Willem Einthoven

THE CLAIM

Willem Einthoven made the rhythm of a living body measurable in absolute units

Willem Einthoven held the chair of physiology at the University of Leiden from the age of twenty five, and in 1901 he published a design that changed what a clinician could know. The paper was Un nouveau galvanometre, in the Archives Neerlandaises des Sciences Exactes et Naturelles (Einthoven 1901). The device was the string galvanometer. In 1924 the Nobel Prize in Physiology or Medicine went to him for the discovery of the mechanism of the electrocardiogram (Nobel Foundation 1924). Read that citation slowly. It does not say the invention of a machine. It says the mechanism. The committee was rewarding an explanation of what the line on the plate means, and the machine was only the means of getting there.

Ask what that commits you to. If the electrical state of a beating heart can be written on a photographic plate in millivolts and thousandths of a second, then the living rhythm of a body is no longer an impression held in a clinician's fingers. It is a quantity. It can be repeated, calibrated, argued over and compared between one hospital and another. Every later measure of autonomic state, including the beat to beat interval analysis now called heart rate variability, sits downstream of that single change. Einthoven did not build a heart monitor. He built a reason to trust a number taken from a living system.

BEFORE THE STRING

The electricity of the heart was known for decades before anyone could read it cleanly

Einthoven never claimed the discovery for himself, and his Nobel lecture is unusually careful about the credit (Einthoven 1925). He names Kolliker and Muller as the first to describe the current the heart develops at each systole. He then names the English physiologist Augustus D. Waller as the man who showed that those potential differences travel out to the surface of the body, and that with a sensitive instrument they can be observed in a living human being. Waller published the first human electrocardiogram in 1887 (Waller 1887), recorded with a capillary electrometer from Thomas Goswell, a technician at St Mary's Medical School in London. Einthoven watched Waller demonstrate the method at the First International Congress of Physiologists in Basel in 1889. He went home with a problem.

The problem was the instrument, not the phenomenon. A capillary electrometer reads the movement of a mercury meniscus, and mercury is heavy and slow. The recorded curve was a smeared version of the real potential change, so the true shape had to be reconstructed afterwards by calculation from the known properties of the tube. Einthoven spent years on that correction and said plainly in Stockholm that it took much time and stood in the way of practical electrocardiography. Force the implication. An instrument that distorts what it measures does not hand you a fact. It hands you a homework problem, and homework does not scale to a ward full of patients.

THE INSTRUMENT

A silvered quartz thread thinner than a wavelength of light turned current into visible motion

The string galvanometer is one idea executed with extreme care. A fine conducting thread is stretched across a magnetic field. Send current through it and it shifts sideways, at right angles to the lines of force, by an amount proportional to the strength of the current. Shine light across the thread, magnify its shadow, and photograph that shadow on a moving plate. That is the whole instrument. Einthoven started from the mirror galvanometer of Deprez and d'Arsonval and stripped out almost all of the moving mass, because the sensitivity he needed was governed by how little there was to move. He replaced the coil with a single silvered quartz filament held in the narrow gap between the poles of a strong electromagnet.

The numbers he gave in Stockholm are still startling. Ordinary working instruments used strings two or three microns thick. Strings of one tenth of a micron worked well, and he showed a photograph taken at a magnification of 1800 of a string he estimated at 0.04 micron, about fifteen times smaller than the wavelength of yellow light. A field magnet of roughly 20,000 gauss could be built without special difficulty. In a vacuum model made by the engineer W. F. Einthoven, a string two centimetres long and around 0.1 micron thick, magnified 1800 times, moved one millimetre in one hundredth of a second under a current of ten to the minus eleven amperes. Later accounts describe the early clinical machine as filling two rooms, weighing some six hundred pounds (Rivera-Ruiz 2008), needing water cooling and five people to run it, with the patient seated and both hands and one foot in jars of salt solution.

TENSION

Tension in the string set both the sensitivity and the speed of the reading

Einthoven's central design problem was a trade, and the variable that governed the trade was tension. Slacken the string and weaker currents will move it, which raises sensitivity. Slacken it and the movements also become slower, which is useless when the signal you want is a fast one. Tighten the string and it answers quickly but only to larger currents. He refused to quote raw current sensitivity for exactly that reason. Instead he defined normal sensitivity, the sensitivity measured at a fixed duration of deflection, and wrote it as a function of the current producing a given displacement and the natural frequency of the freely oscillating string. Sensitivity and response time were bound into one number.

He then set out what normal sensitivity actually depends on: the magnification, the field strength, the mass of the string, and the material the string is made from, with a constant of one three hundred and twentieth when the field is given in gauss, the masses in grams and the resistance in ohms. The engineering lesson is clean. You cannot raise gain without paying somewhere, and the cheapest place to pay is mass. This library reads that trade as a physical rehearsal of something the nervous system does, and the reading is ours rather than his. A system held under tension buys speed with tension and buys sensitivity with slack, and it cannot hold both at full value at once.

THE FLOOR

Below a certain thinness the string never stops moving

Push the design far enough and a limit appears that is not a fault of manufacture. Einthoven made strings so thin that thermal motion in the surrounding molecules kept them permanently in motion. He named the effect Brownian string vibrations and published a record of a string eighteen millimetres long, under 0.2 micron in diameter, with a resistance of 250,000 ohms, in high vacuum, field magnet off, circuit open. It still danced. Air does not damp these vibrations, because the movement of the air molecules is the cause of them. Tighten the string and the frequency rises while the amplitude falls, and the mean energy of the motion stays where it was.

He then worked out why the magnetic field cannot quiet them either. If the field damped Brownian vibrations while those vibrations still induced currents and warmed an external resistance, you would have a perpetual motion machine of the second kind. The resolution is that thermal motion of electrons in the circuit produces its own Brownian currents, and the two effects cancel. Einthoven treated this as a hard practical limit on the measurement of very weak currents. Read it as a general truth about measuring anything alive. There is a floor of restless motion under every recording, and the floor is not error. It is the signature of a system at temperature.

THE WIRE TO THE WARD

In 1905 Einthoven ran a telephone cable from the Leiden hospital to his laboratory

The galvanometer was almost immovable, and sick patients are hard to move. Professor Bosscha suggested linking the hospital of Leiden University to the physiological laboratory by wire, and the Societe Hollandaise des Sciences funded the trial. The overhead sections of the Leiden telephone system were useless, because swaying wires cut the magnetic field of the earth and threw the string about, so Einthoven had a twisted pair sheathed in lead and slung from a steel cable. Even when the wind moved the cable, the reading held. The run was about a mile, roughly 1.5 kilometres. Its resistance was 106 ohms and its capacity 0.075 microfarad, and the circuit was closed at the far end by the body of the patient, between one thousand and two thousand ohms.

The calibration is worth stating, because calibration is the point of the whole enterprise. A step of three millivolts moved the string thirty millimetres, so one millimetre of the record equalled one ten thousandth of a volt. The plate speed made one millimetre equal one thousandth of a second. The first long distance record was taken on 22 March 1905 from a healthy and vigorous man, current led from both hands, showing an R summit of two millivolts, and by his own account it differed in no way from the tracing made beside the galvanometer. The results were published in 1906 as Le Telecardiogramme in the Archives Internationales de Physiologie, volume 4, page 132 (Barold 2003). He also listed the costs honestly. The apparatus needed a special installation, took up much space, and could only be handled by trained staff.

THE ALPHABET

Five deflections named P, Q, R, S and T gave the heart a written vocabulary

In 1893, at a meeting of the Dutch Medical Association, Einthoven used the word electrocardiogram, and he later allowed that Waller may have used it first. In 1895, in Ueber die Form des menschlichen Electrocardiogramms in Pfluger's Archiv, he distinguished five deflections in the corrected curve and lettered them P, Q, R, S and T (Einthoven 1895). The choice of letters was a mathematician's habit rather than a mystery. Curves from the capillary electrometer had already been labeled with early letters of the alphabet, and a corrected curve needed a set of its own, so he took letters from the later part, passing over N and O because both are already spoken for in coordinate geometry (Hurst 1998).

The lettering held because it named something real. P is the atrial event. Q, R, S and T belong to the ventricles. In the 1906 telecardiogram paper he added U for a further peak seen after T in a pathological record, one that ended about half a second after the onset of systole. In the same paper he noted that an interval between atrial and ventricular systole beyond 0.2 second marks a state in which drugs such as digitalis become dangerous, and that the record makes it simple to follow the influence of a drug step by step. A century later those five letters run every cardiology report in the world.

THE TRIANGLE

Lead III equals Lead II minus Lead I

Einthoven standardized three leads: the first from the right and left hands, the second from the right hand and the left foot, the third from the left hand and the left foot. He then stated the relation between them. The potential variation in Lead III equals the difference between the variations recorded at the same moment in Leads I and II. By 1925 his laboratory could record all three at once with three strings, which turned the relation into a running check on the honesty of the instruments. If the arithmetic failed, the recording was wrong, and you knew it without leaving the room.

On 19 March 1912 he addressed the Chelsea Clinical Society in London, and the lecture appeared in the Lancet that year (Einthoven 1912). There he drew the three leads as the sides of an equilateral triangle with the heart at its center, so that the size and direction of the electrical event could be read as a single arrow. Ask what that construction assumes. It assumes you are not measuring the heart at all. You are sampling a field the heart generates in a conducting body, and the geometry lets you work backwards from the surface to the source. Every clinician who reads an axis today is using that assumption, whether or not the name Einthoven comes to mind.

BEYOND THE HEART

The same instrument wrote nerve, muscle, retina and skin

The Nobel lecture spends a long stretch on everything the string galvanometer recorded that was not a heart. Muscle currents. Nerve currents. Skin currents. The currents of sense organs. He shows a record taken from the peripheral stump of the left vagus nerve of a dog, running alongside a pneumogram of the animal's own breathing and a blood pressure curve. When the right vagus is stimulated the heart arrests, the heart waves vanish from the nerve record, and the respiratory waves carry straight on. He shows retinal currents from eyes held in darkness and then struck with a hundredth of a second of strong light. He shows the psychogalvanic reflex, led from the thumb and fingers of one hand, an acoustic stimulus producing a change in skin resistance where one scale division stands for twenty ohms.

None of this was decoration. The same instrument recorded heart sounds against the electrocardiogram and showed that the first sound and the ventricular complex begin together. It served as a wireless receiver, catching signals sent from Bandoeng in the Netherlands Indies across 12,000 kilometres with a six millimetre string tuned to a 7.5 kilometre wave at 40,000 oscillations a second. A one millimetre string could be tightened until its natural frequency reached 300,000 cycles a second. One device, one measured quantity, and it made no difference whether the source was a heart, a nerve trunk, a retina, a sweat gland or an antenna in Java.

WHAT IS DOCUMENTED

Keep what Einthoven demonstrated separate from what this page reads into it

Documented and not in dispute: the string galvanometer design of 1901, the lettering of 1895, the telecardiogram of 1905 and 1906, the three lead relation and the triangle of 1912, the vagal and retinal and skin recordings shown in Stockholm, and the Nobel Prize awarded for 1924. Three details deserve correction because they are repeated carelessly. Einthoven did not record the first human electrocardiogram. Waller did, in 1887, and Einthoven says so in his own lecture (Waller 1887). The string principle was anticipated by the French engineer Clement Ader in 1897 for telegraphy (Burchell 1987), and Einthoven cited him once he learned of the work, with the difference that the wire Ader used was about ten times thicker and would never have caught a heart. He also gave his Nobel lecture on 11 December 1925 rather than in 1924 (Einthoven 1925), because he was abroad when the prize was announced.

One further distinction matters more than the rest on this site. Einthoven did not study heart rate variability. He measured intervals with great precision, and he built the calibrated timebase that makes interval analysis possible, but the reading of beat to beat variation as an index of autonomic state came later, through fetal monitoring work in the 1960s and the international measurement standards published in 1996 (Task Force 1996). Calling him the ancestor of heart rate variability is our framing, not his claim. What he delivered was the precondition. Without absolute units and a trustworthy timebase, variability is only noise that nobody can compare.

EINTHOVEN AND THE MODEL

Einthoven built the instrument that lets tone be measured before the outcome is known

The Unified Model of Tone reads the electrocardiogram as founding equipment. The model holds that tone shows itself wherever a rhythm can be measured for more than its average. It shows in the variability of a signal rather than its mean. It shows in the coupling between two rhythms rather than either alone. It shows in how a system responds to a challenge and recovers from it, rather than how it sits at rest. Heart rate variability and its internal structure is one of the model's named windows, and every reading taken through that window runs back to the string galvanometer of 1901.

The gift was the calibrated unit rather than the tracing. Einthoven put the standard into one sentence in Stockholm. Given an instrument correctly built and fast enough for its purpose, a curve recorded anywhere in the world is directly comparable to any other curve. That comparability is exactly what a window onto tone requires. Without absolute units and a trustworthy timebase, beat to beat variation is noise that nobody can compare between two people or two decades.

What matters clinically is not the average but the structure of the variation. The model states the rule generally. Variability in the act of regulating is health, and variability in the thing regulated is dysregulation. High heart rate variability marks health because it is the trace of a controller adjusting continuously to what it meets. High blood pressure variability marks risk because it is drift in a value the controller is supposed to be holding. Blood pressure variability measured visit to visit or over twenty four hours predicts stroke and cardiovascular events above and beyond mean blood pressure, in a 2016 systematic review and meta analysis (Stevens 2016). Two patients with the same average pressure and different variability profiles face substantially different cardiovascular futures.

The consequence is that these readings can be recorded before the clinical outcome is known. That is what makes tone a measurable quantity rather than a label applied after the fact, and it is the whole force of Einthoven's calibration. A quantity recorded in two people this afternoon can be used to anticipate how differently the same event will land on each of them tomorrow. Einthoven measured the heart. The identification of the beat to beat trace as a reading of nervous system tone is the model's claim, not his.

The instrument also carries its lesson in its own physics. Tension decided how sensitive the string was, how quickly it answered, and how much of its own thermal restlessness reached the record. Those are the same relationships this library tracks in living tissue. Einthoven was building his instrument in the same decade that Santiago Ramon y Cajal was resolving the nervous system into separate cells. One gave the tissue its units. The other gave the signal its numbers. Einthoven closed his lecture by crediting many workers spread over the whole surface of the earth who had not been influenced by political boundaries. No school owns the nervous system, and no technique owns the measurement of it.

WHAT THE RECORD SHOWS

Einthoven left a dated chain of calibrated measurements

• 1895. In Ueber die Form des menschlichen Electrocardiogramms, Einthoven distinguished five deflections in the corrected curve and lettered them P, Q, R, S and T (Einthoven 1895). Those five letters still run every cardiology report in the world.
• 1901. Un nouveau galvanometre described the string galvanometer, whose normal sensitivity could be made a thousand times greater than that of the most sensitive mirror galvanometer then available (Einthoven 1901).
• 0.04 micron. Einthoven showed a photograph taken at a magnification of 1800 of a silvered quartz string he estimated at four hundredths of a micron, about fifteen times smaller than the wavelength of yellow light.
• 22 March 1905. The first long distance electrocardiogram traveled roughly 1.5 kilometres of lead sheathed cable from the Leiden hospital. Three millivolts deflected the string thirty millimetres, and one millimetre of plate equalled one thousandth of a second.
• 1906. Le Telecardiogramme appeared in the Archives Internationales de Physiologie, volume 4, page 132 (Barold 2003). It reported that an interval beyond 0.2 second between atrial and ventricular systole marks a state in which drugs such as digitalis become dangerous.
• 1912. Addressing the Chelsea Clinical Society on 19 March, Einthoven drew the three leads as an equilateral triangle with the heart at its center and stated that Lead III equals Lead II minus Lead I (Einthoven 1912).
• 2016. A systematic review and meta analysis found that blood pressure variability predicts stroke and cardiovascular events above and beyond mean blood pressure (Stevens 2016). That is the mirror image of heart rate variability, and it is why the model separates the regulator from the value being held.