Bio Signal Measurement Flashcards
What are polarizable electrodes?
- No charge across the electrodeelectrolyte interface when current is applied - Current is displacement current - Electrode behaves like a capacitor - Overpotential is due to concentration Vc - Example: Pt - Pt electrodes are used as stimulating electrodes as there is no resistive current
What are non-polarizable electrodes?
-Current passes freely across the electrode-electrolyte interface. - Requires resistive energy to make transition - Perfectly non-polarizable electrodes have no overpotentials π π = 0 - In reality, π π = π π + π π΄, which is much lesser than π π = π π +π πΆ + π π΄ - Example: Ag-AgCl (approximately) - Ag-AgCl electrodes are commonly used in recording as they have a stable potential
What can relative motion between electrode and skin cause?
disturbs the charge layer and gives rise to
voltage.
Whats the stratum Corneum Layer of the skin?
has
high electrical resistance,
hinders charge movement and acts as a capacitor.
Preventive Measures for electrode and skin interface
Limit motion - Remove the stratum corneum - Signal Filtering : High pass filtering can help reduce the capacitance effect of the stratum corneum
Methods of removing the Stratum COrneum
- Soap and water
- Rubbing with isopropyl alcohol
- Abrading with sandpaper
- Using scotch tape
- Scratching with a needle
If blood appears when removing the stratum corneum what does that mean?
That would be the dermis layer - Caused by the complete removal of the epidermis - This will compromise the protective function
Whats the epidermis?
Protective layer
Whats the dermis?
contains the blood
vessels, nerves, and hair follicles.
What happens if theres movement of electrodes during measurement?
-When a polarizable electrode is in contact with the electrolytic gel, a double layer of charge forms at the interface.
β’ Movement of the electrode will disturb the distribution of the charge
and causes a temporary change in the πΈβπ
- The motion artifact is not as pronounced for nonpolarizable electrodes.
- Signal due to motion artifact is low frequency, so it can be filtered out if the biological signal has mostly high frequency content. (like EMG)
- When the signals have a lot of low frequency content (EEG, ECG, EOG) non-polarizable electrodes should be used.
What are types of electrodes?
- Surface Vs Percutaneous
- Non polarizable Vs Polarizable
- Reusable Vs Disposable
- Microelectrodes
Whats the line borne interface?
β’ Capacitive coupling of the body
- To the domestic supply
- To ground
β’ Current flow through body
- Too small to feel
- Safe
β’ Develops potential, Ecm
- Common mode
β’ Common signal is much larger
than the signalWhats a monopolar electrode configuration?
β’ It is not always true that the
inactive region will remain inactive
β’ Will include some power line noise
HAS AN ACTIVE AND INACTOVE CHANNEL
WHats a bipolar electrode configuration?
β’ Focuses just on the area of
interest
β’ Rejects 60 Hz noise
What IA are well suited for bimedical applications?
β’ Ultra high impedance (in the order of 1015 Ξ©) ensures that the electrodes are not loaded (Zout β 100 MΞ©)
β’ High and stable linear gains. A single external resistor can be used to
vary the gain 10-100X
β’ Very high CMRR β allows the measurement of small signals that are
contaminated with high common-mode voltages.
Problem with single op-amp implementation
- Want high input impedance and large Ad
- but impedence and difference gain are inversely proportional
Why do ECG, EEG, EMG, and EOG signals need to be filtered?
- Amplitudes are small
- Match instrumentation bandwidth to that of the desired signal
- To improve SNR of measurement
- Prevent aliasing when we are digitising the signal
Why do we filter?
Filtering clearly improves the signal quality
4 Types of filters
- low pass
- high pass
- bandpass
- band stop
whats a cut-off frequency?
Cut-off frequency delimits the pass and stop band - No attenuation below fc - Infinite attenuation above fc - Abrupt transition between the pass and the stop band (zero width for the transition band, or infinite slope at transition)
What are passband requirements?
β’ Ideally, we want the signal shape to be maintained at the output while still improving the SNR. β’ This requires that the time delay for each component of the signal (within the passband) is the same. Aka β Constant group delay
Whats a butterworth filter
- Attempts to give an ideal amplitude response
- Flat passband and a sharp transition
- But phase response suffers
Whats a bessel filter
- Ideal phase response, with a constant group delay in the pass band
- But the amplitude response suffers- pronounced slope in the passband
whats a tchebyshev filter?
- Ideal transition response
- Both the amplitude and the phase response suffers
What do we want from a filter?
1) Brick wall like amplitude response
2) Linear phase response
3) Steep transition between pass and
stopband
Butter Worth Magnitude Response
- -3dB point is fc β corner/cut off frequency
- Maximally flat response in the pass band
- Gain in stop band β 0 β’ As nβ, the transition region narrows
- Attenuation slope β -20n dB/decade
- As n β, the amplitude response looks more like the ideal brick wall response.
LOOK AT THE THREE FILTERS IN LECTURE 11&12
Comparison of BW, Bessel, and Tchebyshev
β’ BW β flat magnitude response
β’ Bessel β Linear phase response
β’ Tchebyshev β sharp transition
region
Why do we convert to digital formats?
β’ Offline
- Diagnostics
- Biomedical research
β’ Online (real time)
- Control of assistive devices like prosthesis and exoskeletons
- diagnosticsTwo steps of ADC
- Sampling -> continuous to discrete time
* Takes snapshots T seconds apart
* Called Pulse Amplitude modulation - Digitisation -> continuous to discrete amplitude
* Called Pulse code modulation
Frequency of Sampling Routine
β’ Fh - the highest frequency of the input signal
β’ Fs β sampling frequency
β’ Fourier transform of the sampled signal gives a repeated spectrum
(repeats every fs)
How are sampled signals recovered?
-Use a lowpass filter where fc=fh
β’ To avoid spectral overlapping, the fsshould at least be 2fh
β’ This is the Nyquist sampling criterion
β’ Practically, we always take fs >2fh
