Joint estimation of multiple cardiovascular parameters from blood pressure waveforms using multi-task learning

Estimation of cardiovascular parameters using blood pressure waveforms has been widely investigated for cardiovascular health assessment. However, previous studies have mainly focused on individual parameters, while their joint estimation within a unified framework remains underexplored. Given that vascular age, stroke volume, and peripheral resistance jointly shape waveform morphology, we formulated their estimation as a multi-task learning problem.

We proposed a cross-talk architecture, termed residual neural dimension reductor with temporal and channel attention (RNDR-TCA), which preserves main-task representations through residual connections and adaptively modulates transferred features across temporal positions and channels. Experimental validation was conducted using a cardiovascular hardware simulator dataset, with independently controlled parameter values treated as classification targets. RNDR-TCA achieved the highest average Macro F1 score of 0.836 among the evaluated architectures, compared with separate single-task models (0.784), a multi-class classifier (0.771), shared-trunk architectures (0.781–0.816), and other cross-talk architectures (0.792–0.824).

Additional analyses examined numerical estimation errors, statistical uncertainty, attention behavior, class imbalance, generalization to unseen heart-rate conditions, and model complexity. These findings suggest that attention-based inter-task feature exchange can contribute to effective joint classification of multiple cardiovascular parameters from a single blood pressure waveform. This work was supported by the InnoCORE program of the Ministry of Science and ICT (N10260002).

This work was also supported by the BK21 FOUR Program of the National Research Foundation Korea(NRF) grant funded by the Ministry of Education(MOE). Department of Electrical Engineering, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, Republic of Korea Healthcare AI Research Institute, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul, 03080, Republic of Korea Department of Mechanical Engineering, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, Republic of Korea Hong Junki, Adelle Ria Persad, Hyunwoo Song, Yuri Song, Haein Shin, Jaemin Shin & Yong-Hwa Park KAIST InnoCORE PRISM-AI Center, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, Republic of Korea The authors declare no known financial conflict of interest or personal relationships that could have influenced the work reported in this paper. This study used simulator-generated cardiovascular waveform data and did not involve human participants, identifiable personal information, or biological materials.

Therefore, institutional review board approval and informed consent were not required. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material.

You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.