With the recent development of new portable and lightweight EEG systems come numerous reports in popular science outlets but also more and more publications in peer reviewed journals reporting on experiments using such systems in stationary or ambulating participants.
Because proprietary systems do not provide access to the algorithms underlying the relevant output (e.g., detection of emotional state), the results of such studies are highly questionable. This is especially the case for approaches using ambulating participants. Here, gait and other movement-related artifacts distort the signal of interest and it is not clear how these non-brain activities impact the output.
While reports based on easy to use EEG systems without doubt further the increasing popularity of neuroscience in the general public, they might produce expectations that cannot be fulfilled (similar to the failed expectations regarding BCI research).
There has to be an open discussion with the aim to defining rules how to deal with research based on proprietary systems regarding publications.
Would you want EEG results to be published in peer-reviewed journals when there is no access to the algorithms producing the output?
The main scientific article is a new hypothesis that can be properly proved relatively simple methods. There are scientific peer-reviewed journals that are willing to publish such articles.
Given that there are many ambulatory options for EEG recording that are intended to be research-grade, then researchers should favour those options, even if the cost is higher. So the simple answer is no. You've already spelled out the motivation for this answer.
On a more practical level, cost itself can be a relevant factor, e.g. if a wireless ambulatory recording unit will be in danger of being damaged in the experiment, or if some unique form factor is required. It might be feasible to use proprietary gear if the signal outputs have been verified in comparison to a research grade device. We attempted this with EDA recording hardware, see https://www.researchgate.net/publication/280831275_Feasibility_of_an_electrodermal_activity_ring_prototype_as_a_research_tool
The results indicate that the signals are not identical and the proprietary device was not as reliable. But the level of error was usually within acceptable bounds.
However, the cost of a comparative study easily outweighs the cost of buying better hardware. And the study itself is not going to be so easy for EEG!
Conference Paper Feasibility of an electrodermal activity ring prototype as a...
It's a tricky issue to balance the rights of innovative entrepreneurs with scientific openness. Clearly, the algorithms have to become transparent in order to evaluate the science. Perhaps some legal arrangement as in the pharm industry in the US, whereby new formulations become public after a certain number of years? However, that's not a perfect solution because it requires a lot of work by other researchers to go back and re-evaluate anything published in the interim. However, I can't think of anything obviously better. (Incidentally, another example of this might be Porges' algorithm for using HRV to estimate "vagal tone.") Tough issue Klaus! Curious what ideas others may come up with. All the best - G.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Human-Computer Interaction