01 â Upload & Analyze
EEG Signal Analysis
Upload your EEG recording in CSV or EDF format. Our ML pipeline extracts 15+ neurological features and runs your sample through a trained classifier in seconds.
Drop your EEG file here
or click to browse from your device
.CSV
.EDF
.TXT
â File loaded and ready for analysis
File Upload
Signal Processing
Feature Extraction
ML Prediction
PREDICTION RESULT
ADHD DETECTED
Elevated theta/beta ratio and reduced alpha coherence patterns are consistent with ADHD neurological signatures. Confidence: 91.3%. This result is for screening purposes only â consult a licensed clinician.
91%
CONFIDENCE
THETA POWER
2.84 ΞVÂē
BETA POWER
0.91 ΞVÂē
THETA/BETA RATIO
3.12
ALPHA COHERENCE
0.43
SAMPLE ENTROPY
1.76
DELTA POWER
1.12 ΞVÂē
ð Band Power Distribution
ð Entropy Level Trend
ð Feature Insights Panel
THETA/BETA RATIO
3.12
HIGH â
ALPHA POWER
1.21 ΞVÂē
REDUCED â
SAMPLE ENTROPY
1.76
ELEVATED â
SPECTRAL EDGE
18.4 Hz
NORMAL
HJORTH MOBILITY
0.89
ELEVATED â
KURTOSIS
3.41
NORMAL
02 â How It Works
Processing Pipeline
From raw EEG upload to prediction result in under 10 seconds.
EEG Upload
CSV / EDF
â
Preprocessing
Filter + Segment
â
Feature Extraction
15+ EEG Features
â
ML Model
Scikit-learn
â
Result Dashboard
Score + Insights
03 â Technology
Tech Stack
Built with robust, open-source scientific Python libraries and deployable to modern cloud platforms.
ð
Python
Core Language
ðŽ
Scikit-learn
ML Model
ð
Streamlit
Web UI / Dashboard
ðž
Pandas
Data Processing
ð
NumPy
Signal Math
ð
Plotly
Visualizations
ð§Ū
MNE-Python
EEG Processing
âïļ
Render / HF
Cloud Deploy
ð
AWS / GCP
Scale Infra
04 â Security & Privacy
Privacy-First Design
Patient data protection is built into the architecture from day one â not an afterthought.
No Permanent Storage
Uploaded EEG files are processed in-memory and deleted immediately after prediction. Nothing is written to persistent storage.
â ACTIVE
Local Processing Mode
Clinicians can run NeuroSakhi entirely offline on their own machine â no data ever leaves the hospital network.
â ACTIVE
Encrypted Transit
All data in transit is encrypted via TLS 1.3. No raw EEG data is logged or cached by proxies or load balancers.
â ACTIVE
HIPAA Compliance
Roadmap includes full HIPAA compliance: audit logs, role-based access, BAA support for healthcare institutions.
âģ COMING SOON
GDPR Ready
Right-to-erasure, data minimization, and consent management framework planned for international deployment.
âģ COMING SOON
Anonymized Analytics
Only anonymized, aggregated model performance metrics are collected to improve the system. No PII ever.
â ACTIVE