🧠 NeuroQuest (MindMeld)

AI-Powered Holistic Neurodegenerative & Health Analysis Dashboard Developed for the Mayoor 6th Inter School STEM Competition 2025-26

NeuroQuest (also known within its ecosystem as MindMeld) is a comprehensive, multi-modal health monitoring and productivity platform. Its primary goal is the early detection of neurodegenerative decline (like Parkinson’s and Alzheimer's) through passive, non-intrusive monitoring of daily computer usage—combining motor control analysis, ocular/mouse tracking, and linguistic evaluation.

Beyond core neuro-monitoring, the platform includes a full suite of healthcare tools including DermaAI (skin condition analysis), a Symptom Checker, a Doctor Command Center, and a Notion-style productivity workspace with integrated Kanban boards.

---

🌟 Core Features & Modules

1. 🧬 Neurodegenerative Tracking (The Core Engine)

By interacting seamlessly with the user's workflow, NeuroQuest gathers vital biomarkers without needing specialized hardware:

• Keystroke Dynamics (Motor Control): Monitors Flight Time (FT) and Dwell Time (DT) to identify bradykinesia and rigidity, early signs of Parkinson's.
• Mouse Dynamics/Telemetry: Calculates a Jitter Score by comparing actual mouse paths to ideal straight lines, surfacing ataxia.
• Linguistic Analysis (Vocabulary Drift): Evaluates Type-Token Ratio (TTR), syntactic complexity, and semantic loops on the fly. Alzheimer’s patients often suffer from "anomia", causing repetitive language and lower TTR.
• Vocal Biomarkers (Audio Triage): Detects micro-tremors (Jitter) and stammering via localized audio analysis utilizing parselmouth.
• Global Biometric Aggregation: Combines metrics (Ghost Score, Tremor Score, Vocab Score) into a unified Neuro Risk prediction index.

2. 🤖 The "Tri-Brid" Intelligence Architecture

A robust routing system that optimizes for both high-level clinical reasoning and offline privacy:

• Cloud AI (Reasoning): Uses Google Gemini 2.0 (google-genai) for complex medical inferences and differential diagnoses.
• Local AI (Offline Privacy): Employs Scikit-Learn (IsolationForest) for entirely local anomaly detection and VADER for sentiment parsing.
• Heuristic Failsafe: Hardcoded biological thresholds ensure baseline functionality even on weak hardware with no internet connection.

3. 🩺 Full Medical Suite Extensibility

• Derma Analyze (DermaAI): Upload close-up images of skin conditions. Uses localized computer vision models for initial condition severity mapping (e.g., Melanoma, Eczema).
• Symptom Assist: Chatbot interface for medical triage based on an intelligent keyword-mapping and LLM diagnostic workflow.
• Doctor Command Center: Role-based access dashboard for clinicians to oversee patients, track telemetry trends, and access the "Evidence Locker".
• Handover Reports: Auto-generates exportable PDF clinical handover reports.

4. 📝 Productivity & Journaling Integration

• Kanban Board & Workspace: A Notion-style productivity environment.
• Journal Analytics: Uses daily journaling to passively measure cognitive verbal fluency, complexity, and sentiment over time.

---

🛠️ Technology Stack

NeuroQuest utilizes a modern, decoupled architecture connecting a high-speed Python daemon with an interactive, glowing dashboard.

Backend / Core Engine

• Framework: Python, FastAPI, Uvicorn, Multithreading
• Machine Learning & NLP: Scikit-Learn (IsolationForest), TextBlob, Textstat, VADER Sentiment, built-in Neural Network simulations.
• Audio Processing: parselmouth
• Integrations: Google Gemini 2.0 API
• Daemon Process: pystray (enables persistent background monitoring tray icons).
• Data Persistence: MongoDB (pymongo) for cloud aggregation & Local SQLite (healthhub.db, workspace.db) for rapid offline lookups.

Frontend

• Languages: HTML5, CSS3, Vanilla JavaScript (ES6+ integration)
• Styling: CSS variables for a deep, modern glassmorphism aesthetic + Tailwind CSS integrations in medical dashboards.
• Data Visualization: Chart.js for rendering complex biometric trendlines.
• Biometric APIs: Browser-based Web Audio API and manual canvas interactions for Ghost/Tremor tasks.

---

⚙️ How It Works (The "Magic")

1. Passive Ingestion: The background daemon (or the frontend via canvas logic) records microscopic input events (keystrokes, mouse vectors).
2. Feature Extraction: Data is run through mathematical models (e.g., calculating Efficiency = Distance / Sum(Lengths)).
3. Anomaly Flagging: The Multi-Threaded Engine fires non-blocking API calls to IsolationForest or Gemini to judge if the metric is out of bounds for the user's rolling baseline.
4. Live Rendering: Event-driven Server-Sent Events (SSE) stream the results directly back to the active user dashboard or doctor console in real time.

---

🚀 Getting Started

Prerequisites

• Python 3.9+
• MongoDB (optional, can fallback to local SQLite features)

Installation

1. Clone the repository

   git clone https://github.com/yourusername/NeuroQuest.git
   cd NeuroQuest

2. Set up a virtual environment

   python -m venv .venv
   # Windows
   .venv\Scripts\activate
   # Mac/Linux
   source .venv/bin/activate

3. Install Dependencies

   pip install -r requirements.txt

4. Environment Variables Create a .env file in the root directory (or configure via the UI Settings).

   GEMINI_API_KEY=your_gemini_api_key_here
   MONGO_URI=mongodb+srv://... (optional)

5. Start the Engine

   cd backend
   uvicorn main:app --reload

6. Launch Frontend Open http://localhost:8000/ in your browser. (The FastAPI backend securely mounts the static files upon startup).

---

📜 Documentation & Licenses

• Technical Breakdown: For an in-depth dive into the internal formulas, biological correlates, and why we chose Isolation Forests over neural networks for local data, please see the Technical & Mathematical Documentation.
• License: Apache License 2.0 (See LICENSE file for details).

"Preserving minds, one keystroke at a time."