*4.1. Arterial Waveform Measurement Modalities Amenable to Wearable Technology: Obtaining Reliable High Signal-to-Noise Features*

The original studies establishing the basis for CRM [30] used arterial waveforms measured by volume-clamping based continuous blood pressure measurement technology (i.e., Finapres) [73]. Such arterial waveforms have been demonstrated to accurately represent corresponding peripheral blood pressure waveforms obtained using arterial lines [74], and are thereby considered to be a reference standard for non-invasive continuous blood pressure measurement. While the volume-clamping technique is quite accurate at acquiring analog arterial pressure waveforms, the system required is expensive, large, and power-hungry, and thus unsuitable for point-of-care settings. Accordingly, to facilitate translation of CRM outside the lab, investigators have explored other techniques for obtaining analog arterial waveforms that *resemble* blood pressure waveforms; namely, the most commonly employed signal has been the photoplethysmogram (PPG) [39].

Example sensing modalities that provide arterial pressure waveforms (or analogs) that are directly amenable to CRM are summarized in Figure 5. Figure 5a shows volume-clamping based finger cuff BP measurement (i.e., Finapres), which uses an LED and photodiode (PD) to capture the blood volume as a function of time in the finger, and uses a servo controller to modify cuff pressure (Pcuff) dynamically to set the finger blood volume to a constant level. The output pressure required from the controller is thus the hemodynamic pressure inside the artery, and a waveform can be outputted representing the continuous arterial BP signal. The measurement can only be obtained from the finger.

**Figure 5.** Illustration of different waveforms capture techniques that can provide a blood pressure (BP(t)) or blood volume pulse (BVP(t)) signal. (**a**) Volume-clamping based finger-cuff BP measurement (i.e., Finapres). (**b**) Photoplethysmography (PPG) based blood volume pulse measurement. (**c**) Tonometry based arterial pulse measurement. Image created based on Lee and Nam [75]. (**d**) Ultrasound array based arterial pulse measurement. Image created based on Wang and Xu [76].

Note that the approaches besides volume-clamping based BP measurement would result in waveform characteristics that would differ from the existing library of LBNP based arterial waveform measures used for CRM. Thus, a small data collection of approximately ten subjects may be needed with the new modality such that transfer learning or fine tuning methods for retraining the algorithm can be implemented. Following such methodology, the existing database can still be leveraged with the new sensing modality to yield accurate CRM results.
