patients are poorly represented in existing safety databases, so SigVigil prioritizes this exact population.
NeuroTrack (Layer 4)
NeuroTrack models longitudinal treatment response. It integrates stratified phenotypes from TraitStrata with safety signals from SigVigil. The output is a trajectory model: some patients improve and wean off medication while others cycle through multiple drugs with incomplete benefit and accumulating side effects. NeuroTrack asks what distinguishes these paths, and provides an answer that can inform earlier clinical decisions.
MiSOF - A Layman's Guide
MiSOF - A Layman's Guide
You may think of MiSOF as a sorting and prediction system for severe migraine.
The first step organizes patient data into a clean format. The second step finds natural subgroups among patients: i.e. some people might have mostly pain with nausea while others extreme light sensitivity or rare attacks. These subgroups matter because they may respond to different treatments.
The third step looks at medication side effects and asks not just which side effects happen, but when they happen during treatment. This is especially important for teenagers since, biologically, they might react differently than adults. The fourth step puts everything together by modeling how a patient's migraine may change over months or years. Does the person get better on a certain drug? Do side effects force a switch? What does the whole journey look like?
Of course, MiSOF does not treat patients directly. It analyzes public data to find patterns that current research misses with the goal of better stratification today and better treatment guidance tomorrow.
Closing Remarks
Closing Remarks
MiSOF originated from witnessing someone dear navigate a condition that offers no visible markers nor simple resolution. My work here is methodical by design: this is typically what happens when my desire to help meets the limits of what I can actually do.
Adolescent migraine patients are not just downsized adults and thus should not be modeled as such. And holistically, I believe the answers to a person’s suffering should not remain buried in the data.
MiSOF will not produce a cure tomorrow and might not produce one ever. But it does ask the questions that should have been asked years ago, and represents treatment response as a trajectory rather than a single snapshot. It looks for safety signals in populations that pharmacovigilance systems routinely overlook.
I built this for my friend, who might still be waiting for an answer to her affliction. She likely won't ever read this. But if she does:
I hope you know that even if you don’t want to ask for much, everything is still yours. What you carry will never decide what you can become, and the world will always be yours for the taking.
I've always got your back homie 😉
Until the answers arrive,
Angie X.
Page updated
Google Sites
Report abuse