Monday October 12 2020, 9:00-10:00 AM EDT
The video recording of the session is below.
Use the commenting tool to discuss the presentations given in this session. Please register as a member to be able to ask and answer questions https://fnirs.org/account/. You can register for free for this web content.
Boston University, USA
|Neuroimaging in the everyday world||
University College London, UK
|The promise and practicalities of wearable, high-density DOT in newborns, infants and adults||
Polish Academy of Science, Poland
|Multi-channel time-resolved spectroscopy system||
Politecnico di Milano, Italy
|Recent advances in time-domain fNIRS||
Massachusetts General Hospital, USA
|Neuromonitoring with the next generation of NIRS & DCS devices||
University of Texas Southwestern, USA
|Diffuse optical monitoring of spinal cord ischemia in large animal models||
Oslo Metropolitan University, Norway
|Ambient light reduction for fNIRS||
Hardware Development Moderators: Adam Lliebert & Rickson Mesquita
Return to main chat page https://fnirs.org/conferences/fnirs-datablitz-2020-chat/
Rob, great progress with your wearable system! Loved seeing the HRFs at different distances. What challenges do you have getting through the hair?
Thanks David! In the infants it isn’t much of a problem, to be honest – most existing devices use flat probes in infants and barely have time to worry about hair. In the adults it is harder, of course – more optodes = more difficult, but the sensitivity of the devices helps us get past that, to some extent at least.
super cool system. nice to see it at work. How much does the 24 tiles cap weights? How long the battery last?
Hi Mari – I’m not actually sure! Each module is about 7g, so including plastics probably around 300-400g. I have worn 24 tiles myself for hours on end, and it does not feel particularly heavy. I have been running these studies in cabled mode so far.
Rob, how to the “tiles” cover the “inter-tile” space?
I didn’t get a chance to explain this explicitly, but the tiles all talk to each other – so you get intra and inter-tile channels.
I understand that detectors can “see” sources from neighboring tiles. Just a question of managing the multiplexing of sources.
Hi David, can you tell us more about your data sharing vision?
Not really a vision. Many of us in the field have been talking about data sharing for many years. We wrote into our human subject research protocols several years ago that we were going to share anonymized fNIRS data and now that SNIRF is available for standardizing the fNIRS data format, we finally started posting our data on line. There are many repositories out there that you can use. We choose to use nitrc.org because we’ve used it for other things in the past. We want to see others posting their data on whatever repository is best for them. But then they can also put information about their dataset in the meta-data base we created at https://openfnirs.org/data . Given that data can be shared from any of more than a dozen repositories, our thought was that the meta-data base could be a repository to help users find data sets that best fit their needs.
Wonderful work, thank you for your talk. I was wondering how easy hair manipulation is with the tiles? ie how well can you access the hair to move it out of the way if there is a bad optode-scalp connection?
In infants we didn’t have to do anything – the system seems sensitive enough that it is just not an issue (at least at 6 months). In adults it is more tricky – but there are light guide parts that comb through the hair, and by toggling the module back and forth you can usually get good performance.
Hi Rob, nice to see this technology coming along. Do you have any problems with communication between modules? How 24+ tiles would impact temporal resolution?
At the moment we can do 24 tiles at 5 or 10 Hz in the lab. Gowerlabs are working on expanding the commercial device up to as many as ~56 tiles- that’s probably more than we actually need for the whole adult head. The challenge is time, as you suggest – we have to figure out an adaptive multiplexing scheme.
Hi Rob! This is a great system with beautiful signal separation. How did you solve hair-interference problem? Most channels are in hairy regions.
In the infants hair isn’t much of a problem, to be honest – most existing devices use flat probes in infants and barely have time to worry about hair. In the adults it is harder, of course: more optodes = more difficult, but the sensitivity of the devices helps us get past that, as do the light-guides that are designed to comb through the hair.
HI Rob— These infant recordings look incredible. So exciting. How would anatomical co-registration work with a system like this when all of the fibers are enclosed.
Thanks Lauren – the infant data is really nice- even better than we hoped. Luckily we thought of the registration issue: Each module has three little indentations on the back side, which sit a fixed distance exactly above the source optodes. So if you can measure the position of those points (e.g. with polhemus), you can work out your optode locations without approximation (because everything is of known dimensions). You’ll have seen the little fluoro triangular stickers (bright green/pink) – well, those are aligned to those same indentations, so with photogrammetry we can measure their positions and register the probe that way. This is what we did in the infant.
How to be a ninjaNIRS tester? I am new in this field but would like to investigate the feasibility to use the technology to use in my research with people with communication disorders.
Best to email me at Boston University. Easy to find my email address.
What are the two top challenges for developing these systems, and what are a couple that would have been on the list but the field has addressed in the last few years?
As far as I can see from the commercial companies and other academic researchers, it seems that there aren’t really any challenges to developing these systems. Highly sensitive and highly multiplexed systems have already been demonstrated. I guess that is a challenge that has now been addressed for wearable systems. Battery power and long duration of untethered measurements probably still needs to be demonstrated. Ease of application of the probe and resistance to motion are likely still challenges, although again it seems several are saying this is solved. For me, these are still challenges. But coming up next, the real challenges will be related to showing impactful scientific results in the real world. Figuring out the experimental approach and the analysis methods will be challenging. I am inspired by your fNIRS studies during movie watching… seems to me that moving to analyzing data in the everyday world is like analyzing fNIRS data of someone watching a movie.
CulverJoseph I’d love to read your response to this question.
Hi Alessandro! This is a great system. It’s great to have wave forms in TRS. How is the temporal resolution? It seems very rapid and how did you realize it?
Normally 1Hz but we can increase to 10Hz (if SNR allows it)
System looks great, and particularly the retinotopy maps.
What are the new challenges you are working on with the grower lab systems.
Hi Joe – yes, the Culver-lab fanboy in my coming out again.
The next big challenge is scale up. The commercial system is currently at 12 modules – in the lab we have achieved 24. Gowerlabs is now working very hard now on scaling up. The LUMO platform was designed entirely with this in mind, though, so it is basically a firmware challenge at this point. Hoping to have more to show on this in the next few months.
in your opinion, what does the increase of signal magnitude proportional to the source-detector length indicates?
It shows that the HRF is coming from the brain – not the scalp. The relative sensitivity of a channel to the brain increases with distance, so longer channel = greater proportion of brain signal. It’s a great sign that your response really is from the brain.
what are the obstacles to move to humans with this fiber based system?
The introduction of the spinal cord probe into the epidural space has good predicates with other devices (e.g., electrical spinal cord simulators). The FDA has already done an initial review and designated this system as a ‘breakthrough device’ to expedite human translation. We’re confident in the design, but of course there are many obstacles for first in humans use for any invasive device (biocompatibility, sterility, etc).
Hi Rob. I noticed on one of your slides that the HRF showed an initial dip followed by a peak after about 10s. What do you make of the initial dip along with this timing? Seems like a different shape relative to previous studies…
Hi John – yes – great point. In the first infant study with this technology, we wanted to stick as closely as possible to existing data processing approaches – which meant following the Di Lorenzo paper and block-averaging. It turns out the response is nowhere near finished with 10s block-to-block. A good argument to moving to a GLM approach – which I know you are working with I think.
Rob, really nice work. Do your channels have fixed gains, or is there a way to change them on the fly?
Thank you Peyman for the interesting talk. I was curious if your technique can be applied to ambient light artifacts in the near-infrared spectra. Specifically, if you had any suggestions for preventing and correcting during post-processing eye tracker (near-infrared) correlated noise.
Question for Torricelli.
Congratulations. In your TRS measures in walking subjects, cortical HHb changes are quite small respect O2Hb changes. Any explanation?
Thanks. Indeed the actual estimated change in HHb and O2Hb depends on many parameters in the physical models (e.g. backgrund optical properties, depth and extension of absorption change, selection of time windows, …).
Nice study. The results you showed is an average of 3 subjects? Thank you.
Yes, just a pilot test.
Hi Caroline, Sorry for the very late response:-) I think the same suggested strategy of the SHADE method can be applied for the elimination of any other light sources as eye tracker etc. my best regards, P