1 Biomedical Signal Processing and Biomedical Signal Processing and Computation Computation Prof. Zoran Nenadic Outline Outline •What are biomedical signals? •How are they generated? What are biomedical systems? •How do we model/simulate/analyze biological and medical systems? •How do we analyze/process biomedical signals?
36
Embed
Biomedical Signal Processing and Computationcbmspc.eng.uci.edu/GUESTLECTURES/bme1_guest_lecture_2007.pdf · 1 Biomedical Signal Processing and Computation Prof. Zoran Nenadic Outline
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Biomedical Signal Processing and Biomedical Signal Processing and ComputationComputationProf. Zoran Nenadic
OutlineOutline
•What are biomedical signals?
•How are they generated? What are biomedical systems?
•How do we model/simulate/analyze biological and medical systems?
•How do we analyze/process biomedical signals?
2
Biomedical SignalsBiomedical Signals
Biomedical signals are measurements of characteristic variables of biomedical systems.
Biomedical SystemsBiomedical SystemsSystem: a collection of units (elements, parts, devices, organs) , funct ional ly organized to accomplish certain goals by consuming, transforming and exchanging energy, matter and/or information.Biomedical System: a system that describes biological and medical processes (e.g. photosynthesis, respiration) and objects (e.g. cells, tissues, organs).Important to remember:
(A) A system often consists of smaller systems (called subsystems), which work together toward a common function.
(B) No system is isolated, i.e. a system always interacts with the environment and other systems.
13
Example 10 Digestive system
Environment Environment
smallsmallintestineintestine
liverliver
stomachstomach
Digestive systemDigestive system(A)
(B)
Digestive Digestive systemsystem
CardioCardio--vascular vascular systemsystem
Nervous Nervous systemsystem
Endocrine Endocrine systemsystem
Respiratory Respiratory systemsystem
14
15
Therefore, it makes sense to adopt systems approach to biology and medicine.
System science Traditional (experimental) science
systematic approach reductionist approach
study all pieces simultaneously
study one piece at a time
examine interactions(with other systems)
no interactions (system in isolation)
can make forecasts that cannot be extrapolated from data (measurements)
cannot make predictions
16
System theory is about applying mathematical formalism(symbols, relations, operations, etc.) to describe a systemat hand.
Mathematical model is an abstract mathematical description of a system.
This abstraction leads to the representation of a system with a box, with inputs and outputs:
System (*)Input Output
Think of (*) as a system of equations.Finding equation(s) (*) amounts to mathematical modeling.
17
Examples of mathematical models.
Example 11 Eye movements
motorneurons
synapses
muscles
θ
eyeball
18
Example 12 Metabolic process in a cell
membrane
cytoplasm
mitochondria
Cc
RoCm
Demo: cellular_dynamics.m
19
Example 13 Firefly synchronization
Demo: firefly_synch_neighbor.m
20
What can system theory do for us?
- What can be said about the behavior of the system base on the model. Are the outputs oscillatory, aperiodic? How quickly do they reach the steady state? What frequencies is the system tuned to, etc? (analysis)
- Can we figure out how inputs affects outputs? Can we find a set of inputs so that outputs have certain behavior. (control)
- How are the parameters affecting the behavior of the system. (sensitivity analysis)
21
Example 14 Disease outbreak (foot and mouth disease, UK 2001, Ferguson et al., Science, 2001)
y(x, y, t) - degree of infection at place (x, y) at time t.
u(x, y, t) - control variable (slaughter, vaccinate, quarantine, etc.)
What is the best control strategy?
22
Population dynamics
N(t) - population at year tB - birth rate (US: 1 person in 7 sec.)D - death rate (US: 1 person in 13 sec.)I - immigration rateE - emigration rate
Solution: - exponential growth
Does not quite fit the census data! Can we make better predictions?
Example 15
23
Time-varying system: B = B(t), D = D(t), I = I(t), E = E(t)
year λ (source: US Census Bureau)1999 0.90%1992 1.14%1950 2.05% (baby boomers)
Tweaking the parameters B(t), I(t), etc., we can control the growth of population.
These manipulations are performed with the model, and predictions are made.
24
Example 16 Blood pressure dynamics
QL(inflow from the left heart)
Qs(outflow to
systemic tissues)
P
Psa(T) – diastolic pressure
Psa(0) – systolic pressure
tT 3T2T
25
Example 17 Drug delivery dynamics
C - drug concentration in blood [kg/m3]KL - liver constant [1/s]VB - blood volume [m3]Rin - rate of injection [kg/s]
00
1/VB
δ(t) − bolus injection
g(t) = exp(−KL t)/V
B
00
Co
Ro
R0/(V
B K
L)
Rin
(t)
C(t)
I.V. drip bolus injection
26
Example 18 The diffusion of oxygen in a living tissue
- y(ξ,η,ζ,t) ∈ R is the oxygen concentration at a point (ξ,η,ζ) and time t
- D is the diffusion constant
- k is the oxygen uptake constant
If the diffusion constant is known, can make predictions as to how far the oxygen penetrates the tissue.
27
Example 19 Predator-prey equations (Lotka-Volterra)
x - the size of prey population at time t; y - the size of predator population at time t;α, β, γ, δ - parameters representing the interaction of the two species.
28
Example 20 Dynamics of emotions (S. Strogatz, Mathematics Magazine, 1988)
R(t) – Romeo’s love (R>0)/hate (R<0) for JulietJ(t) – Juliet’s love (J>0)/hate (J<0) for Romeo a,b > 0
Romeo is a fickle lover. Juliet’s love echoes Romeo’s. Their ill-fated romance consists of a never-ending cycle of love and hate.
0 2 4 6 8 10−1.5
−1
−0.5
0
0.5
1
1.5
Time
R(t)J(t)
29
Analysis of Biomedical SignalsAnalysis of Biomedical Signals
Recall: Biiomedical signals are generated by biomedical systems, therefore their properties are largely determined by the properties of the generating (biomedical systems) systems:
- non-stationary (time varying)- noisy (stochastic)- limited bandwidth (frequency content)- some of them are periodic- multiple signaling pathways (redundancy)
30
What can engineers do?
- denoising (remove the noise from signals)- spectral analysis(what are the important
frequencies)- compression (squeeze out the redundancy)- pattern recognition (e.g. FBI fingerprint
database, National DNA index system)- and many other things
31
Image denoising (filtering)Example 21
An image of a celebrity scientist garbled beyond recognition and successfully restored.