We use the BPLab (Petr Telegin, Russia) system for assessment the PTT, but I didn't see any comparative study with 2 or more devices.

With tangential i am actually refering to the method described by Chiu et al. 1991(http://www.sciencedirect.com/science/article/pii/0002870391901539) that takes into account two lines .. 1) tangential to the diastolic pressure 2) tangential to the location of the maximum first derivative. The foot of the wave is defined by the intersection of these two lines. To my knowledge it is also used by the commercial system of sphygmocor.

Yes some devices uses the intersecting tangent. Some others uses the maximal point of the second derivative of the signal of the pulse wave. As they are saying in the paper you pointed. We the one or the other.

I completely agree, but has there been a approach do define an non heuristic way of estimating travel time? Intersecting tangents may produce results even in noisier signals but still to my knowledge it has no obvious physical/ physiological background....

There are teams who have studied the pulse wave with MRI... it would be interesting to see this side...

Indeed there have been also studies where the flow wave transmission was studied but i think the question still remains. In this case i would even extend my question to " how can we use the tecniques on pressure and flow waves ?".

Good question !!! since Newton and Poiseuille, the fluid mecanics is still pointing unsolved questions !!! It has been admitted that when you put a sensor on an artery you must accept some imperfections, so the front of the coming wave have been accepted as the foot of the the pulse wave.

To clarify this a bit, i guess the imperfections that you are referring to are the wave reflections that alter each phase velocity to a different extend. Still i think the initial question remains. What is the best method for finding the foot of the wave? or even better to find the time lag between a distal and a proximal wave?

To my knowledge, there has not been a thorough comparison of pulse wave foot estimation (eg. comparing the intersecting tangents method to the use of the peak of the peak of the second derivative or other methods). One of the main reasons behind this is because there is no 100% accurate way of determining where the "foot" of the wave is - a bit like asking where is the corner in a circle. In fact, a great amount of work has been done on this area in image recognition, with much more complex algorithms being developed to try and accurately obtain the focal point of a curve.

Gary Mitchell published a nice paper comparing foot finding techniques in the rat:

Mitchell, G. F., Pfeffer, M. A., Finn, P. V & Pfeffer, J. M. Comparison of techniques for measuring pulse-wave velocity in the rat. Journal of applied physiology (Bethesda, Md. : 1985) 82, 203–10 (1997).

And I outlined and did some minimal comparison work between several foot finding methods in my thesis ( https://www.researchgate.net/publication/224903742_Structural_and_functional_effects_on_large_artery_stiffness_an_in-vivo_experimental_investigation ).

In the end, the various techniques of finding the foot have little influence on transit time and pulse wave velocity if the two sites being measured are a long distance apart. The only troubles come when the sites are very close together (such as in MRI studies where the local stiffness of an artery is being assessed, see attached).

Article The Relationship of Age With Regional Aortic Stiffness and Diameter

Thesis Structural and functional effects on large artery stiffness:...

Dear Mark

Much time has passed since i started this thread.

However please have a look in our 2013 work where we showed that for fully controlled signal conditions ( in silico waveforms) different techniques result in really differnet PTT values.

http://ajpheart.physiology.org/content/304/11/H1558.long

kind regards

-Orestis

Many thank's. Orestis

Are you still intersted in real files of data with pulse waves ?

Are you attending the Artery 14 ?

Regards

Thanks Orestis. I had already found and read the paper as it is exactly the study to answer this type of question. A very well designed and conducted experiment. The next step will be to see that manufacturers implement this improved algorithm in their PWV devices.