Acquisition and Analysis of Biosignals

What is the origin of electrical, optical, mechanical, and acoustic biosignals? Where the signals are measured? Give some examples.

Electrical biosignals come from ion movements in cells and we measure them with skin electrodes, such as an electrocardiogram (ECG) measured on the chest to check the heart.

Optical biosignals come from light changes when blood pulses through vessels and we measure them with light sensors, such as a photoplethysmography (PPG) measured on a finger.

Mechanical biosignals come from body movements and heart vibrations and we measure them with motion sensors like accelerometers, such as an seismocardiography (SCG) measured on the breastbone.

Acoustic biosignals come from sounds made by heart valves and blood flow and we measure them with stethoscopes, such as a phonocardiography (PCG) measured on the chest.

What is an action potential, what are its phases, and how does it function?

An action potential is a fast electrical pulse across a cell membrane that happens in five simple steps: first, the cell starts at a resting state of -70 millivolts because ion pumps keep more positive ions outside the cell; second, an incoming signal hits a threshold of -55 millivolts and opens sodium channels, so sodium rushes in and pushes the voltage up to 30 millivolts during depolarization; third, sodium channels close and potassium channels open, so potassium rushes out and lowers the voltage back down during repolarization; fourth, potassium channels close slowly, which makes the voltage dip down to -80 millivolts during hyperpolarization so the cell cannot fire again right away; and fifth, ion pumps fix the ion levels and bring the cell back to its normal resting state.

Define stationary and non-stationary biosignals, and provide examples of both.

A stationary biosignal is a signal whose basic statistical numbers, like its average value and variance, stay the same over time, and a real-life example is a short recording of an electromyography (EMG) muscle signal while a person holds a steady weight. In contrast, a non-stationary biosignal is a signal whose average, variance, and frequency content change all the time, and a clear example is a phonocardiography (PCG) heart sound signal because it has loud sound bursts from heart valves mixed with quiet breaks.

What is an ECG, where does it originate, how are leads defined, and what is Einthoven's triangle? How many electrodes are used, and what are the components of the waveform?

An electrocardiogram (ECG) is a test that measures the electrical activity of the heart muscle from the skin using ten physical electrodes (four on the arms and legs, and six on the chest) to calculate twelve different electrical views called leads. Einthoven's triangle is a simple model that treats the body as a triangle between the right arm, left arm, and left leg, where Lead one measures the voltage between the arms, Lead two between the right arm and left leg, and Lead three between the left arm and left leg, following the rule that Lead one plus Lead three equals Lead two. ECG’s P-wave shows atrial activation, a PR-interval shows the delay between the upper and lower heart chambers, a sharp QRS-complex for ventricular pumping activation, an ST-segment where the ventricles stay active, a T-wave for ventricular recovery, and a QT-interval for the total pumping time.

Compare the ECG and PPG. What are their differences, and how many leads does a standard ECG have?

A standard electrocardiogram (ECG) uses ten electrodes on the body to measure twelve electrical leads that show the electrical signals of the heart with very high timing accuracy, while a photoplethysmography (PPG) is an optical test that uses only one small sensor with a light and a light detector on a finger or wrist to see blood pulses; this means an ECG is great for diagnosing serious heart problems like irregular beats and heart attacks, while a PPG is simpler and great for smartwatches to count heart rate, check heart rate changes, and measure blood oxygen levels, even though a PPG is more easily ruined by hand movements.

What biosignals other than the ECG can be used to measure heart rate and respiratory rate, and what are their advantages and disadvantages compared to the ECG?

Besides an electrocardiogram (ECG), we can measure both heart rate and breathing rate using a photoplethysmography (PPG) (finding heart rate from pulse peaks and breathing from small wave drifts), a phonocardiography (PCG) (finding heart rate from lub-dub sounds and breathing from sound volume changes), or an seismocardiography SCG (finding heart rate from chest jumps and breathing from chest motion); the main advantage of these methods compared to an ECG is that they are much more comfortable to wear in watches or can even work without touching the skin inside smart beds and chairs, but their main disadvantage is that any body movement or room noise easily disturbs them, and they cannot show the electrical health of the heart.

What is the origin of the EEG, and what are alpha, beta, theta, delta, and gamma waves?

The electroencephalogram (EEG) comes from the action potential of millions of brain cells, and we measure it with electrodes placed on the scalp to see five main brain wave speeds: slow delta waves (0.5 to 4 Hz) that happen during deep sleep, theta waves (4 to 8 Hz) that happen during light sleep and deep relaxation, alpha waves (8 to 13 Hz) that happen when resting peacefully with eyes closed and drop when eyes open, faster beta waves (13 to 30 Hz) that happen during active thinking and problem-solving, and very fast gamma waves (over 30 Hz) that happen when the brain connects different senses and works on hard tasks.

Compare the PCG, EEG, and EDA. What is their origin, what sensors are used, where are they attached, and what are their components?

Phonocardiography (PCG), electroencephalogram (EEG), electrodermal activity (EDA) are three very different body signals: a PCG comes from heart valve sounds and blood flow, uses an acoustic microphone on the chest, and contains the first heart sound (S1), the second heart sound (S2), and heart murmurs; an EEG comes from brain cell signals, uses electrodes on the scalp, and contains delta, theta, alpha, beta, and gamma waves; and an EDA comes from sweat gland changes controlled by stress nerves, uses two small electrodes on the fingers or palm, and contains a slow baseline sweat level (tonic level) plus fast sweat jumps caused by sudden feelings or events (phasic responses).

Is the PCG stationary or non-stationary, and why? What parameters are used to detect this, why is segmentation important, what segment widths should be used, and how is noise removed?

A PCG is a non-stationary signal because its loudness and frequencies change constantly as loud heart sounds alternate with quiet resting times; we can detect this change by tracking moving variance (which spikes during sounds and drops in between) and spectral flux (which tracks fast frequency shifts); segmentation is important because cutting the signal into separate heartbeats and quiet parts lets doctors find hidden murmurs, using windows that are 20 to 50 milliseconds long so the signal stays steady without cutting off low 20 to 40 Hz sounds; and we remove unwanted noise by using a 20 to 500 Hz bandpass filter to cut breathing drift and room noise, plus wavelet filtering to clean background noise while keeping the sharp sound clicks intact.

What is the Pan-Tompkins algorithm, and what are its stages? What frequency filters are used in biosignal processing, and what are two examples of their operation?

The Pan-Tompkins algorithm is a computer method that finds heartbeat spikes in an ECG through five basic steps: first, a 5 to 15 Hz bandpass filter cleans off slow baseline drift and fast muscle noise; second, a derivative step highlights the steep rise of the heartbeat; third, a squaring step makes all values positive and turns small bumps into tiny bumps and big spikes into huge spikes; fourth, a 150-millisecond moving integrator smooths the spike into one clean hill; and fifth, an adaptive threshold automatically adjusts its height to count the heartbeats correctly; and for filters in general, a low-pass filter lets low speeds pass and blocks fast noise (like a 40 Hz filter blocking muscle tremors), while a high-pass filter lets fast speeds pass and blocks slow drift (like a 0.5 Hz filter removing breathing movements).

What is the Short-Time Fourier Transform (STFT), how does it differ from the standard Fourier transform, what is an envelope, and how is an envelope extracted using synchronized averaging?

The standard Fourier transform tells us what frequencies exist in a signal but loses all information about when they happened, whereas the Short-Time Fourier Transform (STFT) chops the signal into small sliding time windows and checks the frequencies of each window to show a map of frequencies over time. An envelope is a smooth outer line that traces the top peaks of a bouncing signal to show how strong it is without showing every tiny bounce; and to pull out this envelope using synchronized averaging, we find a repeating trigger like a heartbeat spike, cut out equal-sized slices around each trigger, line the slices up so they match perfectly, calculate the average slice by adding them together (which cancels out random noise), and trace the final smooth outline using a Hilbert math step or a low-pass filter.

Explain EMG acquisition, testing, and analysis.

An EMG measures the electrical signals of working muscles using sticky electrodes on the skin over a muscle or small needle sensors pushed inside the muscle; testing usually includes a maximum voluntary contraction (MVC) test where the person pushes as hard as possible against a wall or strap so doctors have a 100 percent score to compare other tests against, along with nerve speed tests using mild electric shocks; and analyzing the signal involves cleaning it with a 20 to 450 Hz filter and calculating its average power to see muscle strength, as well as checking the frequency speeds, because as a muscle gets tired, its conduction speed drops and its average frequency clearly shifts down..

What is cross-correlation, why is it important in biosignal analysis, and what are two examples of its use? What is obtained when the Fourier transform is applied to cross-correlation, and what does it reveal?

Cross-correlation is a math tool that slides one signal across another to measure how similar they are at different time delays, which is very useful in medicine for measuring signal delays, aligning multiple sensors, and finding hidden heartbeats in messy data; two common examples are measuring pulse transit time (finding the time delay between an ECG heart peak and a finger PPG pulse to check blood pressure and artery stiffness) and finding bad heartbeats by matching new beats against a library of saved normal and abnormal beat shapes; and if we run a Fourier transform on this cross-correlation result, we get the Cross-Power Spectral Density (CPSD), which shows what shared frequencies the two signals have and gives the exact phase delay between them at each frequency.