You can see also our work on the E4 validation:

Article Stressing the accuracy: Wrist-worn wearable sensor validatio...

Our results are slightly different from those of Matthijs Noordzij as we found no visual ressemblance between wrist and finger SC, and we found accurate HRV metrics compared to ECG only in motion- and stress-free conditions. Having an adaptation period (e.g., 1 min) before the recording intervals of interest seems to increase the measurement accuracy.

Hi Mateusz,

Modern monitors for a task like this are all good enough - miniaturisation and modern hardware engineers have progressed very well in the last few years. The Empatica monitor is OK, but obviously does only SCR. The Zephyr has reasonable accuracy and takes a variety of measurements, but is a lot more expensive and is reasonably obtrusive for long term measurement. I never used it for more than a few hours.

To find more devices (and there are many), Profs. Wilhelm and Grossman reported a long list of portable monitor manufacturers in the attached paper:

https://www.unispital-basel.ch/fileadmin/unispitalbaselch/Bereiche/Medizin/Psychosomatik/Lehre_Forschung/Publikationen/wilhelm_grossman_emotions_biol_psychol_2010.pdf 

There are many more - Mindware, for instance, make a full-suite portable physiological monitor:

https://www.mindwaretech.com/product_detail.asp?ItemID=2054

You should be able to find instrumentation easily enough.

Finally:Your main problem isn't measuring EDA or HR over time, your main problem is successfully defining what measurements from these represent when discussing 'stress' - which isn't a unitary concept. 

I am using an Empatica E4 at the moment. It is a wristband device that records BVP, GSR and temperature. Data can be recorded on the device itself (I have recorded up to 8 hours of data, but maybe you can get more). There is also an Android app which is quite useful.

Most useful aspect of this device is that it can be used in conjunction with a BlueTooth dongle and TCP/IP streaming whereby you can sample data in your software. I have done this on Windows OS. You can also write your own Android code to do this on a mobile.

What I haven't done so far is test the veracity of the data from E4 against data from some other device, such as Biopac.

We are currently testing the Empatica E4 wristband against the VU-AMS (www.vu-ams.nl).

First results show that the Empatica measures are quite different from the VU-AMS results. The VU-AMS measures seem to be more precise, i.e., showing more variation and variability.

I would really be interested in experiences of other Empatica/VU-AMS users!

We are also running various validation experiments and feasability/ usability studies on the E4. I cannot provide you with a conclusive answer yet, but i think it is also very important to consider your own use case: What is good enough? The HR measruments are very sensitive to movement so if you are planning to do heartrate variability in the field with th E4 you might have serious problems, whereas if you are only interested in gross estimates of heart rate over large time frames it might be good enough.

It depends on your research question and what information you want to collect. 

Do you want to include physiological measures or not? Do you want to have high information density? Do you want to include a large sample? 

The most precise is the VU-AMS when it comes to physiological measures (ECG, HRV, etc) but you can only wear it several days (the electrodes irritate the skin). The Zephyr seems to be accurate, but is not easy to wear for women, and you loose data when the sensor doesn't make contact with the skin. There is also the Actical, which is used a lot in our department. Is placed around the wrist or ankle (and waist, but I don't know how), waterproof, validated and the battery works a lot longer than 24 hours. Nowadays you also have the fitbit, garmin and smartwatches. If you want to include a really large sample, this might be interesting as well.   

For me it is very important to have an instrument that is sensitive to (small) changes over time, as we want to do analyses on the link between physiology and behavior (both coded continuously, i.e., every second)... I am not sure whether the Empatica, or other devices, are sensitive enough for this purpose...

Hi. Have you considered the Shimmer GSR unit? It's a very decent research tool, costs half as much as empatica E4 and seems to have good quality signals.

Also try:

Microsoft band2 https://www.microsoft.com/en-us/band/techspecs Much cheaper and it has an impressive list of sensors including: Optical heart rate sensor 3-axis accelerometer/gyro Gyrometer GPS Ambient light sensor Skin temperature sensor UV sensor Capactive sensor Galvanic skin response Microphone Barometer I have built my own app to collect the data and link it to other sensors such as Noise monitor on the phone, self report labels, and the Phone's GPS (the band we have doesnt has GPS)..etc.

Eiman Kanjo this band looks nice, such a lot of sensor (even EDA).

If I good understand, there is a possibility to access to the raw data and download it or collect with other apps?

@Mateusz I found unofficial SDK forms band.https://github.com/JHamberg/unisensors-band2.

Did you find other similar devices? I am interested in physics body measuring and correlated data witch psychological signals.

Have you read about Hesvitband S3?

Empatica is excellent, but it cost a lot.

We live in the same city. We can do the research and development together.

We are testing Everion at this moment, it is an interesting device :)

https://www.biovotion.com/everion/

We are still processing and analyzing the data so, we don't have an answer on the quality yet. But you can get raw and processed measures for PPG and EDA.

This device looks great. Do you have information about pricing and availability?

Mateusz Soliński I built my own app to log the data and timestamp it along with Location and envromentl noise. Happy to share based on collaborative projects. Also, we have collected a valuable dataset (see my publications).

Hey Kamil, sorry for the delay!

I am sending you the information about pricing in a private message.

Cheers

Ariane

Ariane Morassi Sasso can you send me also information about price and availability. We can not find anything online.

Thanks

Alex

See also:

Article A standardized validity assessment protocol for physiologica...

You can see also our work on the E4 validation:

Article Stressing the accuracy: Wrist-worn wearable sensor validatio...

Our results are slightly different from those of Matthijs Noordzij as we found no visual ressemblance between wrist and finger SC, and we found accurate HRV metrics compared to ECG only in motion- and stress-free conditions. Having an adaptation period (e.g., 1 min) before the recording intervals of interest seems to increase the measurement accuracy.

For most of our SC signals visual ressemblance was also often weak, but perhaps slightly better than in your study Luca Menghini? I need to check. I would say both our studies (and earlier studies) would warn against assuming SC ambulatory wrist measurements will show high cross-correlations to finger measurements. Importantly, I also think you do need to always consider your particular use case and what construct you want to operationalize by using the physiological measurements. Then you need to make an informed decision on whether the measurement tool is appropriate for a given level. So try to assess the signal, parameter and / or effect level you are interested in, and then make an argumented decision.

I agree with you Matthijs Noordzij and it is interesting that both of our studies suggested that wrist SC is more sensitive to high-intensity social stressors like public speaking or singing a song. However, wrist SC seems to be more informative of the thermoregulatory origins of EDA, which can be very important in certain applications (e.g. discriminating REM and non-REM sleep, see article below). In any case, I found your analytical approach very interesting and exhaustive, and I will surely take it into account for new validation studies.

Article Quantitative analysis of wrist electrodermal activity during sleep

Hi Mateusz,

we are currently starting a large scale project which is aimed to link behavioral, neuroimaging and real-life data to examine social information processing in lonely individuals, so we will be looking for a wearable device for HRV measurement. Any recommendations?

If you are looking for HRV only then try Oura ring https://ouraring.com/?utm_source=google&utm_medium=cpc&gclid=EAIaIQobChMIt6OGjcKD5QIVCLLtCh3VKARAEAAYASAAEgJAI_D_BwE

Otherwise you you could try bioharness zephyr belt which is more expensive but it comes with many sensors including HR, breathing rate motion...etc.