EEG Alzheimer's Analysis

CNN-BiLSTM
83.8% Acc
0.95 AUC
OpenNeuro
AWS Lambda
Drop Clinical EEG
Supports .set (19-Channel)
Primary Classification
--
Alzheimer's 0%
FTD (Frontotemporal) 0%
Healthy Control 0%
Analysis Insights

Detailed neural pattern analysis will appear here...

Methodology & Processing

Signals undergo a 0.5-45 Hz band-pass filter and are segmented into 2-second epochs (1s overlap). Welch's method computes the PSD to extract Relative Band Power across 5 frequency bands (Delta, Theta, Alpha, Beta, Gamma) over 19 channels.

Model Performance

Trained on the OpenNeuro dataset, the CNN-BiLSTM architecture achieves a Test Accuracy of 83.81%, a Cohen's Kappa of 0.7520, and an average AUC of 0.9546, ensuring robust diagnostic separation across classes.

DISCLAIMER: This tool is for research and demonstration purposes only. It does not provide medical diagnosis. Always consult a qualified healthcare professional for neurological assessments.