Every so often a report says that a study „showed what happens in the brain“ during touch, during a conversation between therapist and client, or when two people „tune in to each other“. Often the method behind it is fNIRS — functional near-infrared spectroscopy. Before we judge how relevant the interpretations of its results are — especially in newspaper and popular-science articles — we need to understand what the method actually measures and how it works physically. And within what limits and with what precision. Only then do we have at least a sense of how far the proxy data the method delivers speak about the phenomenon the study investigates and describes.

What it measures: blood, not nerves; change, not level

A cap with light sources and detectors is placed on the head. Near-infrared light passes through skin and skull further than one would expect — Frans Jöbsis described this transparency of tissue as early as 1977. Part of the light is absorbed on the way, mainly in blood: haemoglobin with and without oxygen absorbs different wavelengths differently. From how much light comes back at two wavelengths one computes how much oxygenated and deoxygenated blood under the cap has changed.

Two things follow that get lost in popular reports:

  • fNIRS does not measure nerve-cell activity, but blood supply. That an „active“ part of the brain receives more oxygenated blood is a well-supported relation, but it is an inference, not a measurement. Animal experiments have also shown that the direction of change in deoxygenated haemoglobin is set not only by whether a region is „active“ but also by what happens in the veins (Hoshi et al. 2001).
  • What is measured is a change from baseline, not an absolute value. To say exactly how much blood was added, one must know how far the light travelled inside the head. Common instruments cannot measure that and plug in an estimate. The founding 1988 paper (Delpy et al.) measured that pathlength in a rat. The result is therefore more „how much more, in units that may differ between people“ than an exact number.

Where it fits

Where magnetic resonance cannot go. The measurement is harmless, tolerates body movement, the instrument is highly portable, and it can be used from newborns to the elderly and outside the laboratory (Pinti et al. 2018). That is why it is used with infants, during walking, in conversation between two people, during manual work — exactly the situations that interest body therapy.

Where the catch is

On its way to the brain and back, the light also passes through the scalp and skull. And the scalp is richly supplied with blood and responds to breathing, pulse, blood pressure and arousal. fNIRS experts who develop the method themselves decompose the measured signal into six components according to where it comes from (brain or head surface), whether it is related to the task, and whether it is neural activity or general physiology — and the measurement targets only one of them; the others, in their words, are noise that confounds (Tachtsidis and Scholkmann 2016).

Importantly, these confounds are not random. Speaking, mental arithmetic or raising the arms demonstrably change breathing and blood pressure — exactly when the „brain response“ is supposed to be measured. And the typical picture of „activation“ (more oxygenated, less deoxygenated blood) is, according to the same authors, not sufficient as evidence that no such distortion took place.

With measurements of two people at once (so-called hyperscanning — „brains in sync“) this is particularly delicate: two people in conversation often share rhythms of breathing, speech and arousal. Tachtsidis and Scholkmann therefore recommend naming such „connectivity“ honestly — as vascular connectivity, the sum of what happens inside and outside the brain.

What to ask

One question is enough, and it is not ours: it stands in the checklist for publications issued by the international fNIRS society itself (Yücel et al. 2021). How did the authors separate the influence of the scalp and general physiology (breathing, pulse, pressure) from the signal of the brain?

A good answer is, for example: they used short channels that measure only the head surface and subtracted it from the signal; they recorded breathing and pulse and included them in the analysis; or they had an instrument that can resolve depth. If a study is silent about this, it does not mean it is bad — it means we do not know from it how much of the „brain response“ is brain.

What do we take from this in practice? Three conclusions, each modest and well supported.

  1. fNIRS is a real and useful method for situations where a person moves, talks or touches — the ones that matter for body therapy.
  2. What it measures is blood supply under the cap, not a thought or an emotion. The link to what the brain does is an inference that needs control of surface and general influences.
  3. With reports of „what happens in the brain during touch or conversation“, ask that one thing — how breathing, pulse and scalp were filtered out. It is the question the field itself asks.

We are not yet writing about what fNIRS has said about body therapy: we have not yet examined such a study in detail. When we have, we will write about it separately.

References

  • Jöbsis, F. F. (1977). Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science, 198(4323), 1264–1267. https://doi.org/10.1126/science.929199 · PMID 929199
  • Delpy, D. T. et al. (1988). Estimation of optical pathlength through tissue from direct time of flight measurement. Physics in Medicine and Biology, 33(12), 1433–1442. https://doi.org/10.1088/0031-9155/33/12/008 · PMID 3237772
  • Hoshi, Y., Kobayashi, N., Tamura, M. (2001). Interpretation of near-infrared spectroscopy signals: a study with a newly developed perfused rat brain model. Journal of Applied Physiology, 90(5), 1657–1662. https://doi.org/10.1152/jappl.2001.90.5.1657 · PMID 11299252
  • Tachtsidis, I., Scholkmann, F. (2016). False positives and false negatives in functional near-infrared spectroscopy: issues, challenges, and the way forward. Neurophotonics, 3(3), 031405. https://doi.org/10.1117/1.NPh.3.3.031405 · PMC4791590
  • Pinti, P. et al. (2018). The present and future use of functional near-infrared spectroscopy (fNIRS) for cognitive neuroscience. Annals of the New York Academy of Sciences, 1464(1), 5–29. https://doi.org/10.1111/nyas.13948 · PMC6367070
  • Yücel, M. A. et al. (2021). Best practices for fNIRS publications. Neurophotonics, 8(1), 012101. https://doi.org/10.1117/1.NPh.8.1.012101 · PMC7793571

Our reference documentation (working versions in .md): fNIRS substrate (neuro-assessment/nirs/SUBSTRAT-NIRS-v0.1-EN.md) with reading notes on both full texts; physics-of-measurement assessment, Part 2 §3c. Currently in a private working repository; publication is in preparation. How well we know what we know is summarised on [What we know and how well](/evidence/).

Forthcoming. A technical article for mim2000.cz on where fNIRS came from — from the transparency of tissue to infrared light to today’s instruments — and what can and cannot be said about measuring two people at once.

Method and limits. The text rests on abstracts of the founding papers and on two fully read methodological texts (Yücel 2021 only in part). These materials do not rest on studies motivated by or grounded in body psychotherapy, nor on studies of clients of any of its schools; they are a methodological foundation not tied to therapy as such. This text was written on 12 Oct 2026 with the Claude Opus language model in Cowork mode; oracle: the sources listed above; domain: fNIRS literature valid and current as of 12 Oct 2026.