Introduction. Editorial on 'Signal processing in vital rhythms and signs'.

On *A Eng Signal processing has become a core technology in medical devices today, as it is crucial in a wide range of applications. It is used to automate the measurement of various signal characteristics, which previously was done manually and, as a result, subjectivity is reduced and reliability is increased. Another purpose is to filter out undesired signal components of technical or physiological origin so that the signalto-noise ratio is improved and subsequent analysis facilitated. Signal processing is also of key importance when uncovering components with low amplitude and/or subtle variations in frequency, which are very difficult, if not impossible, to observe by the naked eye. For example, signal processing plays a crucial role for the detection of the alternans phenomenon occurring during cardiac repolarization. Information about alternans has turned out to be the most useful for risk stratification of certain cardiac patients. The analysis of heart rate variability represents another application where signal processing is essential for the characterization of tiny oscillations in rhythm; the resulting measurements convey unique information about the autonomic nervous system. Considering that today’s recording equipment and computers are quite powerful and available at affordable prices, limitations related to implementation are rather procedural than computational. The role of signal processing can therefore be expected to further increase in the future as novel techniques are developed and used to uncover and characterize hidden activity which may be present in a signal. Traditionally, biomedical signal processing has been largely synonymous to ‘unimodal’ analysis in which only one type of signal has been subjected to analysis at a time. Countless approaches have been developed to unimodal analysis of the most common bioelectrical signals, i.e. the electrocardiogram (ECG), the electroencephalogram (EEG) and the electromyogram (EMG). However, multimodal signal modelling and processing are currently receiving considerable attention as it is highly desirable to account for the interdependence that exists between different Phil. Trans. R. Soc. A (2009) 367, 207–211 doi:10.1098/rsta.2008.0239 Published online 20 October 2008