player259 - вот получил сегодня ответ от дочки насчет подходов к регистратору дыхания. Она как раз занимается разработкой всяких медицинских приборов. Привожу ее ответ без изменений:
> Do you have a sample picture how the blood presure changes with breath?
Here is the deal physiologially: An average heart rate is about 60-80 pulses/min at rest. Your breath rate is around 12 pulses/min at rest. There is a slight drift due to breath that you see if you monitor blood pressure continuously - it is riding on top of the 12 p/min sine, essentially. You can see that on both pressure and electrical measurements, as the impedance across your chest changes. The best example picture I can find is here:
http://www.sambasensors.com/preclinical ... -pressure/
Above is an invasive recording (needle in blood vessel) of a mouse, which has a much faster heart rate and breath rate than a human.
The thing is, we have currently 2 methods to record blood pressure continuously:
1) Invasively (something within a blood vessel, either implanted or inserted temporarily)
2) Non-invasively - stuff like pulse oximetry (that little light on your finger), or impedance measurements, which all have a baseline drift strongly dependent on contact (aka if a person is sitting still we can measure no problem, otherwise it's hard). Here is a pulse oximetry paper that deals with removing the breathing artifact, look at page R10 for a graphical example:
http://new.biomedix.com/UserFiles/PPG%2 ... 0Meas..pdf
An ECG shows similar drift with breath, as I mentioned before, because the impedance across your chest changes when your lungs expand. It has similar contact issues, and can also pick up muscle activity that will override the heart signal.
http://www.physionet.org/physiotools/edr/cic85/
If you measure the chest impedance directly, you have the same contact and muscle activity issue. Electrically speaking, your skin is about a 10MOhm resistor, and changes in moisture level can impact you greatly. TI has actually a paper out on this:
http://www.ti.com/lit/an/sbaa181/sbaa181.pdf
With all these methods, some very clever signal processing will probably have to be employed. Currently the way they measure your breath rate directly that I'm aware of is to have flow sensors in your breathing tube (for example exercise studies), or with patients in intensive care you can extract it from ECG or pulse oximetry, since they are not moving, though most of the time chest impedance is used, I think:
http://onlinelibrary.wiley.com/doi/10.1 ... 908.x/full
The method I've seen that works best is to strap a cloth band that has wire wrapped within it to make an inductor. When the chest expands, the inductance of the band changes. This is not dependent on contact.
http://en.wikipedia.org/wiki/Respirator ... ysmography
Here is a more modern example, they are using piezo instead of inductance, but same idea:
http://www.aastweb.org/Resources/FocusG ... _Intro.pdf
Finally the most esoteric idea: During surgery, they insert a thin white tube with a thermometer and acoustic holes
http://www.freepatentsonline.com/4577638.pdf
which has hole at the top for attaching a stethoscope - this way the doctor can hear the breath sounds and the heart sounds (which are distinct and therefore would have different frequency components as well). Perhaps there is some clever way to implement this principle noninvasively, but I think it would be very much subject to noise. Noninvasively this is done with a regular stethoscope:
http://en.wikipedia.org/wiki/Stethoscope
Call me if you are confused
