Key result
Kernel Spectral Regression and Random Forest algorithms maintained acceptable peak detection accuracy (RMSE ≤ 0.15) in intracranial pressure waveforms even in the presence of high noise levels.
Why the study?
Monitoring quasi-periodic biological signals plays an important role in early detection of adverse patient events, requiring quantitative evaluation of computational frameworks for automatic peak extraction in ICP waveforms.
Answers are grounded in this paper and cite their sources.
Machine learning models, particularly Kernel ridge regression and peak tracking methods, provide robust peak detection in intracranial pressure waveforms even in the presence of noise and missing data.
Background : The monitoring and analysis of quasi-periodic biological signals such as electrocardiography (ECG), intracranial pressure (ICP), and cerebral blood flow velocity (CBFV) waveforms plays an important role in the early detection of adverse patient events and contributes to improved care management in the intensive care unit (ICU). This work provides a quantitative evaluation of existing computational frameworks for the automatic extraction of peaks within ICP waveforms. Methods : Peak detection techniques based on state-of-the-art machine learning models were evaluated in terms of robustness to varying levels of noise. Evaluation was performed on a dataset of ICP signals assembled from 700 hours of monitoring from 64 neurosurgical patients. The groundtruth of the peak locations was established manually on a subset of 13,611 pulses. Additional evaluation was performed using a simulated dataset of ICP with controlled temporal dynamics and noise. Results : The quantitative analysis of peak detection algorithms applied to individual waveforms indicates that all techniques provide acceptable accuracy (RMSE <= 0.15) without noise. In presence of higher level of noise, however, only Kernel ridge regression and Random forest remains below that error threshold while the performance of other techniques significantly deteriorates. Our experiments also demonstrated that tracking methods such as Bayesian inference and LSTM can be applied in a continuous fashion and provide additional robustness in situations where single pulse analysis methods tend to fail such as in presence of missing data. Conclusion : While machine learning-based peak detection methods require manually labeled data for training, these models outperform conventional signal processing ones based on handcrafted rules and should be considered for peak detection in modern frameworks. In particular, peak tracking methods that incorporate temporal information between successive periods of the signals have demonstrated in our experiments to provide more robustness to temporary artifacts that commonly arise as part of the monitoring setup in the NICU.
No takes yet. Share an insight, caveat, or question.
Wei et al. (2023) studied Neurosurgical patients requiring intracranial pressure monitoring (n=64). Machine learning-based peak detection algorithms (e.g., Kernel Spectral Regression) vs. Other machine learning and conventional signal processing methods was evaluated on Root Mean Square Error (RMSE) of peak detection. Kernel Spectral Regression and Random Forest algorithms maintained acceptable peak detection accuracy (RMSE ≤ 0.15) in intracranial pressure waveforms even in the presence of high noise levels.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: