Famous signal #10 — Electrocardiogram
Some signals are everywhere in engineering, but only a few cross over into everyday life. The electrocardiogram, or ECG, is one of them. It lives at the intersection of physiology, electronics, and signal processing. A clinician sees a diagnostic trace. An engineer sees a low-amplitude, noisy, quasi-periodic waveform carrying layers of information about an underlying dynamical system. A patient simply sees proof that the heart is being watched.
An ECG is the recorded electrical activity of the heart as measured from the body surface. It is typically represented as a time-domain waveform with repeating structures such as the P wave, QRS complex, and T wave.
Name
Electrocardiogram (ECG)
Category
Biomedical
Where it comes from
The ECG comes from the heart’s electrical conduction system. Every heartbeat is initiated and coordinated by electrical impulses that propagate through cardiac tissue. As cardiac cells depolarise and repolarise, they generate tiny voltage differences that spread through the torso and can be measured using electrodes placed on the skin.
Those electrodes do not measure the heart directly in a mechanical sense. They measure the electrical consequences of the heart’s activation pattern. In other words, the ECG is not the pumping action itself, but the electrical script that makes the pumping possible.
What the physical quantity is
The physical quantity measured in an ECG is the electric potential difference, usually in the millivolt (mV) range, as a function of time.
This is important from a signal processing perspective. ECG is a low-amplitude signal, which makes it especially vulnerable to noise and artefacts. Motion, muscle activity, poor electrode contact, power-line interference, and baseline drift can all distort the waveform. So even before interpretation begins, the problem is already an engineering problem: how do we recover a meaningful physiological signal from imperfect measurements?
What signal processing techniques can be used to extract the information
ECG is one of the classic playgrounds for signal processing because almost every stage matters.
Filtering is usually the first step. A high-pass filter can reduce baseline wander caused by breathing or electrode movement. A low-pass filter helps suppress high-frequency noise, including muscle artefacts. A notch filter is often used to remove power-line interference at 50 or 60 Hz.
Peak detection, especially QRS detection, is one of the most important processing tasks. The QRS complex is the most visually prominent part of the ECG and usually serves as the anchor for heartbeat segmentation. Once QRS complexes are reliably detected, heart rate and beat-to-beat intervals can be estimated.
Segmentation allows the signal to be broken into individual cardiac cycles. This makes it possible to measure clinically relevant intervals such as PR, QRS duration, QT, and RR interval.
Get Bogdan-Adrian Enache’s stories in your inbox
Join Medium for free to get updates from this writer.
Time-domain feature extraction is used to quantify waveform amplitudes, durations, slopes, and intervals. These features support both rule-based analysis and machine learning pipelines.
Frequency-domain analysis can be useful in specialised settings, especially for studying heart rate variability or identifying spectral characteristics of noise and rhythm disturbances.
Wavelet transforms are particularly effective for ECG signals because they contain sharp transitions and transient events. Wavelets can preserve localised features better than many traditional frequency-only methods.
Template-matching and classification algorithms are used to identify abnormal beats, arrhythmias, and morphological changes in long recordings.
In modern systems, ECG analysis may also involve machine learning and deep learning, especially for automated screening, rhythm classification, and anomaly detection. But even the most advanced models still depend on the same foundation: clean acquisition, robust preprocessing, and reliable feature structure.
What information it carries
ECG carries information about both rhythm and cardiac electrical morphology.
At the simplest level, it tells us whether the heart is beating regularly and how fast. But it goes much further than that. The waveform shape and timing relationships can reveal conduction delays, ectopic beats, atrial or ventricular arrhythmias, ischemic patterns, chamber enlargement, and other abnormalities.
It also carries indirect system-level information. For example, variations in the RR interval can be used to study autonomic nervous system activity through heart rate variability (HRV) analysis. In long-term monitoring, ECG can reveal transient events that a snapshot examination might miss.
So the ECG is not just a heartbeat counter. It is a compact electrical summary of how well the cardiac system initiates, conducts, and recovers from each beat.
Why it became famous
ECG became famous because it turned an invisible physiological process into a visible, interpretable signal with immediate medical value.
Few signals have had such a strong impact across both science and practice. It gave medicine a repeatable, non-invasive way to observe cardiac activity. It became standard in hospitals, ambulances, intensive care units, outpatient clinics, wearable devices, and consumer health technology. Even people with no technical background recognise the characteristic spike pattern as a symbol of life, urgency, and monitoring.
From an engineering standpoint, it also became famous because it is a perfect example of a real-world signal: structured, meaningful, noisy, and critical. It is simple enough to introduce fundamental concepts, yet rich enough to support serious research and advanced algorithm development.
The lesson
The ECG teaches an essential lesson about signals: a waveform is never just a waveform. It is a physical process translated into measurements, shaped by sensors, corrupted by noise, and decoded through analysis.
That means signal processing is not an optional extra. It is the bridge between raw acquisition and useful interpretation.
ECG also reminds us that the most valuable signals are often not the largest or cleanest ones. A few millivolts measured on the skin can contain life-saving information when the right methods are used to extract it. That is the heart of biomedical signal processing: finding structure in complexity, and turning subtle measurements into actionable insight.
The electrocardiogram became iconic because it does exactly that. It transforms the hidden language of the heart into a signal we can measure, process, and understand.
