Cardio Respiratory Disorders Using Ppg Signals Biology Essay

Detection of cardio-respiratory upsets is important for the forecast of several medical conditions. The present techniques involve complex diagnostic processs with demand for automatic categorization of acquired signals. This paper presents a fresh attack for sensing of cardio-respiratory upsets utilizing the photoplethysmograph ( PPG ) signals acquired utilizing simple medical apparatus which reduces cost and uncomfortableness to the topics. Signals are classified based on several clip and frequence characteristics extracted from the PPG signals. Decision tree based categorization has been implemented and truth of 94.44 % and 97.19 % has been achieved for the induced cardiac emphasis and apnea conditions severally. Based on the consequences, a preliminary diagnosing can be performed to observe the cause of abnormalcy in the recordings.


Photoplethysmography ( PPG ) is a non-invasive measuring technique, suited for mensurating blood volume alterations in the micro-vascular bed of tissue. [ 1 ] PPG has been extensively used in different clinical scenes such as monitoring of blood O impregnation, bosom rate, cardiac end product, blood force per unit area and respiration. [ 2 ] The measuring is performed by projecting seeable or infra-red visible radiation on the surface of the tegument and observing the transmitted or reflected visible radiation from the blood vass. The fluctuations in signal strength may either be periodic or non-periodic, originating due to combined influence of perfusion force per unit area and sympathetic vascular control. [ 3 ] PPG signals can be acquired at assorted sites of the organic structure such as fingers, ear lobes, brow or toes, leting different possibilities for informations acquisition protocols. [ 4 ]

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

The PPG consists of two constituents – a slow, changing DC beginning due to clamber and electrode response, and an AC constituent, typically around 1 Hz, which reflects blood volume pulsings. [ 5 ] The amplitude fluctuations in the PPG signal are influenced by respiration and the activity of sympathetic nervous system which, in bend, are attributed to independent control of peripheral vass. The forward force per unit area moving ridge is created due to ventricular contraction. [ 6-8 ] It flows from the bosom to the aorta and other smaller arterias, and is called the lifting stage or systolic constituent. These moving ridges are so reflected from the fringe at chief subdivision points and they constitute the diastolic constituent. The point of contemplation of the moving ridge is characterized by the dicrotic notch, whose tallness is considered to be a step of peripheral force per unit area moving ridge contemplation. [ 9-10 ] Pulse passage clip ( PTT ) is the clip interval between the systolic and diastolic extremums. Qualitatively, the systolic extremum ( anacrotic stage ) corresponds to the bosom status and the diastolic extremum ( catacrotic stage ) is used to find snap and other characteristics of the vascular system. [ 4 ] Fig. 1.1 shows the morphology of a typical PPG wave form.

Figure.1: Morphology of typical PPG wave form

The PPG signal form contains certain coded information sing the cardiovascular and respiratory province of the topic and a elaborate form analysis finally provides clinical informations for early sensing of cardiovascular and respiratory abnormalcies. The possible for pull outing diagnostic information from the PPG has been reviewed. [ 11 ] Ageing and arterial diseases are said to hold an consequence on the fluctuations of the AC constituent in PPG wave form. The peripheral pulsation is used to measure the province of wellness and disease in topics. [ 12 ] A weak or delayed response indicates marks of occlusive arterial diseases. [ 13 ] During apnea, vasoconstriction occurs and it is reflected in the PPG signal by a lessening in the fluctuation of amplitude. Several such quantitative steps that define the pulsation form have been used in the analysis of PPG wave forms. Time sphere characteristics such as PPG rise clip, peak-to-peak clip, peak amplitudes, form and variableness have been investigated. Pulse Rate Variability ( PRV ) exhibits frequency constituents from 0-0.5 Hz which are associated with the autonomic nervous system subdivisions. Frequency constituents in the 0.15 -0.4Hz represent vagal tone and these frequences are known as High Frequency ( HF ) constituents. Frequencies from 0.04-0.15 Hz manifest the activation of parasympathetic and sympathetic nervousnesss and are labeled as Low Frequency ( LF ) constituents. The ratio between LF and HF is defined as the sympatho-vagal balance. [ 14 ] This ratio has been termed Power Ratio ( PR ) . In this work, elaborate spectral and clip sphere analysis has been carried out for categorization of PPG.

Decision tree algorithm is a information excavation technique that recursively partitions a information set utilizing different methods until all the informations point are classified. This stage is known as the tree edifice stage and is performed in a top-down mode. Another stage of categorization is the tree-pruning stage which is performed in a bottom-up mode and is used to better the categorization truth of the algorithm. Using this technique, abnormalcies associated with the respiratory and cardiac system can be detected with a good grade of truth.



Data acquisition was performed on 45 healthy, non-smoking and non-athletic voluntaries ( 25 Male and 20 Female topics ) without symptoms of cardiac or respiratory diseases. The topics were seated during the scrutiny with their custodies laid comfortably on the thighs and were encouraged to maintain their fingers still to forestall gesture artefacts. First stage of acquisition was carried out for all topics in resting province. The 2nd stage of informations acquisition was carried out for the topics under different experimental conditions. For respiratory and cardiac instance surveies, 25 topics were asked to execute breathe clasp exercising and the staying 20 topics were asked to undergo physical exercising for 5 proceedingss prior to informations acquisition. The topics ‘ physical properties such as tallness and weight were besides recorded for post-acquisition analysis. Surveies were performed with blessing of the Centre for Biomedical Research and Signal Processing, SSN College of Engineering, Chennai and consent was obtained from all voluntaries before informations was acquired.

Data Acquisition

The PPG signal was acquired utilizing BIOKIT Physiograph ( Version 4.1 Build 3 ) , TekSys Electronics. LED and LDR, optical transmit-receive type finger detector of wavelength 940nm, with input electric resistance of 1Ma„¦ and a addition of -5K upper limit was used for geting the information from the topics. Frequency response was recorded at 2-40Hz. Casing included PCB mounted sender and receiving system in a Velcro belt. The most ideal location for the PPG detector was found to be the index finger of the manus because of the high signal strength and comfort of the topics [ 15 ] . The PPG signal was acquired from the right index finger. Subjects were made to sit in an unsloped place with the forearm placed in a relaxed place on the thigh. Care was taken to cut down gesture artefact due to respiration. After a short resting period for stabilisation, the PPG information was acquired post-prandial ( about 30 proceedingss after nutrient ) . A complete scheme of the informations acquisition process is presented in Fig. 2.1. During the process, the topics breathed spontaneously at more than 12 cycles/min and the signals were recorded at 1000 Hz trying frequence. Room temperature was regulated at 28 degree Centigrade with humidness at 50 % .

Figure.1: Timing diagram for informations acquisition

Feature extraction

Feature extraction is really of import for sensing of unnatural forms in the recorded bio-signals. Although several characteristics from clip and frequence spheres can be extracted, the designation of cardinal characteristics which provide concrete grounds of fluctuation from normal is indispensable for accurate categorization. Uniting several characteristics from both spheres makes the categorization system more accurate and fool-proof.

2.3.1 Pre-processing: PPG signals have lesser sophisticated morphology when compared to other cardinal physiological signals and therefore feature extraction and peak sensing are comparatively simpler. But baseline impetus and deformation may happen more often due to motion of the topics or their physiological status. It has besides been demonstrated that fluctuations caused by respiration, sympathetic activity and even arousal alterations such as sleepiness may do baseline impetus. [ 16 ] The major interventions impacting the PPG signals are gesture artefacts, respiration and low perfusion.

These interventions highlight the demand for preprocessing of PPG signals prior to the application of characteristic extraction algorithms. Baseline roving can be removed utilizing additive de-trending. Noise due to external factors and electrical intervention is removed by using a digital FIR band-pass filter of order 8. The cut-off frequence lies between 0.01-40 Hz. The recorded informations are so segmented in 15 2nd intervals prior to have extraction. All signal processing phases have been implemented utilizing MATLAB ( The Math Works Co. MATLAB© version 7.0 ) .

2.3.2 Time sphere: The morphology of the pulsatile constituent in PPG signal is said to alter with physiology. [ 17 ] The analysis of signal morphology is important as it is believed to incorporate information on the cardiovascular system and gives critical grounds for placing clinical conditions such as diabetes, coronary artery disease and arterial stiffness. [ 18 ] Several clip sphere characteristics have been extracted for the analysis of PPG signal. They can be classified as clip and amplitude indices. The typical clip sphere indices of the PPG signal is shown in Fig. 2.2.

Figure 2.2: PPG waveform exemplifying clip sphere indices

Stiffness Index ( SI ) is a step of the arterial dispensableness and is used to happen age related jobs such as arterial stiffness. It is besides considered to be a alternate to pulsate moving ridge speed ( PWV ) measurings, which is used as a marker to bespeak vascular amendss. [ 19 ] SI is calculated in footings of topic ‘s tallness and PTT. [ 8 ] The equation is as follows:

PTT is the clip interval between the systolic and diastolic extremums of the PPG signal. It reflects the theodolite clip of force per unit area moving ridges from root of subclavian arteria to the point of contemplation and back. The SI tends to increase with age as the PTT is quicker due to big arteria stiffness.

The Reflection Index ( RI ) is chiefly used to qualify alterations due to apnea conditions as the amplitude differences between systolic and diastolic extremums is found to be suppressed. It is besides considered to be a non-invasive marker for vascular appraisal. RI is derived as a ratio of diastolic extremum over systolic extremum. The equation for RI is given by

Where, a and B are systolic and systolic and diastolic extremum amplitudes severally. RI can be used as a diagnostic tool for vascular age and arterial conformity. [ 18 ]

Pulse rate and Body Mass Index ( BMI ) were besides taken into consideration as surveies have shown that the physical properties play a critical function in the belongingss of bio-signals.

2.3.3 Frequency sphere: In add-on to the clip sphere characteristics, several frequence sphere parametric quantities were besides considered for the intent of categorization. The combination of clip and frequence characteristics helps in accurate categorization. Power Ratio ( PR ) is a bench grade parametric quantity in measuring the power distribution in the acquired signal. [ 20 ] The equation is given by:

Low frequence power is considered to be a quantitative marker for sympathetic transitions and sympathovagal activity. High frequence power is a step of the cardiac parasympathetic vagal nervous activity. The different frequence sets have been illustrated in Fig. 2.3.

Very low frequency: 0.003 to 0.04 Hz

Low frequency: 0.04 to 0.15 Hz

Hafnium: 0.15 to 0.40 Hz

Figure 2. : Power Spectral Density ( PSD ) of PPG wave form with dominant frequence sets illustrated

Other characteristics such as breadth in the lower and higher frequence sets have been used to mensurate the denseness of the different frequence elements of the signal. Peak frequence and amplitude in both frequence sets have been recorded for farther analysis.

Decision tree classifiers

Decision tree is a really popular informations excavation technique used for categorization undertakings. A determination tree is a classifier that can be stated as a recursive divider of the case infinite which adopts a top-down acquisition system scheme [ 21-22 ] . It consists of root node, zero or more internal nodes, and one or more terminal nodes. Root node has no entrance edges. A node with outgoing borders is referred to as internal node, while all other nodes are called terminal nodes. In the determination tree, each internal node splits the case infinite into several sub-spaces based on the property values. Each terminal node is assigned to one category stand foring the most appropriate mark value. Cases are classified by voyaging the nodes from the root node down to a terminal node, harmonizing to the result of the trials along the way. A categorization regulation in the determination tree represents the way from the root node to that specific terminal node. The of import facet to build an efficient determination tree is to choose the good splitting standards. Gini diverseness index is chosen as splitting standards. The Gini dross step vitamin D ( T ) at node T is given as follows:

where S ( the dross standards ) = ? p2 ( j | T ) , for j=0,1,2, ….k. K denotes the figure of categories bing in that node and P ( j | T ) corresponds to the comparative frequence of category J in node t. The Gini diverseness index of a node is biggest when all the category in the node occurs with equal chance and is minimum when the node contains merely one mark category [ 23-24 ] . The extraction of categorization regulations from the determination tree is shown in the Fig. 2.4.

Figure 2. : Extraction of regulations from a determination tree [ 25 ]


The voluntaries take parting in this survey have a average age of 20.43 old ages ( run 19-22 ) and the average Body Mass Index ( BMI ) was 21.1977 ( run 19-25 ) . The physical feature of the topics from whom information has been acquired has been shown in Table 3.1.

Table 3. : Information on topics participated in the survey




Number of topics



Age ( old ages )

20.76 ± 0.44

20.09 ± 0.83

BMI ( kg/m2 )

21.57 ± 1.97

20.83 ± 1.28

Height ( m )

1.74 ± 0.07

1.60 ± 0.06

The signals were pre-processed to take any artefacts. Peak sensing algorithm was implemented to observe PPG composites for clip and frequence parametric quantity calculation. The ocular review of PPG recorded under different conditions revealed certain alterations in morphology. The signals and their PSD are presented in Fig. 3.1.

Figure 3.1: PPG wave forms under different emphasis conditions

By ocular review, it can be observed that there occurs a important alteration in wave form continuance and amplitude fluctuation between normal and induced conditions. These morphology alterations straight reflect on the SI and RI clip sphere characteristics. There is besides a important alteration in energy distribution in the PPG signals under different conditions, as seen in the PSD secret plans. Tables 3.2 and 3.3 summarize the different characteristics that have been exploited in this survey.

Table 3. : Comparison of clip sphere characteristics under different emphasis conditions



Respiratory Condition

Cardiac Condition

Stiffness Index ( SI )

6.89 ± 0.88

6.37 ± 1.35

7.96 ± 0.86

Reflection Index ( RI )

0.54 ± 0.12

0.69 ± 0.17

0.63 ± 0.17

Table 3. : Comparison of frequence sphere characteristics under different emphasis conditions



Respiratory Condition

Cardiac Condition

Power Ratio ( PR )

2.11 ± 0.32

1.01 ± 0.08

0.61 ± 0.07

Width in LF Band

0.07 ± 0.10

0.19 ± 0.11

0.18 ± 0.12

Width in HF Band

0.03 ± 0.04

0.076 ± 0.01

0.06 ± 0.03

PCR in LF Band

0.13 ± 0.12

0.02 ± 0.01

0.06 ± 0.15

PCR in HF Band

0.06 ± 0.05

0.02 ± 0.01

0.02 ± 0.01

The PR, SI and RI characteristics were selected for categorization intents utilizing the determination tree classifier. The categorization regulations for induced cardiac emphasis and induced apnea is shown in Fig. 3.1 and 3.2 severally.

Cardiac Decision Tree.png

Figure 3.1: Decision tree for categorization of PPG recorded under induced cardiac status

Respiratory Decision Tree.png

Figure 3.2: Decision tree for categorization of PPG recorded under induced apnea status

The figure of nodes, figure of regulations and the truth of the classifiers is given in the tabular array below.

Table 3. : Decision tree parametric quantities

Induced cardiac emphasis status

Induced apnea status

Number of nodes



Number of regulations




94.44 %

97.19 %

From the categorization regulations, PR was found to be an of import forecaster for the cardiac status and RI for the respiratory status. The derived regulations can non be explained to the full based on the standard medical cognition, since determination tree can besides detect unimportant regulations.

This paper discusses a fresh sensing mechanism for cardiac-respiratory upsets utilizing a determination tree informations excavation attack. Time and frequence sphere characteristics were applied and an automated regulation based system utilizing a determination tree was implemented. From the proposed survey, it can be observed that informations mining attack found to be giving promising consequences compared to nervous web based categorization attack. A categorization truth of 94.44 % and 97.19 % was obtained for cardiac and respiratory conditions. In order to better the categorization efficaciousness, several cross proof process with other informations excavation attacks are presently under probe. Further, the algorithm needs to be validated under existent life state of affairss to understand the efficiency of the current categorization algorithm.


The writers would wish to admit Dr. Mahesh V. , Department of Biomedical Engineering, SSN College of Engineering, Chennai, India for the PPG acquisition process.