(Version 0.2 — Harmony–Imagery–Logic Edition)
ReMind is an open-source cognitive wellness platform designed to help seniors, individuals with Mild Cognitive Impairment (MCI), and those in the early stages of Alzheimer’s disease reconnect with memory, creativity, and reasoning.
ReMind engages the brain’s three fundamental domains through Harmony (auditory–emotional), Imagery (visual–spatial), and Logic (executive–reasoning) pathways — three universal languages of human cognition that transcend literacy and spoken language.
By merging interactive design (Unity), AI-driven personalization, and data-validated research pipelines, ReMind provides an accessible and scientifically grounded framework for cross-cultural cognitive care.
Conventional cognitive therapies rely heavily on literacy, text, or linguistic understanding, excluding millions of elders in dialect-based or illiterate communities. Alzheimer’s disease currently affects more than 55 million people worldwide, yet accessible early interventions remain scarce.
By reawakening the senses — through sound, image, and reasoning — we can restore the natural harmony of memory, imagination, and logic in the aging brain.
ReMind proposes that stimulating the Harmony, Imagery, and Logic pathways together can strengthen neuroplasticity, delay cognitive decline, and foster emotional resilience.
| Project | Domain | Evidence | Inspiration |
|---|---|---|---|
| Sea Hero Quest | Spatial navigation | Behavioral markers for early Alzheimer’s detection | Data-driven gamification |
| Akili EVO | Attention & multitasking | FDA-cleared digital therapeutic | Clinical validation framework |
| VR Reminiscence Therapy | Emotion & recall | Improves affect and engagement | Emotional comfort design |
| Pathway | Target Function | Brain Regions | Empirical Support |
|---|---|---|---|
| 🎵 Harmony Pathway | Auditory processing, rhythm, emotional synchronization | Temporal lobe, limbic system | J Neurol Sci (2020): rhythm training enhances working memory and mood regulation |
| 🎨 Imagery Pathway | Visual-spatial reasoning, imagination, autobiographical recall | Occipital, parietal & prefrontal cortices | Front Psychol (2022): visual art boosts memory and emotional well-being |
| 🔢 Logic Pathway | Executive control, reasoning, daily decision-making | Frontal & parietal regions | Neuropsychol Rehabil (2021): arithmetic training improves cognitive flexibility |
ReMind follows the multi-domain principles of the FINGER and U.S. POINTER trials, demonstrating that integrated lifestyle and cognitive engagement can yield measurable protection against dementia.
ReMind App (Unity: iOS / Android)
├── Harmony Pathway
├── Imagery Pathway
├── Logic Pathway
├── DataCollector.cs
├── EmotionDetector.cs
└── APIClient.cs
↓
ReMind Cloud (FastAPI / AWS / Render)
├── MongoDB (Behavioral data)
├── /upload → session logs
├── /analyze → ML inference
└── /report → dashboards & summaries
| Layer | Tools | Purpose |
|---|---|---|
| Frontend | Unity (C# / URP) | Interactive cognitive modules |
| Backend | FastAPI + Python | Data collection & analysis |
| Database | MongoDB / Firestore | Session and metric storage |
| AI Services | Gemini / Stable Diffusion / TTS | Generative art, narration, personalization |
| Emotion Detection | MediaPipe / ONNX | Non-intrusive affect recognition |
| Visualization | React Dashboard / Plotly | Behavioral analytics |
- Task type: melody matching, rhythm tapping, tonal memory
- Goal: strengthen auditory attention, rhythm perception, and emotional resonance
- Metrics: timing accuracy, tonal recognition rate, emotional stability index
- Task type: guided drawing, color association, visual storytelling
- Goal: evoke imagination and memory through visual-spatial creation
- Extension: AI-generated “memory paintings” or narrated image stories
- Metrics: drawing duration, color variance, line diversity, emotional valence
- Task type: daily arithmetic, logic puzzles, decision-making tasks
- Goal: train reasoning, planning, and executive control
- Metrics: accuracy rate, response latency, adaptive difficulty trend
[Unity Module] → [DataCollector.cs] → [POST /upload]
↓
FastAPI Receiver → MongoDB
↓
Behavior Analyzer (Python ML)
↓
Visualization Dashboard
{
"user_id": "anon_001",
"session_id": "2025-11-01T18:24:03",
"module": "harmony",
"metrics": {
"accuracy": 0.82,
"response_time_ms": 940,
"emotion_score": 0.73
},
"device": "iPad_Pro_M2"
}Pilot Study (Phase I)
- Participants: 5–10 seniors, ages 60–85
- Duration: 6 weeks
- Endpoints: usability, adherence, and emotional engagement
- Methods: behavioral metrics + facial/voice sentiment analysis
Expanded Study (Phase II)
- Participants: ≥ 50 (community sample)
- Duration: 6–12 months
- Endpoints: MoCA scores, working memory, mood scales
- Analysis: longitudinal trend modeling
Each participant develops an adaptive profile:
{
"harmony_focus": 0.7,
"imagery_expression": 0.85,
"logic_strength": 0.55,
"recommendation": "increase tonal recall sessions"
}- Inputs: rolling average of task performance
- Outputs: personalized pacing, difficulty, and encouragement
- Model: reinforcement learning or Bayesian adjustment
| Domain | Principle | Implementation |
|---|---|---|
| Privacy | GDPR + HIPAA compliance | anonymized IDs, opt-in consent |
| Data Minimization | No raw audio/video storage | derived metrics only |
| Security | AES-256 + HTTPS | encrypted end-to-end communication |
| Transparency | Open data dictionary | accessible documentation |
Formal studies will undergo IRB approval prior to deployment.
ReMind/
├── unity-client/
├── fastapi-backend/
├── dataset-schema/
├── analytics/
├── docs/
└── community/
- Code (Unity C# / Python ML)
- Research (validation and statistics)
- Localization (dialect and accessibility support)
- Clinical Collaboration (IRB partnerships)
| Category | License |
|---|---|
| Code | MIT |
| Data | CC BY-NC 4.0 |
| Documentation | CC BY 4.0 |
| Dimension | Metric | Source |
|---|---|---|
| Engagement | Average sessions per week | App analytics |
| Cognition | Accuracy and latency trend | In-game metrics |
| Emotion | Sentiment stability index | Affect recognition model |
| Usability | SUS / qualitative feedback | Pilot survey |
| Research Impact | Citations, dataset reuse | DOI / GitHub stats |
| Phase | Duration | Focus | Deliverables |
|---|---|---|---|
| Phase 1 | 0–3 mo | Prototype & Pilot Data | ReMind App v0.1 + Report |
| Phase 2 | 3–9 mo | Adaptive AI + Dataset v1 | ML API + Dashboard |
| Phase 3 | 9–18 mo | Longitudinal Study | White Paper v2 + Publication |
| Phase 4 | 18+ mo | Clinical Partnerships | Open Dataset v2 + Trials |
“To re-mind the world — not merely to remember, but to rediscover harmony, imagination, and logic as one.”
ReMind envisions cognitive care that is beautiful, inclusive, and scientifically grounded — uniting art, emotion, and reasoning into a holistic approach to aging and well-being.
- Ng et al., J Neurol Sci, 2020.
- Park et al., Front Psychol, 2022.
- Smith et al., Neuropsychol Rehabil, 2021.
- Ngandu et al., Lancet, 2015.
- Livingston et al., Lancet Commission on Dementia Prevention, 2024.
| Variable | Description | Range |
|---|---|---|
| accuracy | Task correctness | 0–1 |
| response_time_ms | Reaction time | 100–5000 |
| emotion_score | Sentiment valence | -1 to 1 |
| engagement_index | Session frequency norm | 0–1 |
- MCI: Mild Cognitive Impairment
- BPSD: Behavioral and Psychological Symptoms of Dementia
- MoCA: Montreal Cognitive Assessment
- TTS: Text-to-Speech
- Project Lead: Haoda “Daaa” Zhao
- GitHub: github.com/zzhhdaaa/ReMind
- Email: [to be added]
- Community: Discord / Slack – ReMind Open Collective
- License: MIT + CC BY-NC 4.0