Mind the Cap - When the Experimental Limit Was in the Researcher, Not the Participant

Electroencephalography is often presented as a relatively accessible, noninvasive, and broadly applicable technique.

But accessible to whom?

This question becomes especially important when we observe that Black participants remain underrepresented in EEG research, while a methodological justification is repeatedly invoked: certain hair types, volumes, and styles would make cap placement more difficult and compromise data quality.

For years, this explanation seemed reasonable enough to shape procedures, exclusions, and even recommendations made to participants themselves.

But Jen Lewendon, İlayda Özdemir, Anna Binabdullah, and Anne Maass did something scientifically more interesting:

they decided to test the justification itself.

Published in 2026 in Psychophysiology, the article Mind the Cap: Inclusivity Gaps in EEG Research asks whether race and gender actually produce measurable differences in EEG data quality — or whether part of the difficulty attributed to the participant was, in fact, located in laboratory experience, methodological design, and the expectations of those collecting the data.

This is an excellent scientific question because it shifts the location of the problem.

When a methodological assumption becomes a fact without being sufficiently tested

Previous reviews had already drawn attention to the systematic exclusion of people with curly, coiled, dense, or high-volume hair from psychophysiological methods that require direct contact with the scalp.

Equipment manuals, training images, and classical protocols have historically been developed around a limited range of hair phenotypes. This can lead operational difficulties to be interpreted as natural properties of participants when they may, at least in part, reflect technologies and practices poorly adapted to human diversity.

The problem becomes even greater when this assumption begins to reinforce itself.

Researchers have little experience with certain hair types.

That lack of experience is interpreted as technical difficulty.

Technical difficulty reduces recruitment.

Lower recruitment further reduces laboratory experience.

And then the absence of participants begins to look like confirmation that the method “does not work well” in those bodies.

The merit of Lewendon and colleagues was to interrupt this cycle with a very simple question:

are the data actually worse?

The elegance of the design lies in comparing different systems

The authors analyzed data obtained using two types of EEG systems.

In the dry-electrode system, 60 participants were included: 30 women, 30 men, 30 Black participants, and 30 White participants.

In the wet-electrode system, previously collected data from 53 participants were analyzed: 27 Black and 26 White participants, including 36 women and 17 men.

This combination is particularly important.

If the effect had appeared only with one specific type of electrode, it would have been possible to argue that the conclusion depended strongly on that technology.

Instead, the researchers evaluated the question across different systems and used five measures of data quality:

• number of bad electrodes;

• quality of ICA decomposition;

• artifact rejection rate;

• baseline standard deviation;

• standardized measurement error.

The question therefore no longer depended on the researcher’s subjective impression.

“It was difficult to fit the cap” would not be equivalent to “the data were poor.”

That distinction is central.

The result was uncomfortably simple

The authors found no evidence of poorer data quality in Black participants, across any of the five evaluated metrics or either EEG system.

This does not mean that cap placement was always equally easy.

Qualitative notes from researchers, available for the dry-electrode system, suggested that preparation could require more effort in some Black participants.

But this operational difficulty did not translate into measurably poorer final data quality.

The distinction between “it was more difficult for me to do” and “the resulting data were worse” may be the most important contribution of the article.

Because those statements are not equivalent.

And a science that confuses the two risks transforming the researcher’s lack of experience into a characteristic of the participant.

When the method produces the exclusion it later attributes to the Body

Here we begin to approach the concept of Body-Territory.

Not because the study was designed to investigate it.

But because it shows, in a very concrete way, that what we call a “limitation of the body” may actually be produced in the relationship between body, equipment, protocol, technical experience, and expectations.

The person arrives at the laboratory with a particular type of hair.

The cap has a particular architecture.

The electrode has a particular shape.

The researcher has a particular training history.

The protocol establishes a particular sequence.

The institution has accumulated a particular form of experience.

And the result that emerges belongs to the relationship among all these elements.

If the system encounters difficulties, it is premature to conclude that the problem “lies in the participant.”

Perhaps it lies in the relationship.

The Body does not arrive at the experiment incorrectly

This deserves to be stated clearly:

the Body does not arrive at the experiment incorrectly.

The experiment may have been designed around a narrower range of bodily variation than the one that exists in the world.

This inversion may seem small, but it profoundly changes scientific responsibility.

Instead of asking:

“how do we make this participant fit our method?”

we can ask:

“is our method capable of adequately encountering the diversity of Bodies it intends to represent?”

The article’s own impact statement is particularly direct: the authors argue that responsibility for greater diversity in EEG research lies with researchers, and that inexperience with the hair types of Black participants should not automatically be treated as a barrier to inclusion.

Another result appeared — and it also requires a question

Interestingly, although the authors found no poorer data quality associated with race, they observed lower quality metrics in women compared with men across measures from both systems.

This finding should not be quickly transformed into a biological explanation about “female differences.”

It should produce another question.

What exactly is generating this difference?

Hair length?

Volume?

Hair products?

Electrode contact?

Cap architecture?

Preparation differences?

Some other unmeasured variable?

The study itself demonstrates why we must be careful before turning a methodological association into a property of the Body.

The answer may lie in an interaction that has not yet been described.

Inclusion is not only a moral issue — it is also a matter of scientific validity

The underrepresentation of particular groups creates more than a question of fairness.

It creates an epistemological problem.

If we build models of neural activity, ERP responses, biomarkers, or EEG norms using systematically restricted samples, to what extent can we later call those conclusions “human”?

Recent studies reinforce that EEG data of equivalent quality can be obtained in racially diverse samples when appropriate preparation, open communication, and inclusive procedures are used. In 2026, Farkas and colleagues showed equivalent data quality in Black women using contemporary recruitment and preparation practices. Another study using mobile EEG and dry electrodes also reported comparable data loss across racial groups during seated and standing tasks.

The problem, therefore, may be far less inevitable than it once seemed.

5D Consciousness: what we measure also depends on how we manage to encounter it

Within our proposal of 5D Consciousness, what is materially happening in the Body-Territory does not depend on the instrument being able to measure it.

EEG does not create brain activity.

It transduces part of it.

But what we manage to transduce depends on the relationship between biological system and measurement system.

This distinction is fundamental.

When an electrode records poorly, it does not mean that neural activity has stopped existing.

It means that our interface with that activity is limited.

Therefore, before attributing a difference to the Body, we need to ask:

are we observing a difference in the phenomenon, or a difference in our ability to measure it?

That question goes far beyond EEG.

It applies to fNIRS.

To neuroimaging.

To wearable sensors.

To algorithms.

To databases.

And also to artificial intelligence.

Those who are absent from the data are also absent from the models

There is another contemporary consequence.

Today’s neurophysiological data can feed statistical models and, increasingly, artificial intelligence systems.

If certain Bodies appear less often in data collection, they appear less often in training data.

If they appear less often in training, the system learns a less complete representation of human diversity.

And when that model returns to the world, it may be less accurate precisely for those who were underrepresented from the beginning.

The problem therefore stops being only:

“who was able to wear an EEG cap?”

It becomes:

“who participated in the computational definition of what we will call normal?”

Perhaps the limit was not in the Body

Lewendon and colleagues deserve recognition because they did not accept a methodological explanation simply because it sounded plausible.

They turned it into a hypothesis.

They measured it.

They compared it.

And they found something that shifts responsibility.

Data collection may be more labor-intensive in some cases.

But that does not necessarily mean poorer data.

The researcher’s effort is not a measure of the participant’s neural data quality.

This distinction seems obvious once stated.

But perhaps many scientific changes happen exactly this way: when someone decides to test what everyone had already begun to treat as known.

The study therefore enables a question that goes far beyond the EEG cap:

how many differences attributed to the Body were produced, amplified, or simply made invisible by the way we chose to measure it?

Within Body-Territory, the instrument also enters the territory of the encounter.

It is not neutral.

It has architecture.

It has limits.

It has history.

And it has users with different levels of experience.

Perhaps, then, before asking why a particular Body “does not work” in our experiment, science needs to learn to ask:

what Body did we imagine when we built the experiment?

And perhaps even more importantly:

how many other Bodies were left outside because we confused the limits of our method with their limits?

References

Lewendon, J., Özdemir, İ., Binabdullah, A., & Maass, A. (2026). Mind the Cap: Inclusivity Gaps in EEG Research. Psychophysiology, 63(8), e70371. https://doi.org/10.1111/psyp.70371

Farkas, A. H., et al. (2026). Open Communication Can Lead to Equivalent EEG Data Quality for Black Women: Multilevel Modeling Interindividual Differences on Emotional Scene and Face Perception. Psychophysiology. https://doi.org/10.1111/psyp.70343

Webb, E. K., Etter, J. A., & Kwasa, J. A. (2022). Addressing racial and phenotypic bias in human neuroscience methods. Nature Neuroscience, 25, 410–414.

Etter, J. A., et al. (2022). Hair me out: Highlighting systematic exclusion in psychophysiological methods and recommendations to increase inclusion. Frontiers in Human Neuroscience, 16, 1051400.

Inclusive mobile brain-body imaging achieves equivalent EEG data quality across racial groups. (2026). Study using inclusive dry-electrode Mobile Brain/Body Imaging procedures in racially and ethnically diverse older adults.