FactorAdjust: FactorAdjust

In stw32/PulseWaveform: Extraction of Photoplethysmography Waveform Features and Decomposition Modeling of Pulse Waveforms

FactorAdjust

Description

FactorAdjust identifies the most appropriate factor value to be used in undetrending an ISO study ppg time series (for inputting into the Undetrend function). It focuses on a single beat segment of the ppg time series and adjusts the factor value to correct morphological distortions common to detrending in the ISO data. These are 1. abnormal tail decay (e.g positive gradient to tail) 2. values significantly below the approximated baseline (such that notch features are negative and therefore meaningless). The function operates within the FindUndetrendingParams function.

Usage

FactorAdjust(data, factorCutoff, ppg, u, beat, a., test, gs=gs, beatTime, nextTime, plot = T)

Arguments

data	an individual beat segment of raw ppg time series data
factorCutoff	the maximum acceptable tail gradient (default 0)
ppg	the ppg time series
beat	dataframe of all detected peaks in the ppg time series
u	Undetrend function
a.	index number indicating which beat (relative to the calculated minimum IBI of the time series) to sample
test	index number indicating the beat identified as the minimum IBI of the time series
gs	model2.GetSegment function
beatTime	index value for the start of the beat to be assessed
nextTime	index value for the end of the beat to be assesssed
plot	logical, if TRUE, plots the beat segment for each iteration of the while loop, demonstrating the change in morphology as it is corrected

Value

factor_value

the appropriate factor value for restoring proper morphology to the ppg time series, to be inputted into the Undetrend function to complete ISO study data preprocessing

Note

A minimum IBI section of the time series is relevant to this function given its purpose in preprocessing ISO study data. The morpological change associated with isoprenaline results in a lowered notch (and IBI), and thus any alterations to morphology to ensure notches above 0 should be targeted at the waves with the lowest notches.

Within the ISO data four types of waveform endings are possible:

1. prolonged negative slope (morphologically normal) 2. prolonged positive slope (due to heavy detrending) 3. ~5 positive slope coming out of the notch, followed by ~5 negative slope as a rather short tail i.e an n shape (these tend to be ISO waves (they are short due to high HR)) 4. ~5 positive slope at the end of the tail, preceded by a prolonged negative slope i.e a v shape (due to tail noise)

1. and 2. will be processed the same regardless of which of the last 10 datapoints the gradient is taken from 3. In the initial 'last 5' datapoint assessment, these will be found to be negative and thus will continue being iterated on based on the last 5 datapoints. 4. In the initial 'last 5' datapoint assessement, these will be found to be positive, and so will be iterated on based on NOT the last 5 datapoints.

Examples

FactorAdjust(data, factorCutoff, ppg, u, beat, a., test, gs=gs, beatTime, nextTime, plot = T)

stw32/PulseWaveform documentation built on Dec. 6, 2022, 2:50 a.m.