Scientific pitch

A compact case for longitudinal health signal research.

Butterfly Effect studies whether repeated sleep, overnight physiology, movement, context, and daily self reported state can reveal signals relative to baseline that relate to next day burden. The project is built as a research only system for future professional review, useful only when signal, limits, evidence, and return pathway stay visible together.

Read validation Open live surfaces Study workflow
Research only todayThe current version makes no diagnosis, medical advice, or disease risk claim.
Audited engineRoutes carry quality, stability, dependency, confidence, and provenance checks.
Professional reviewOutputs are designed for bounded interpretation, not consumer scoring.

Core story

A governed signal system.

The important message is not that sleep predicts everything. The important message is that repeated night to day patterns can be studied, bounded, validated, and returned for professional review without pretending they are ready for clinical use today.

01

The scientific question

Can changes relative to baseline in sleep timing, continuity, cardiovascular load, movement, context, and daily labels reveal interpretable shifts before the person experiences next day burden?

02

The system answer

The site separates nightly routes, event stress, daily challenge layers, non stress support, participant return, and institutional pilots so each claim keeps its own evidence standard.

03

The practical return

When data quality is sufficient, the output is not a consumer score. It is a longitudinal pattern review with confidence, missingness, provenance, and professional context kept close.

Interpretation engine

What sits under the public surface.

The motor is a layered interpretation stack: collection, feature contracts, quality gates, route specific inference, baseline comparison, route stability, feature dependency audit, confidence scoring, evidence attachment, governance, participant return, professional review, diagnostic readiness, and staged clinical validation planning.

Data

Prospective longitudinal intake

Daily self reports, context, subjective cognitive state, intervention events, and later sleep and physiology linkage are kept as a time-aligned participant record.

Control

Quality and feature contracts

Missingness, timing, signal quality, preprocessing contracts, leakage controls, and route eligibility are checked before any interpretation is trusted.

Inference

Route specific interpretation

The engine does not collapse everything into one model. Nightly, event, daily, non stress, and portable family routes keep their own target, horizon, and evidence standard.

Explanation

Baselines, drivers, and evidence

Outputs carry personal or cohort baselines, local drivers, domain findings, confidence, interval logic, provenance, and linked literature where available.

Audit

Stability and ablation checks

The engine audits whether a route is stable enough to trust and whether its signal comes from sleep, activity, physiology, self report context, or portable shared domains.

Confidence

Research confidence score

Predictions can be accompanied by a confidence score that combines quality, validation strength, history, uncertainty, route stability, and feature dependency.

Review

Participant and professional surfaces

The same structured payload can become a participant research return or a professional companion without changing the underlying claim boundary.

Validation

Diagnostic readiness layer

Internal gates now map signals to phenotype candidates, including neurocognitive, neurological, and rehabilitation contexts, while keeping diagnostic claims blocked in the current phase.

Research pathway

From feasibility to supervised validation.

The pitch should make the development path legible: first prove that the participant workflow and signal quality hold, then move into a small supervised pilot with predefined outcomes and clear ethics.

Prospective daily collection

Daily ratings capture sleep quality, stress, anxiety, mood, fatigue, energy, context, and subjective cognitive state.

Wearable derived linkage

Non invasive wrist worn sensing is treated as consumer grade longitudinal physiology, with timing and quality checks.

Quality first analysis

Missingness, outlier nights, schedule disruption, alcohol, caffeine, illness, and notes are handled before modeling.

N of 1 feasibility readout

The first participant case tests adherence, interpretability, variable stability, and whether signals deserve expansion.

Small supervised pilot

The next phase is a bounded N=10-30 study with academic review, predefined endpoints, and no current diagnostic claims.

Ground truth review

Clinical or professional anchors can be added later to test whether research phenotypes have meaning beyond self report.

Collaboration fit

Specific academic collaboration.

The project is not looking for generic endorsement. It needs methodological pressure, domain supervision, validation standards, and a responsible path toward a pilot that can survive scientific review.

Methodological review

Help define the study contract before scale: variables, endpoints, missing-data rules, participant burden, and what counts as a valid readout.

  • Prospective N of 1 design review.
  • Criteria for moving from one participant to a supervised small cohort.
  • Predefined interpretation limits for participant-facing outputs.

Pitch script

One-minute scientific pitch.

A compact version for researchers, industrial liaison offices, and clinical or data science teams: ambitious enough to be worth a meeting, restrained enough to be credible.

Opening

Butterfly Effect is a research only platform for studying repeated sleep, physiology, movement, context, and daily state as signals relative to baseline for future professional review.

Problem

Most wearable sleep products either stay descriptive or overstate the clinical meaning of a single night. This project focuses on longitudinal patterns and keeps the claim boundary explicit.

Method

The system links night data, daily labels, validation metrics, evidence references, participant reports, professional companions, and internal diagnostic readiness gates through separate layers rather than forcing everything into one model.

Evidence

Some within cohort and event stress routes already show defensible signal. Cross cohort transport remains treated cautiously, which is why the public surface says what holds and what remains research only.

Direction

Prospective studies are in process. Until those readouts are ready, the project keeps public claims attached to validated layers and uses new data to strengthen the pathway toward professional review and later supervised validation.

Claim boundary

Ambitious, but bounded.

This is the line that protects the project when universities, clinicians, and data scientists read it: the page can be ambitious without sounding like a diagnostic product today. A future clinical pathway would require supervised studies, external validation, ethics approval, and regulatory review.

Allowed Research signal language
  • Longitudinal patterns relative to baseline.
  • Sleep, physiology, movement, context, and self report as partial evidence streams.
  • Professional review surfaces with uncertainty attached.
  • Diagnostic readiness language when it clearly means validation planning, not diagnosis.
Reserved Under study
  • Prospective participant readouts until data quality is reviewed.
  • Wearable linkage until timing, coverage, and signal reliability are audited.
  • Small cohort conclusions until a supervised pilot exists.
  • Clinical ground truth and review packets until a protocol defines who can assess them and how.
Current limit Not claimed today
  • No diagnosis, treatment, prevention, or disease risk statement in the current research surface.
  • No cognitive impairment, Alzheimer, or neurological prediction claim before disease specific clinical validation.
  • No replacement for clinical evaluation or medical decision making.