Clinical Decision Support System
AI-Powered Early Alzheimer's Detection Using Multi-Modal Biomarker Integration
System Status
Online
Model Accuracy
75.51%
Avg Processing
4.3min
Total Patients
1270
Tiered Diagnostic Approach
1
Stage 1: Primary Screening
Clinical Assessment (MMSE, CDR)
Digital Handwriting Analysis
SVM-based Risk Stratification
1270 patients screened
Low-cost, rapid assessment for all patients
2
Stage 2: Advanced Analysis
Structural MRI (CNN Analysis)
Blood Biomarkers (Random Forest)
Ensemble Meta-learner
635 patients analyzed
Only moderate-to-high risk cases from Stage 1
Multi-Modal AI Performance
MRI CNN Model
Accuracy: 75.47%
Macro F1: 78.95%
Architecture: ResNet50
Genomic XGBoost
Accuracy: 100.00%
Macro F1: 100.00%
Features: 1000+ variants
Ensemble Model
Accuracy: 75.51%
Macro F1: 80.26%
+0.04% vs MRI
Key Research Findings:
- • Ensemble achieves 75.51% accuracy, outperforming MRI CNN (75.47%)
- • Macro F1 improvement of 1.31% shows better class balance
- • Multi-modal approach successfully combines complementary information sources
- • Meta-learner demonstrates superior performance across all evaluated metrics
Recent Cases
P00130m ago
Demographics:
F, 72y
Status:
Complete
Stage 1 Risk:
65%
Diagnosis:
Mild AD
Stage 2 Risk:78%
P00245m ago
Demographics:
M, 68y
Status:
Stage 1
Stage 1 Risk:
25%
Diagnosis:
Normal
P0031h ago
Demographics:
F, 75y
Status:
Complete
Stage 1 Risk:
82%
Diagnosis:
Moderate AD
Stage 2 Risk:89%
| Patient ID | Demographics | Stage 1 Risk | Stage 2 Risk | Diagnosis | Status | Time |
|---|---|---|---|---|---|---|
| P001 | F, 72y | 65% | 78% | Mild AD | Complete | 30m ago |
| P002 | M, 68y | 25% | N/A | Normal | Stage 1 | 45m ago |
| P003 | F, 75y | 82% | 89% | Moderate AD | Complete | 1h ago |
Live Dataset Preview
MRI Brain Scans Dataset - All Categories
Normal Cognition
nonDem2540.jpg
nonDem2541.jpg
00a98422...87ee.jpg
Healthy brain structure
Very Mild (MCI)
verymildDem1776.jpg
verymildDem1777.jpg
000a074f...7018.jpg
Early cognitive decline
Mild Dementia
mildDem701.jpg
mildDem702.jpg
000cdcc4...b564.jpg
Moderate atrophy
Moderate Dementia
moderateDem36.jpg
moderateDem37.jpg
00a4080b...b5f5.jpg
Severe tissue loss
Real MRI scans showing progressive structural brain changes across all dementia stages
Handwriting Analysis Dataset - Multiple Tasks
TASK_02: Connect Dots
Healthy Control
T02_HC_001
Alzheimer's
T02_AD_001
TASK_03: Vertical Lines
Healthy Control
T03_HC_001
Alzheimer's
T03_AD_001
TASK_05: Draw Spirals
Healthy Control
T05_HC_001
Alzheimer's
T05_AD_001
TASK_04: Trace Circle
Healthy Control
T04_HC_001
Alzheimer's
T04_AD_001
TASK_21: Complex Drawings
Healthy Control
T21_HC_001
Alzheimer's
T21_AD_001
TASK_24: Clock Drawing
Healthy Control
T24_HC_001
Alzheimer's
T24_AD_001
Real handwriting samples across 6 different tasks showing motor control and cognitive differences between healthy controls and Alzheimer's patients
Dataset Statistics
MRI Dataset
- • Total Images: 6,400+ brain scans
- • Normal Cognition: 3,200 cases
- • Very Mild Dementia: 2,240 cases
- • Mild Dementia: 896 cases
- • Moderate Dementia: 64 cases
Handwriting Dataset
- • Total Samples: 1,800+ handwriting tasks
- • Healthy Controls (HC): 900 samples
- • Alzheimer's Disease (AD): 900 samples
- • Task Types: 6 different cognitive tasks
- • Features: Velocity, pressure, tremor analysis