A software sensor using neural networks for detection of patient workload

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Pacemaker function monitoring is crucial for patient management.
• Assessing patient workload and body posture non-invasively is challenging.
• Intracardiac electrograms (IEGMs) offer detailed cardiac electrical information.

Purpose of the Study:

• To investigate the potential of IEGM morphology for estimating pacemaker patient workload.
• To explore the correlation between IEGM morphology and body posture.
• To evaluate the utility of neural networks in classifying IEGMs based on morphology.

Main Methods:

• IEGMs were recorded from pacemaker patients during exercise and at rest in various postures.
• Morphological analysis of IEGMs was performed visually and via computer-simulated neural networks.
• A neural network was employed as an automatic IEGM classifier based on waveform characteristics.

Main Results:

• IEGM morphology demonstrated significant changes related to patient workload and body posture.
• The ST segment of the IEGM was identified as a key area containing relevant diagnostic information.
• Neural networks proved effective in automatically classifying IEGMs based on their morphology.

Conclusions:

• IEGM morphology serves as a viable indicator for assessing patient workload and body posture.
• Neural networks show promise for integration into active cardiac devices for enhanced monitoring.
• Further research into neural network applications in cardiac device diagnostics is warranted.