Neural method extracts varying heart and breath signals from data
PI-AMFM: Permutation-Invariant Learning for Variable-Cardinality AM-FM Mode Decomposition in Biomedical Signal Analysis
Machine Learning
Summary
Biomedical signals like heartbeats and breathing are complex and change over time, making them hard to separate into parts. The authors created a new neural network approach called PI-AMFM that can flexibly identify and separate these changing parts without knowing their number beforehand. They tested it on both fake and real biological data, showing it can accurately capture important patterns like heart and respiratory rhythms. This approach helps analyze physiological signals more effectively when their components vary.
What this means in practice
- •For medical device engineers: Build more accurate devices that extract changing cardiac and respiratory signals from physiological recordings without fixed component assumptions.
- •For wearable health technology teams: Improve algorithms in wearable sensors to dynamically track variable heart and breathing patterns using learned signal decompositions.$Commercial implications: Enables more precise health monitoring products by accurately separating time-varying physiological signal components in real-time.
Authors
Youngsun Kong, Ki H. Chon
Abstract
Physiological recordings often contain nonstationary oscillatory components whose number and dynamics vary across signals. Amplitude- and frequency-modulated (AM-FM) representations are well suited to characterizing such dynamics and have shown broad utility in biomedical signal analysis. Recent approaches have incorporated neural networks to learn mode decomposition patterns from data, but component cardinality is often predefined or determined through separate stopping or selection mechanisms. We propose a permutation-invariant neural framework for variable-cardinality AM-FM mode decomposition (PI-AMFM). PI-AMFM combines a multiscale temporal encoder, Mamba backbone, and component-presence estimation, with permutation-invariant Hungarian matching during training. On synthetic AM-FM signals, PI-AMFM achieved lower decomposition, instantaneous-frequency, reconstruction, and mode-count errors than the compared methods while preserving the overall trajectory pattern in a crossing-chirp example. On photoplethysmographic recordings, recovered modes captured cardiac and respiratory dynamics despite training only on synthetic signals. These results support the feasibility of PI-AMFM for variable-cardinality decomposition of nonstationary biomedical signals.