alzheimer_brain_morphology_mental_health

Alzheimer’s Disease: Brain Morphology, Daily Functioning, and Clinical Severity

A clinically grounded data science project

This project investigates how brain structural metrics, daily functional abilities, and clinical severity indices vary across Alzheimer’s disease diagnostic groups.
It combines rigorous statistical analysis with a clean, modular engineering architecture suitable for both scientific and recruiter-facing audiences.

---

🔍 Project Overview

Alzheimer’s disease is a complex neurodegenerative condition where cognitive decline, functional impairment, and brain atrophy evolve at different rates.
This project explores three core questions:

1. Do brain volumes, daily activities, and symptom severity differ across diagnostic groups?
2. Are brain morphology metrics correlated with clinical severity?
3. Can structural brain measures predict individual symptom severity?

The analysis is grounded in clinical reasoning and supported by a fully reproducible data science pipeline.

---

🧠 Key Findings

1. Strong group-level differences

ANOVA and MANOVA analyses reveal clear differences across diagnostic groups in:

• Daily functioning
• Symptom severity
• Several brain volume measures

2. Weak correlations between brain morphology and severity

Correlations between structural volumes and clinical severity are modest, suggesting that morphology alone does not explain symptom burden.

3. Very low predictive power

Baseline models (Linear Regression, Random Forest) show:

• Low R² values
• Poor generalization
• High error variance

➡️ Brain morphology alone cannot predict individual clinical severity.

This aligns with clinical evidence: Alzheimer’s symptoms emerge from a combination of structural, functional, and neurobiological factors.

---

🧪 Methods & Pipeline

The project follows a fully modular, reproducible workflow:

1. Preprocessing

• Deterministic cleaning
• Numeric conversion
• Missing value handling
• Export of clean datasets

2. Statistical Analysis

• Pearson correlations
• ANOVA & MANOVA
• Tukey HSD post‑hoc tests
• Chi‑square tests for categorical variables

3. Modeling

• Linear Regression
• Random Forest Regressor
• Model comparison and evaluation

4. Visualization

• Group mean plots
• Side‑by‑side bar charts
• Correlation heatmaps
• Diagnostic group comparisons

All figures and tables are generated automatically and stored in: reports/figures/ reports/tables/

---

📁 Repository Structure

alzheimer_brain_morphology_mental_health/ │ ├── notebooks/ # Clean, modular Jupyter notebooks ├── src/ # Reusable analysis modules ├── data/ # Raw and processed datasets ├── reports/ │ ├── figures/ # Generated plots │ └── tables/ # Generated statistical outputs └── README.md # Technical project description

---

📊 Figures & Tables

All visual outputs and statistical tables are available directly in the repository:

• reports/figures → diagnostic plots, group comparisons, volume charts
• reports/tables → ANOVA results, correlations, Tukey tests, model metrics

These are generated automatically when running the notebooks.

---

🎯 Purpose of the Project

This project is designed to:

• Demonstrate clinical data science expertise
• Showcase statistical rigor and interpretability
• Provide a clean engineering architecture aligned with industry standards
• Communicate insights clearly to both technical and clinical audiences

It is part of a broader portfolio focused on clinical analytics, brain health, and evidence‑based modeling.

---

👩‍💻 Author

Patri
Clinical Data Analyst & Data Scientist
Focused on clinical modeling, and reproducible scientific workflows.

---

📬 Contact

For collaboration or inquiries, feel free to reach out via GitHub.

This site is open source. Improve this page.