Cardiomyopathy100How is cardiomyopathy diagnosed and treated?Property of Prof. Paolo Brenner
All inputs zero. Low or high activations are not a diagnosis.
Inputs · 20 features
Drag, tap or use arrow keys. 0 = absent, 100 = present; intermediate values are abstract (original was on/off).
No matching features or topics.
Outputs · 20 model topics
Network activation 0–1. Not a diagnosis probability, triage level or treatment recommendation.
No matching features or topics.
Model, scope and scale
The demo has 20 inputs and 20 selected outputs (no therapy topics). The other 80 inputs are fixed at zero; the computation equals the original network exactly for that case. Complete weights are not included.
The template only had on/off inputs. Intermediate values from the draggable bars are computed correctly but are not clinically scaled.
Training, privacy and limitations
The template claims 10 million training cases but supplies only weights and five examples. The weights behave as untrained: example cases do not produce matching diagnosis topics. No new training was performed; stock-market data would be unsuitable. Do not enter real patient data; inference is local without storage or transmission.
How is cardiomyopathy actually assessed?
Independent of the model: history and examination, echocardiography, cardiac MRI, ECG and Holter ECG, laboratory and genetic testing, and endomyocardial biopsy where needed (DHZC Charité, German).
Scenario print view
Use Ctrl+P or Cmd+P to print. Records a model state, not a medical finding.