Pulsatile physiological signals, characterized by rhythmic fluctuations, are vital for assessing health conditions and are widely used in wellness devices and medical equipment. Despite their significance, models addressing domain-specific unmet needs and considerations have not been developed as much as in other fields. Therefore, building on the foundation of variational autoencoders, we introduce VABAM, a novel model for the amplitude-based synthesis of pulsatile physiological signals. The uniqueness of VABAM lies in its ability to maintain the morphological identity of signals throughout the synthesis process, achieved by integrating pass filter effects within the variational autoencoder architecture. To assess the effectiveness of the model, we developed three novel metrics based on joint mutual information. These metrics were aimed at evaluating the disentanglement of latent spaces, influence of ancillary information on signal morphologies, and controllability of amplitude-based synthesis within morphological identities. Comparative analyses demonstrated that VABAM and its variants were notably effective at preserving morphological integrity, highlighting their potential to minimize morphological distortions in physiological signal processing and their compatibility with artificial intelligence models employing frequency and amplitude features. Additionally, the proposed metrics, compatible with probabilistic models, were empirically proven to capture the characteristics of various models from multiple perspectives.

Pulsatile physiological signals, characterized by periodic waveform traits, are essential for patient monitoring. However, artifacts considerably compromise signal fidelity, thereby rendering the signal cleansing process essential. Although previous studies have focused on algorithmic advancements, domainspecific factors in this field are yet to be comprehensively understood. Therefore, this study conducts an empirical investigation into four critical signal cleansing aspects in the medical field. First, we verified the performance enhancement of incorporating domain knowledge into cleansing techniques. Second, we emulated real-time cleansing requirements from clinical environments. Third, we explored the similarities and differences among three evaluation approaches: quantitative assessment of the cleansing performance, qualitative assessment of the same by clinical experts, and a cleansing effect assessment on downstream tasks. Finally, we evaluated whether the willingness to accept these techniques varied according to the knowledge incorporation mode and professional groups comprising clinicians and engineers. The results revealed that domain knowledge incorporation enhances cleansing performance, visually appealing cleansing results can mislead experts, and the willingness to accept cleansing techniques varies depending on the knowledge incorporation method, professional group, and experience level.

Monitoring arterial blood pressure (ABP) in anesthetized patients is crucial for preventing hypotension, which can lead to adverse clinical outcomes. Thus, several efforts have been made to develop an artificial intelligence-based hypotension prediction index. Nevertheless, the use of these indices is limited because they may not provide a convincing interpretation of the association between predictors and hypotension. Herein, we developed an interpretable deep learning model that forecasts hypotension occurrences 10 min before a given 90 s ABP record. Internal and external validations of model performance reported the area under the receiver operating characteristic curve (AUC) as 0.9145 and 0.9035, respectively. Furthermore, the hypotension prediction mechanism can be physiologically interpreted by using predictors representing ABP trends that are automatically generated in the proposed model. Ultimately, we demonstrate high-applicability of a deep learning model that has a high accuracy performance and provides an interpretation of the association between ABP trends and hypotension in clinical practices.