Hi,
In a study on the effects of breathing frequency manipulation on blood pressure variability we are going to perform spectral analysis of blood pressure. Each breathing block is 3 minutes long, mean number of beats = 219 +- 28 in each block.
There are some specifications for such spectral analysis that I need to know:
- Which method is more appropriate for such data? AR or FFT?
- Method of interpolation (e.g. linear) and re-sampling rate (1 or 2 Hz)?
- Recommended segment length/number and % of overlap?
Tnx.
AR provides better spectral estimation for short segments. 219 samples is kind of short.
I would use cubic spline interpolation and resampling at 2 Hz, or maybe even 4 Hz. Try both. 1 Hz is too course.
the overlap compensates for the antileakage window effect of attenuating the ends of the segment. 50 % overlap is adequate, but you might want to try 75 % too to have a finer time sequence (but know that you are merely interpolating in time when you use more than about 50 % overlap).
Have fun!
If applying splines and FFT then 1024 samples will work out.
If removed beats remain uncorrected by any spline then unevenly sampled spectral methods of Lomb will work .
Sounds like you are sampling respiration and bp together- if so then a spline will probably be needed and 1024 samples spread over over a few minutes sounds about right.
Sounds interesting Good luck
Dear Ali,
We usually use cubic interpolation and resampling at 4 Hz for BP and RRI analysis. What frequency method is better to analyze data depends on the goal you set. What respiration manipulation your participants performed?
For example, If participants paced breathed at various frequencies and you want estimate how amplitude of BP oscillation related to respiration frequency is right to apply the Fourier filtration analysis which gives amplitude of the response and phase shift between BP and respiration for each stimulation frequency. This analysis is described in our paper. I attached it please see p. 9 if you interested.
Measure the BP and calculate the variability via SD or other measures of variability. Merrill
Dear All, Tnx for your guides.
Evgeny Tnx for sharing your paper, interesting indeed.
Our study aim was to test the effect of various breathing exercises on Vagal activity and Baroreflex Sensitivity. Subjects performed uncontrolled breathing (NB), paced breathing at 0.23 Hz (PB) and 0.1 Hz. In breathing at 0.1 Hz we applied an inspiratory threshold load (SB0-SB10). We hypothesized that it helps deep breathing to be more efficient, creating higher oscillation in BP and subsequently higher influence on HR through the baroreflex.
I did spectral analysis of both HR and BP based on FFT and the attached picture are the results. It seems that the intervention could increase the amplitude of BP and HR oscillations at the desired frequency. To better understand the mechanism, I'll try to perform transfer function analysis, as you suggested, also including the tidal volume. We can also perform same analysis on stroke volume and peripheral vascular resistance derived from the Modelflow.
It was very interesting, and not expected, that in your biofeedback study there was no consistent respiratory pattern despite clear oscillations in HR/BP at the desired frequencies. This I think is of clinical importance since it suggests that the mechanisms of increased RSA, BRS, Vagal activity by breathing exercises may be different from that induced by biofeedback. These two different methods may have different long-term effects and different clinical efficacy in various situations.
BW
Ali
Blood pressure variability includes rhythmic and non rhythmic fluctuation that appear as clear peaks or broadband power with spectral analysis. Spectral analysis methods commonly used are fast Fourier transform versus autoregressive modeling, time varying versus broadband spectral analysis.
Please let me know if the following references/sites are helpful to you:
1. From the first spectral analysis of the blood-pressure
...
www.biomed.cas.cz/physiolres/pdf/prepress/932533.pdf
From the First Spectral Analysis of Blood Pressure Variability in the World to the Present Time: Contribution of the Department of Physiology of the Faculty of Medicine,
Spectral Analysis of Blood Pressure Variability in Heart ...
hyper.ahajournals.org/content/25/4/643
Abstract The cardiac transplant patient provides a unique model for the study of blood pressure variability in the absence of heart rate variability. We examined the ...
2. Spectral Analysis of Arterial Blood Pressure Variability ...
cinc.mit.edu/archives/2003/pdf/505.pdf
Spectral Analysis of Arterial Blood Pressure Variability: Assessing the Accuracy of the Finometer Device R Maestri, GD Pinna, E Gobbi, E Robbi, MT La Rovere
Dennis
Dennis Mazur
Here the results of our study on the effect of using inspiratory threshold load during slow deep breathing on cardiovascular parameters, details of spectral analysis of blood pressure are provided in supplementary file:
https://doi.org/10.1111/psyp.13447
- Badges
- Science topic
- Similar topics
- Psychology