Digital Processing of Photoplethysmographic Blood Volume Pulse (BVP) for Exercise Evaluation

Objective: 
A finger photoplethysmograph (PPG) is a non-invasive transducer to measure the relative changes of blood volume in a subject's finger .  The wave form obtained from such transducer represents the Blood Volume Pulse (BVP) of the subject. This signal has traditionally been used as a safe, non-invasive mechanism to assess the  Heart Rate (HR) of the subject, by focusing on the maxima of the waveform, possible following some differentiation to emphasize those peaks. However, only recently the representation of overall cardiovascular system changes by more specific features of this wave form has been considered. This research attempts to use Digital Signal Processing to quantify the typical changes in the BVP wave form through an exercise session, so that the resulting numeerical indices can be used as an objective, personalized gauge of the level of exercise achieved by an individual at any given time during the session.

Hypothesis:
We propose that the typical morphological changes that a BVP signal suffers during an exercise session reflect the level of modification in the subject's caridiovascular system caused by the exercise. Thus, we seek a Digital Signal Processing technique capable of quantifying those changes to provide a result that can be employed as an objective, personalized monitoring index for exercise.

Experimental Method:
Experiemntal subjects were asked to perform the following protocol:

BVP observations:
The typical changes observed in the BVP signals are described and illustrated below:

Blood Volume Pulse Quantification Methods

Methods:
The morphological modifications observed in the BVP signal through an exercise session have been evaluated by using two approaches:
 

I. Average Beat Histogram Analysis:
BVP beats comprised in 8-second data segments are automatically aligned according to their onset and a time-aligned average beat is obtained. Then the histogram of the average beat is evaluated and the relative size of two regions in the histogram are compared by forming a ratio H[50,75] / H[75,100]. This ratio is the single numerical index that represents the status of the BVP wave form.

In the subjects studied this ratio decreased from the REST condition (STAGE A), to the measurement made immediately after exercise (STAGE B), and  showed a tendency to increase towards its initial value after RECOVERY (STAGE C).

II.  BVP Harmonic Composition Changes:
At rest (STAGE A) the presence of the Dicrotic Notch in the BVP wave form implies a significant contribution of the second harmonic (f2) to its spectrum. Immediately after exercise (STAGE B) the BVP is more rounded, with a smaller Dicrotic Notch (if any). So, the first harmonic (f1) dominates the BVP spectrum in STAGE B. This disappearance of the Dicrotic Notch is captured by a marked increase in the ratio f1/f2 comparing the spectral contributions  of the fundamental and the second harmonic in the BVP wave form. This ratio increased from its REST value (STAGE A)  immediately after exercise in the subjects studied, showing a tendendency to go back to its original value after recovery (STAGE C).