Download - DEPARTMENT OF BIOMEDICAL ENGINEERING
DEPARTMENT OFBIOMEDICAL ENGINEERING
Advisor: Mitch TylerClient: Dr. Christopher Green, MD
University of Wisconsin-Madison
Anna KarasAnna KarasMeghan Olson Meghan Olson Blake HondlBlake HondlZachary CohenZachary Cohen
• PTT is not calculated with the current Sleep Apnea Equipment (Flaga system)
• Design a solution which interfaces with the current system and is capable of detecting and measuring changes in pulse transit time which occur due to vasoconstriction during a micro arousal
Problem Statement
• Must interface with existing Sleep Apnea Study Equipment– Cannot void current system’s warranty
• Accurately detect the ECG and Pulse Oximetry peaks– Must reject signals due to noise or motion artifact
• Calculate PTT• Correlate PTT with other sleep study
measurements
Design Constraints
• Affects 12 million Americans• Obstructive Sleep Apnea
– Is caused by obstruction of the trachea (windpipe)
• From tongue, tonsils, uvula, fatty tissue in the throat, and/or relaxed muscles
– Individual can stop breathing for 10-30 sec.– Low blood oxygen level activates “flight or
fight” response • releases adrenaline• causes vasoconstriction
– Individual brought out of REM sleep– Can occur every 1-4 minutes (~400 times a
night)• Results in health problems and poor sleep
Sleep Apnea: Biological Aspect
Time will decrease with a micro arousal Measure from peak
Signal from ECG machine
Pulse Transit Time Theory
• Obstruction of breathing causes release of endorphins monitored by cerebral cortex
• The blood vessels constrict causing higher blood pressure (blood flows faster)
• The amount of time from heart pumping blood for the pressure wave to reach an extremity decreases
• PTT is correlated with micro arousals
Signal from Pulse Oximeter
t
PTT Correlates to Micro Arousals
• Blood pressure rises imply arousals (through PTT) are as good as EEG micro-arousals at predicting subjective sleepiness
• PTT detects subcortical/autonomic arousals, currently undetected by EEG monitoring
• ΔPTT is a sensitive indicator of inspiratory effort variance due to obstructions
• PTT detects events of a respiratory obstruction, but cannot differentiate types of events
» v = K*√{V*ΔP/ΔV} «
v=Pulse Wave Velocity V= Initial Vessel Volume K= Constant ΔP= Pressure Delta ΔV= Vessel Volume Delta
FLAGA / Somnologica System
• Incorporates a variety of sleep study signals: EOG, EEG, ECG, Pulse Oximeter, Respiratory CO2 , Noise/Vibration (snoring), Visual monitoring
– Signals can be sampled at 2-2000 Hz
• Somnologica program process signals and displays information in GUI format
• Data can be extracted in European Data Format (EDF)
Design 1: Spaced Transducer Technique
• Original method to test PTT• Pressure sensors placed on
skin consecutively on an artery
• Sensors measure the pulse moving in the artery below
• Sensors placed known distance apart (d)
• Analog to Digital conversion of signal
• Calculate the time between when the first sensor feels the pulse and the second
• Real time PTT calculation• Store information for later
reference
Pressure Sensor
Pressure Sensor
Pressure Sensor
d
d
Design 1: Spaced Transducer Technique
• Advantages– Simple design– Low cost– Independent of Flaga system
• Disadvantages– Noisy signals– Difficult to use with sleeping patients– Inaccurate compared with other techniques
Design 2: Automated Ambulatory Blood Pressure
• Blood pressure can be measured and correlated with PTT
• Measure blood pressure with piezoelectric microphones• Measures systolic and diastolic blood pressure• Portable, lightweight, can run on batteries• Used from 24-48 hr. periods
– measurements taken at set intervals (~every 5-60 min.)
• Tiba Medical – ~$1800 for a unit– $225/month to rent– Connects to PC– 512 MB storage capability
• Advantages– Lightweight– Portable– Interface with computer for
post analysis– Past studies show it does
not effect sleep of child– Another way to measure
micro arousals (correlated to blood pressure increase)
• Disadvantages– Costly– Dependent on cuff inflation– Data can only be taken at
intervals of 5- 60 min.– Not a continuous
measurement– Micro arousals may not be
detected
Design 2: Automated Ambulatory Blood Pressure
Design 3: PTT Software Solution
• Interface with existing Flaga System• Stored data interfaces with Matlab• Programming
– Signal analysis• Rejection of motion artifact• Detection of QRS and Pulse Oximeter wave
– PTT calculation
• Correlation of PTT with collected data from Flaga– Via Epic Number
Design 3: PTT Software Solution
• Advantages– Feasibile– Cost– Meets client requirements– Interfaces with existing system
• No additional signal measurements required
– Provides accurate PTT
• Disadvantages– Not real time
– Not portable
Design Matrix
Criteria Transducer Blood Pressure SoftwareCost 2 3 1
Feasibility (time restraints)
2 1 1
Satisfies Client Requirements
3 3 2
Ease of Use 2 2 1
Interfaces with Flaga system
3 3 1
Average Score 2.4 2.4 1.2
Good Good Best
Rated 1-3: 1=Best, 3=Worst
Potential Problems/Future Work
• Obtaining real-time signal from FLAGA system to output PTT in real-time
• Motion artifacts are introduced into ECG, influencing PTT– Need processing algorithms to account for
this
Convert EDF to text file
Using Matlab develop algorithms
to detect wave peaks
Create rejection algorithm for
motion artifact
Calculate PTT & display
Appendix / References
• http://www.biopac.com/AppNotes/app117PTT/ptt.html
• http://www.tibamedical.com/ABPM.htm• http://www.medicinenet.com/sleep_apnea/
page5.htm• http://www.disabilitysecrets.com/medicine-
medication-prescription-drugs-sleep-apnea.html• http://www.aafp.org/afp/20030601/2343.html