This paper is especially useful because it moves the wearable question from “what can these devices measure?” to “which type of wearable is actually most useful when I care about signal quality, movement, raw data and comfort at the same time?”

That is exactly where my project is now.

I am currently between a ring and a wrist-worn form, and I am also trying to decide whether the final sensing architecture really needs EDA + PPG / heart rate + accelerometry, with skin temperature as a possible secondary signal.

This study does not give me one universal “winner”, but it does make the trade-offs between devices much clearer.

What the paper compared

Brandebusemeyer, Georgi and Arnrich compared four wearable systems:

• Shimmer3 GSR+
• Empatica E4
• Empatica EmbracePlus
• Google Pixel Watch 2

The first study compared Shimmer3 GSR+, Empatica E4 and Pixel Watch 2 across baseline, cognitive-load, startle, relaxation and physical-activity conditions.

A second study then compared the older Empatica E4 with its successor, EmbracePlus, across selected cognitive, affective and relaxation conditions.

The paper looked at:

• heart rate;
• HRV-related measures;
• electrodermal activity (EDA);
• skin temperature;
• motion robustness;
• and wearer comfort.

The most useful overall finding

All of the wearables could detect large physiological changes between clearly different conditions.

The devices were much less reliable at distinguishing subtle differences, such as smaller changes between cognitive-load levels.

That distinction is very important for my project.

It suggests a wearable is more defensible as a system that notices:

“something has changed”

than one that claims:

“I know exactly what state this person is in.”

Key lesson for my project

Wrist wearables can reliably notice larger physiological changes, but subtle states are harder to distinguish. That fits a check-in system much better than an automatic state classifier.

EDA looks particularly relevant

One of the strongest technical findings for me is that EDA provided more consistent task differentiation than PPG-derived measures under motion.

That matters because I have been questioning whether EDA deserves to remain in the sensing architecture.

This paper gives me a reason to keep it under serious consideration.

Both the Shimmer3 GSR+ and Empatica E4 captured clear EDA changes across conditions, including during physical activity. The Shimmer produced larger absolute responses and more SCR peaks, but that was partly because its electrodes were placed on the fingers, where sweat-gland density is higher.

So the larger signal does not automatically make it the better design choice.

Comfort changes the decision

The comfort results are particularly important for an airport application.

The Pixel Watch 2 received the highest positive comfort ratings in study 1, with 85% of participants rating it comfortable or very comfortable.

The Empatica E4 also performed well, with 80% positive comfort ratings.

For EmbracePlus, 16 of 18 participants reported that it did not disturb them during normal work activity.

By contrast, the Shimmer3 GSR+ was often rated uncomfortable because its finger electrodes interfered with natural hand movement.

That immediately makes the Shimmer less attractive for my project.

An airport traveller needs to:

• hold a phone;
• carry luggage;
• handle documents;
• open bags;
• use ticket gates;
• interact with security trays;
• and use their hands constantly.

Finger electrodes may produce a stronger EDA signal, but the form itself could create exactly the sort of burden I am trying to avoid.

Did the paper identify one device that is best for me?

Not definitively.

The paper itself concludes that device choice should depend on the research question, required transparency, signal type and expected movement conditions.

However, for my current project requirements, one device now stands out as the strongest practical candidate from this paper:

Empatica EmbracePlus — strongest current wrist candidate

Not because it “won” every measurement, but because it offers the strongest balance of the things I currently care about:

• wrist-worn and relatively unobtrusive;
• PPG / heart-rate sensing;
• EDA;
• skin temperature;
• 3-axis accelerometry;
• raw sensor access;
• HRV-related analysis;
• successor to the discontinued E4;
• and good reported comfort / unobtrusiveness.

Current design interpretation

If my final sensing architecture remains EDA + PPG / heart rate + accelerometry, EmbracePlus currently looks like the strongest all-round research wearable for prototyping the system.

But this is still an interpretation, not something the paper proves.

Why not the Pixel Watch 2?

The Pixel Watch 2 is interesting because it performed well for heart rate and skin-temperature trends in dynamic settings, and it received the strongest comfort ratings.

However, this paper exposed an important limitation.

Although the Pixel Watch 2 contains PPG and EDA sensors, the researchers found that the exported PPG and EDA data were not interpretable enough for the analysis, so they could not use them for HRV or detailed EDA metrics.

That makes the Pixel attractive as an everyday consumer object, but weaker for a project where I need to understand and explain exactly what the physiological sensing is doing.

So for my current design:

Pixel Watch 2 = strongest consumer comfort / processed trend option

but

EmbracePlus = stronger research / transparent sensing option

Why not the Shimmer3 GSR+?

The Shimmer3 GSR+ has technical strengths.

Its higher sampling rates and finger electrodes produced strong EDA signals and allowed more detailed phasic analysis.

However, almost half of the participants rated it uncomfortable, largely because of the finger sensors.

For a laboratory study that may be acceptable.

For an airport accessibility wearable, it is a major problem.

The strongest signal is not automatically the best design if the measurement method interferes with ordinary behaviour.

Why not the Empatica E4?

The E4 still performed reasonably well and has been widely used in research.

However, it has been discontinued and replaced by EmbracePlus.

The paper's exploratory comparison suggests the newer EmbracePlus can provide comparable overall performance, while remaining a current platform.

That makes the E4 more useful as a research precedent than as the device I would now design around.

Did any of these beat the Oura Ring?

This paper cannot answer that directly because Oura was not included in the comparison.

So I should not say that any of these devices “beat” Oura based on this study.

However, there is an important design difference.

The Oura Ring study I read earlier showed that a ring can provide:

• heart rate;
• inter-beat interval / HRV-related information;
• skin temperature;
• and movement/activity data

in a very discreet form.

What it did not give me was the EDA channel I am currently considering.

So the comparison is becoming:

Oura-style ring

Advantages

• extremely compact
• discreet
• familiar consumer object
• cardiac + temperature + movement data
• strong precedent for long-term personal baselines

Current limitation

• no EDA in the paper I reviewed

EmbracePlus-style wristband

Advantages

• EDA
• PPG / heart rate
• HRV-related analysis
• skin temperature
• accelerometry
• raw data access
• unobtrusive wrist form

Current limitation

• potentially more visible / physically noticeable than a ring

So I do not think the question is “which one beats the other?”

The more useful question is:

Do I actually need EDA strongly enough to justify a wrist wearable?

If the answer becomes yes, EmbracePlus currently looks more compatible with my sensing architecture than the Oura Ring.

If later research shows that PPG / pulse + movement + context is sufficient, then the ring becomes much more competitive because of its smaller and more discreet form.

One important limitation before I choose EmbracePlus

I need to be careful not to over-read the result.

The EmbracePlus comparison in this paper was exploratory, and the second study did not reproduce every movement condition from the first study.

So this paper does not prove that EmbracePlus is the most motion-robust device in an airport-like environment.

The earlier multicentre EmbracePlus paper is stronger evidence for pulse-rate accuracy during motion.

Together, the two papers make a stronger case than either paper alone:

Paper 34: EmbracePlus can estimate pulse accurately during movement using PPG + accelerometry.

This paper: EmbracePlus provides EDA, PPG, temperature and accelerometry in a comfortable wrist form, with raw data access and broadly comparable performance to E4.

That combination is why it is becoming a serious candidate for my prototype architecture.

What this means for my airport system

This paper strengthens the case for:

EDA + PPG / pulse + accelerometry

as the most coherent current sensing combination.

But the system should still use these signals to notice change, not diagnose a state.

A plausible architecture is becoming:

personal baseline + physiological change + movement + airport context → cautious interpretation → transparent check-in → traveller decides

The paper also reinforces something equally important:

wearability is part of sensing quality.

A technically stronger sensor can still be the wrong design if it interferes with the person's hands, movement or comfort.

What changed in my thinking

This paper is the first one that makes me feel I can begin narrowing the physical sensing hardware, rather than only narrowing the abstract sensor types.

My current position is:

EmbracePlus is now the strongest wrist-based research candidate for my prototype, but I am not ready to rule out a ring until I decide whether EDA is genuinely necessary.

That gives me a much clearer next comparison:

EmbracePlus-style wrist sensing vs Oura-style ring sensing

not as brands, but as two different sensing/form architectures.

Source

Brandebusemeyer, C., Georgi, F. and Arnrich, B. (2026). Reliability Assessment of Wearable Technologies for Physiological Measurements: An Evaluation of Shimmer3 GSR+, Empatica E4, EmbracePlus, and Pixel Watch 2 Across Cognitive, Affective and Physical Activity Tasks. Sensors, 26(14), 4376.

Key points from my reading: wrist versus finger sensing; EDA robustness during movement; weaker sensitivity to subtle cognitive-load differences; Pixel Watch 2 strong for heart-rate / temperature trends but limited raw-data usefulness in this study; Shimmer technically strong for EDA but intrusive; E4 discontinued; EmbracePlus current research-grade successor with EDA, PPG, temperature, accelerometry and raw data access; wrist-worn devices generally more comfortable than finger-based sensors.