Galvanic Skin Response

Abstract

Galvanic skin response (GSR), which MeSH indexes among the psychological techniques used to read the body's state, is the transient change in the electrical conductance of the skin produced by the eccrine sweat glands under control of the sympathetic branch of the autonomic nervous system. Because those glands receive no parasympathetic supply, the signal is an unusually clean peripheral index of sympathetic arousal, which is why it became one of the founding measures of psychophysiology. The recorded signal separates into a slowly drifting tonic level and fast phasic responses that follow salient or emotionally significant events. This article sets out the sweat-gland mechanism, the tonic and phasic components and how they are measured and decomposed, the orienting response and its habituation, and the modern move to wearable, ambulatory recording. Interactive demonstrations reproduce the decomposition, the response shape, and habituation.

Keywords: galvanic skin response, electrodermal activity, skin conductance, sympathetic arousal, orienting response

Few physiological signals have carried as many names as this one. The same measurement has been called the galvanic skin response, the psychogalvanic reflex, the skin conductance response, and, in the modern standardised vocabulary, electrodermal activity. The names differ but the referent is constant: a change in how easily the skin conducts a small electrical current, driven almost entirely by how much the sweat glands are secreting at that moment. Because sweating of the palms and fingertips tracks emotional and cognitive engagement rather than thermal load, the signal offers a continuous, involuntary readout of arousal that a person cannot easily suppress (Boucsein, 2012).

The Electrodermal Signal

The skin's ability to conduct electricity is governed by its sweat glands. Human skin carries two to four million eccrine sweat glands, densest on the palms and soles, and each behaves like a variable resistor in parallel: as a gland fills with electrolyte-rich sweat, a conductive column rises through the resistive outer epidermis and the skin conducts more readily (Boucsein, 2012). Applying a small, imperceptible voltage across two electrodes on the fingers or palm and measuring the resulting current therefore yields a continuous estimate of sweat-gland activity, conventionally expressed as conductance in microsiemens rather than as its reciprocal, resistance.

What makes the measure so useful to psychology is the innervation. Eccrine sweat glands are controlled by the sympathetic branch of the autonomic nervous system through cholinergic fibres, and they receive no parasympathetic supply at all. Every other classic autonomic index, heart rate chief among them, reflects a tug-of-war between sympathetic and parasympathetic drive; electrodermal activity does not. A rise in conductance can be read as a rise in sympathetic outflow with none of that ambiguity (Critchley, 2002). The central origin of that outflow is not a single structure: electrodermal responses are shaped by the amygdala, which drives responses to threat and emotional salience, and by the prefrontal cortex, which modulates them, so the peripheral signal reflects the activity of the brain systems that assign significance to events, and it forms part of the bodily arousal that interoception makes available to feeling (Critchley, 2002).

Tonic and Phasic Activity

A raw electrodermal recording is not one signal but two superimposed. The slowly changing baseline, the skin conductance level (SCL), drifts over tens of seconds to minutes and indexes general, sustained arousal; it can wander from below two microsiemens in a drowsy, relaxed person to well above twenty in an agitated one, and its absolute value varies so much between individuals that it is usually interpreted only as a within-person change. Riding on top of it are the fast skin conductance responses (SCRs), transient bumps that rise within one to five seconds and then decay (Dawson et al., in Boucsein, 2012).

Phasic responses come in two kinds distinguished only by what precedes them. An event-related SCR follows an identifiable stimulus within a short response window, conventionally one to three seconds of latency, and is attributed to that stimulus. A nonspecific SCR (NS-SCR) appears with no external trigger; the rate of these spontaneous fluctuations, counted per minute, is itself a tonic index of arousal that rises with anxiety and effort. The distinction is operational rather than physiological: the same sweat-gland burst is called event-related or nonspecific purely according to whether a stimulus can be assigned to it within the window.

Measuring and Decomposing the Signal

There are two ways to derive the signal, distinguished by whether current is imposed on the skin. Exosomatic measurement, by far the more common, passes a small external voltage or current between two electrodes and records conductance; endosomatic measurement records the skin's own electrical potential with no applied current. This division is as old as the measurement itself. The exosomatic method descends from Féré, who in 1888 passed a current across the skin and saw its resistance fall under emotional and sensory stimulation; the endosomatic method descends from Tarchanoff, who in 1890 recorded a potential from the skin with no applied current at all. Through the early twentieth century the phenomenon was studied as the psychogalvanic reflex, notably in Jung's word-association experiments, before the modern vocabulary settled on electrodermal activity (Neumann & Blanton, 1970). The modern standard is a constant-voltage exosomatic method reporting conductance in microsiemens, a convention argued for on the grounds that conductance, unlike resistance, is linearly related to the number of active sweat glands (Lykken & Venables, 1971). Sites, electrode paste, and units were standardised by successive committee reports precisely so that values could be compared across laboratories (Fowles et al., 1981; Boucsein et al., 2012).

Decomposing the phasic responses is harder than it looks, because a fresh response often begins before the previous one has recovered, so overlapping SCRs sum into a shape that peak-picking mis-scores. Two families of method address this. Deconvolution treats the measured signal as the sweat-gland driver convolved with a standard response shape, and recovers a sparse driver whose discrete impulses are the underlying responses, giving a continuous phasic measure that separates overlapping events (Benedek & Kaernbach, 2010). Model-based analysis goes further and treats the response as the output of a biophysical, generative model whose parameters are estimated by inversion, so that the quantity reported is an estimate of the underlying sympathetic input rather than of the skin signal itself (Bach & Friston, 2013). Convex optimisation offers a third route, jointly estimating the tonic and phasic components and the sparse driver as a single well-posed problem (Greco et al., 2016).

The Orienting Response and Habituation

When a novel stimulus of any modality arrives, an organism turns toward it and its physiology shifts to take it in; this orienting response includes a reliable skin conductance response, and the electrodermal channel is one of the standard ways of measuring it. The orienting response is not a reflex to a particular stimulus but a response to novelty and significance, which is why it appears whenever a stimulus is unexpected or matters to the task, and why it is bound up with the allocation of attention (Lykken & Venables, 1971).

Its most studied property is what happens on repetition. Present the same stimulus again and again and the orienting SCR shrinks, trial by trial, in an approximately geometric decline until it disappears: the response has habituated. Habituation is a form of non-associative learning, the nervous system's way of ceasing to spend arousal on a stimulus that has proved inconsequential. Two diagnostic features separate it from simple fatigue. First, changing the stimulus restores the full response, a recovery called dishabituation that shows the decline was specific to the repeated stimulus, not a depletion of the sweat glands. Second, the rate of habituation is itself informative: fast habituators and slow habituators differ in temperament and in clinical status, and the count of trials to habituation is a standard individual-difference measure. Because the SCR here is an index of a conditioned or unconditioned autonomic reaction, the same recording method underpins the study of autonomic learning in classical conditioning.

Worked Example

Consider a habituation session in which the same neutral 70-decibel tone is presented every thirty seconds. The first tone evokes an orienting skin conductance response of 0.80 microsiemens. Each identical repetition evokes a smaller response, declining geometrically as SCR(n) = SCR(1) x e^(-k(n-1)) with a habituation rate of k = 0.35. The orienting response is scored as habituated once its amplitude falls below the conventional 0.05 microsiemen detection criterion. Setting 0.80 x e^(-0.35(n-1)) below 0.05 and solving gives n - 1 greater than ln(0.80 / 0.05) / 0.35 = ln(16) / 0.35 = 2.773 / 0.35 = 7.92, so n = 9. The ninth tone is the first whose response falls below criterion: 0.80 x e^(-0.35 x 8) = 0.049 microsiemens, just under the line, while the eighth still clears it at 0.069. The Orienting and Habituation demonstration reproduces this exactly: stepping through the trials drives the response beneath the criterion line at trial nine, and presenting a novel stimulus restores the full response through dishabituation.

Discussion

The galvanic skin response occupies a peculiar place in psychology: it is at once one of the oldest instrumental measures of mind, dating to the turn of the twentieth century, and one of the most actively re-engineered. Its longevity rests on the anatomical accident that makes it valuable, the purely sympathetic innervation of the eccrine glands, which gives psychology a rare window onto autonomic arousal uncontaminated by opposing parasympathetic drive. That same purity is also its limit. The signal reports that arousal has changed, not why: a startle, a lie, a hard sum, and a sudden interest all raise conductance, and nothing in the waveform distinguishes them. This is the deep reason the polygraph remains contested. It measures arousal faithfully and reads guilt into it only by inference, and inference from an undifferentiated arousal signal is exactly what a signal detection analysis shows to be error-prone when the base rates and the overlap of innocent and guilty arousal are taken seriously.

The measure's methodological history is a case study in earning comparability. Because absolute conductance depends on electrode placement, skin hydration, temperature, and the individual, early results were nearly impossible to pool, and the field's response was not a new sensor but a series of committee reports fixing units, sites, and procedures so that a number from one laboratory meant the same as a number from another. The subsequent shift from hand-scored peaks to deconvolution and model-based estimation is the same impulse pursued mathematically: to recover the underlying sympathetic events from a signal that blurs them together, and to report a quantity that means the same thing every time it is computed.

Current Directions

The most visible change is where the signal is now recorded. Electrodermal activity has moved off the laboratory palm and onto the wrist, carried by inexpensive wearable sensors that record continuously for days outside any laboratory. This has opened ambulatory and naturalistic study of arousal but has also imported new problems: wrist recordings are noisier than palmar ones, motion and loss of skin contact produce artefacts that mimic responses, and long unsupervised recordings need automated quality control, for which transparent, shareable procedures have been proposed and validated (Kleckner et al., 2018). A systematic review of these innovations catalogues the new electrode materials, recording sites, and indices, including frequency-domain measures of sympathetic tone intended to be more robust than the traditional time-domain scores (Posada-Quintero & Chon, 2020). In parallel, the analysis side continues to mature: convex-optimisation and model-based pipelines aim to make the decomposition reproducible and to tie the reported quantity to an explicit generative model rather than to a scoring convention (Greco et al., 2016; Bach & Friston, 2013). The open question uniting these fronts is whether the cheap, ambulatory signal can be scored to the standard the century-old palmar method reached, so that field data and laboratory data can finally be compared on equal terms.

Common Misconceptions

A rising skin conductance response reveals a specific emotion.

It reveals only that sympathetic arousal has risen. Fear, excitement, mental effort, and surprise all produce responses, and the waveform does not distinguish them; emotional meaning must come from the eliciting context, not the signal alone (Critchley, 2002).

The measure is a lie detector.

Electrodermal activity is one channel of the polygraph, but it indexes arousal, not deception. Because innocent and guilty arousal overlap, inferring a lie from a response is a detection problem with a real false-positive rate, not a direct readout of honesty (Boucsein, 2012).

Skin conductance is driven by general sweating and body heat.

The palmar and plantar eccrine glands that dominate the signal respond to psychological and sensory significance rather than to thermoregulatory need, which is why the measure tracks arousal and not room temperature (Boucsein, 2012).

A declining response across trials means the sensor is failing.

A geometric decline over repeated identical stimuli is habituation, a genuine learning phenomenon, not equipment drift. The proof is dishabituation: a novel stimulus instantly restores the full response (Lykken & Venables, 1971).

Glossary

Arousal.

A state of heightened physiological and psychological activation; sympathetic arousal is what electrodermal activity most directly indexes.

Autonomic Nervous System.

The division of the nervous system controlling involuntary organ function, comprising sympathetic and parasympathetic branches; only the former innervates the eccrine sweat glands.

Deconvolution.

A signal-processing method that recovers the underlying sweat-gland driver from the measured response by inverting a standard response shape, separating overlapping phasic responses.

Dishabituation.

The restoration of a habituated response when the stimulus changes, demonstrating that the prior decline was stimulus-specific learning rather than fatigue.

Eccrine Sweat Gland.

The most numerous type of sweat gland, densest on palms and soles, whose secretory activity changes skin conductance and generates the electrodermal signal.

Electrodermal Activity.

The modern standardised umbrella term for all electrical activity of the skin, encompassing the galvanic skin response and its tonic and phasic components.

Endosomatic Measurement.

Recording the skin's own electrical potential with no externally applied current, as distinct from exosomatic measurement.

Exosomatic Measurement.

Recording skin conductance by passing a small external voltage or current between two electrodes; the standard modern method.

Habituation.

The progressive decline of a response, such as the orienting SCR, to a repeated inconsequential stimulus; a form of non-associative learning.

Microsiemens.

The conventional unit of skin conductance, one millionth of a siemens; the reciprocal of resistance and linearly related to the number of active sweat glands.

Nonspecific Skin Conductance Response.

A phasic response occurring with no identifiable eliciting stimulus; its rate per minute is a tonic index of background arousal.

Orienting Response.

The set of physiological changes, including a skin conductance response, evoked by a novel or significant stimulus as attention is directed toward it.

Phasic Activity.

The fast component of electrodermal activity: discrete skin conductance responses rising within a few seconds and decaying, tied to specific events.

Polygraph.

A multi-channel recorder combining electrodermal, cardiovascular, and respiratory measures; it indexes arousal, from which deception can only be inferred.

Skin Conductance Level.

The slowly changing tonic baseline of electrodermal activity, indexing sustained arousal and interpreted as a within-person change.

Skin Conductance Response.

A discrete phasic increase in conductance following a stimulus or occurring spontaneously; the elementary unit of phasic electrodermal activity.

Sympathetic Nervous System.

The autonomic branch mediating arousal and the fight-or-flight response; its cholinergic fibres are the sole autonomic control of the eccrine sweat glands.

Tonic Activity.

The slow, sustained component of electrodermal activity, captured by the skin conductance level and the nonspecific response rate.

Key Researchers

Mathias Benedek. Professor of psychology at the University of Graz; co-developed the continuous deconvolution method for decomposing phasic electrodermal activity that underlies much modern SCR scoring. Faculty Page - ORCID - Google Scholar

Wolfram Boucsein (1944-2012). Professor of physiological psychology at the University of Wuppertal; author of the standard monograph Electrodermal Activity and lead of the 2012 measurement-recommendations committee. Faculty Page - Wikipedia - Wikidata

Hugo D. Critchley. Chair in psychiatry at Brighton and Sussex Medical School, University of Sussex; mapped the brain systems that generate and regulate electrodermal responses and their role in interoception. Faculty Page - ORCID - Google Scholar - Wikipedia

Michael E. Dawson. Professor emeritus of psychology at the University of Southern California; co-author of the definitive handbook treatment of the electrodermal system and its measurement. Faculty Page - Google Scholar

Don C. Fowles. Professor emeritus of psychological and brain sciences at the University of Iowa; lead author of the 1981 committee report that standardised electrodermal measurement. Faculty Page

Hugo F. Posada-Quintero. Assistant professor of biomedical engineering at the University of Connecticut; authored the systematic review of modern electrodermal data-collection and signal-processing methods. Faculty Page - ORCID - Google Scholar

Frequently Asked Questions

What is the galvanic skin response? It is the change in the electrical conductance of the skin caused by eccrine sweat-gland activity, an involuntary index of sympathetic arousal also known as electrodermal activity or skin conductance (Boucsein, 2012).

Why is skin conductance a good measure of arousal? The eccrine sweat glands are controlled only by the sympathetic nervous system, with no opposing parasympathetic input, so a rise in conductance reflects sympathetic arousal without the ambiguity that affects heart rate (Critchley, 2002).

What is the difference between tonic and phasic electrodermal activity? Tonic activity is the slow skin conductance level indexing sustained arousal; phasic activity is the fast, discrete skin conductance responses tied to specific stimuli or occurring spontaneously (Boucsein et al., 2012).

Is the galvanic skin response a reliable lie detector? No. It measures arousal, not deception. Because innocent and guilty arousal overlap, inferring a lie from a response carries a genuine false-positive rate rather than reading honesty directly.

Why should conductance be reported in microsiemens rather than resistance? Conductance is linearly related to the number of active sweat glands, which makes it the more interpretable unit and the reason it was adopted as the measurement standard (Lykken & Venables, 1971).

What is habituation of the orienting response? When a novel stimulus is repeated, the orienting skin conductance response it evokes declines geometrically until it disappears; a change of stimulus restores it, a recovery called dishabituation (Lykken & Venables, 1971).

How are overlapping skin conductance responses separated? Deconvolution recovers a sparse sweat-gland driver from the signal, and model-based analysis estimates the underlying sympathetic input from a generative model, both separating responses that peak-picking would merge (Benedek & Kaernbach, 2010; Bach & Friston, 2013).

Can electrodermal activity be recorded outside the laboratory? Yes. Inexpensive wrist-worn sensors record continuously for days, though they are noisier than palmar recordings and require automated quality control to reject motion artefacts (Kleckner et al., 2018).

References