Medical devices such as ECG monitors, EEG headsets, pulse oximeters, and blood pressure sensors measure signals produced by the body. These signals can be extremely weak, often only microvolts or millivolts, so they must be carefully cleaned up before a doctor can use them. Signal processing turns noisy sensor readings into clear graphs, numbers, and alarms.
This matters because better signal quality can lead to faster and more accurate diagnosis.
A typical medical signal-processing pipeline starts with a sensor that converts a body signal into a small electrical voltage. An amplifier boosts the signal, filters reduce unwanted noise, an analog-to-digital converter changes the signal into numbers, and software displays the result. Each stage must preserve the useful medical information while removing interference from motion, muscles, power lines, and electronics.
The final display may show a waveform, heart rate, oxygen level, or warning message for a clinician.
Understanding Medical Technology: Signal Processing in Medical Devices
Many body measurements use a pair of electrodes or light sensors rather than one sensor alone. With ECG recording, two electrodes pick up electrical activity from slightly different places on the skin. The device focuses on the difference between those readings.
Interference that reaches both leads in a similar way, such as mains electricity from nearby wiring, can then be reduced. This is called common mode rejection. Good electrode contact matters as much as clever electronics.
Dry gel, sweat, body hair, loose leads, or moving skin can change the contact resistance. The result may look like a heart problem even when it is only a poor connection.
Not every unwanted part of a recording should be removed. Breathing can slowly shift an ECG baseline. Muscle activity can add fast, uneven signals.
A patient turning in bed can create a large motion artifact. Filters are chosen by considering the feature that clinicians need to see. If filtering is too strong, it can flatten a real change in the waveform.
If it is too weak, noise can hide the change. Some filters can shift the timing or shape of a pulse.
That matters when software measures intervals between ECG waves or detects a sudden rhythm change. Signal processing is therefore a balance between a cleaner trace and a faithful trace.
Digital processing makes it possible to inspect a signal in small time steps. The sampling rate must be selected before recording begins. A rate that is too low can create aliasing.
This means rapid changes are recorded as a false, slower pattern. Once aliasing has happened, software cannot reliably recover the original detail. Devices often use an analog filter before digitizing to block frequencies that would cause this problem.
After digitizing, software may smooth the data, locate peaks, compare several beats, or reject sections affected by motion. A monitor should not treat every sharp spike as a heartbeat. It checks features such as spacing, width, and consistency over time.
Pulse oximeters show why measurements need both physics and careful algorithms. They shine red and infrared light through a fingertip or earlobe. Blood changes how much of each light reaches the detector.
The device looks for the repeating change caused by each pulse, then uses the relative changes to estimate oxygen saturation. Nail polish, cold hands, poor circulation, bright room light, and movement can disturb this estimate. A number on a screen is not automatically a perfect fact.
Students should learn to ask what the sensor directly measured, what the software calculated, and what conditions could make the result unreliable. Medical devices are tested against reference measurements, yet trained staff still check whether the displayed result matches the patient and the rest of the evidence.
Key Facts
- Amplification increases signal size: Vout = gain × Vin.
- Signal-to-noise ratio compares useful signal to unwanted noise: SNR = signal power / noise power.
- A low-pass filter lets low frequencies pass and reduces high-frequency noise.
- A high-pass filter lets high frequencies pass and reduces slow drift or baseline wander.
- Digitizing uses sampling, where the sampling rate fs must be high enough to capture the signal clearly.
- Nyquist rule: fs ≥ 2fmax, where fmax is the highest frequency in the signal.
Vocabulary
- Biosignal
- A biosignal is a measurable signal produced by the body, such as the electrical activity of the heart or brain.
- Sensor
- A sensor is a device that detects a physical or biological quantity and converts it into an electrical signal.
- Amplifier
- An amplifier is an electronic circuit that makes a small input signal larger without ideally changing its shape.
- Filter
- A filter is a circuit or algorithm that keeps desired parts of a signal and reduces unwanted frequencies.
- Analog-to-digital converter
- An analog-to-digital converter is a device that changes a continuous voltage signal into a sequence of digital numbers.
Common Mistakes to Avoid
- Assuming amplification removes noise, which is wrong because an amplifier usually boosts both the useful signal and the noise unless filtering is also used.
- Choosing a sampling rate that is too low, which is wrong because important waveform details can be lost or distorted by aliasing.
- Using the same filter for every medical signal, which is wrong because ECG, EEG, and other signals have different useful frequency ranges.
- Ignoring sensor placement and contact quality, which is wrong because poor contact can create motion artifacts and weak or unstable measurements before processing even begins.
Practice Questions
- 1 An ECG electrode measures a 2 mV signal. If the amplifier gain is 500, what is the output voltage in volts?
- 2 A medical device must measure signal frequencies up to 150 Hz. What is the minimum sampling rate required by the Nyquist rule?
- 3 A heart monitor waveform is drifting slowly up and down because of patient breathing while the sharp heartbeats are still present. Which type of filter would help reduce the slow drift, and why?