PE2-3.1 Introduction to Biomedical Signals

What this is and why it exists

Every signal in this subject starts in the same place. A cell changes the voltage across its membrane, and that change spreads.

Understand that one mechanism and the rest of the course stops being a list of acronyms. Each recording is the same event, seen from a different distance through a different amount of tissue.

The vocabulary

• Action potential — the rapid voltage change across a cell membrane that carries excitation.
• ECG — the electrocardiogram, the heart's electrical signal recorded at the skin.
• EEG — the electroencephalogram, the brain's signal recorded at the scalp.
• EMG — the electromyogram, the signal from active muscle.
• PCG — the phonocardiogram, the sound the heart makes.
• ENG — the electroneurogram, recorded from a nerve.
• ERP — an event-related potential, a response to a known stimulus.
• EGG — the electrogastrogram, from the stomach.
• Non-stationary — having statistics that change over time.
• Artefact — anything in the recording that is not the signal you wanted.

The mental model

A resting cell holds a voltage across its membrane. When it is excited, channels open, ions move, and that voltage swings rapidly and then recovers. That is the action potential. One cell produces a tiny effect. Many cells acting together produce something measurable at a distance. What you record at the skin is the sum of a great many of them, filtered by the tissue between.

That picture explains the differences between the recordings. The heart's cells act nearly together and the heart is large, so the signal reaching the skin is comparatively big. Brain activity is less synchronised, the source is smaller, and the skull lies between, so the signal at the scalp is much smaller.

Learn each signal by two numbers: its amplitude and its frequency band. Those two numbers decide the entire amplifier design in a later topic, and they explain why some recordings are routine and others are difficult. A heart signal at the skin is measured in millivolts. A brain signal at the scalp is measured in microvolts, which is a thousand times smaller.

The naming convention is worth knowing early. The heart's waveform has named parts, one for the atria contracting, a sharp group for the ventricles, and one for recovery. Almost everything clinical is stated in terms of those names and the intervals between them.

The remaining signals extend the picture. Muscle activity gives a noisy signal whose size grows with effort. Nerve recordings are small and fast. The heart also makes sound, which is mechanical rather than electrical. Relating the two in time is what makes both interpretable. Stomach activity is very slow. Event-related potentials are responses to a known stimulus, and they are so small that they cannot be seen in a single recording at all.

Now the honest framing, and it deserves real attention.

These signals are hard for reasons that ordinary signal processing does not face. They are small and the interference is large. The interference is often another physiological signal, so frequency alone will not remove it. Muscle activity contaminates a heart recording, and the heart contaminates a brain recording. They are non-stationary, so statistics measured in one minute may not describe the next. And they are patient-specific, so a threshold tuned on one person may fail on another.

That is why methods that work well on radio signals often disappoint here. It also explains the shape of the whole subject. State what you are trying to achieve first, then choose a technique, rather than reaching for a familiar one.

The destination is a system that supports a clinician rather than replacing one. That framing is not politeness. It decides how results should be presented, because a number offered without its uncertainty invites a decision it cannot support.

What you should now be able to explain or do

Describe the action potential and say why it explains every signal here. Say why a scalp recording is so much smaller than one from the chest. Give the amplitude and frequency band of the main recordings and say what that dictates. Name the parts of the heart's waveform. State the four difficulties of biomedical signal analysis and say why interference cannot always be filtered by frequency.

Check yourself