fNIRS Datablitz 2020 Q&A: Neurodevelopment and Aging

Tuesday October 13 2020, 10:00-11:00 AM EDT

The video recording of the session is below.

Use the commenting tool below 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.

Tu8 Clare Elwell
University College London, UK
Brain imaging for global health
Q&A
6
Question to Dr Elwellx
Tu9 John Spencer
University of East Anglia, UK
Examining the early development of working memory in India using fNIRS and structural MRI
Q&A
4
Question to Dr Spencerx
Tu10 Lauren Emberson
Princeton University, USA
Infant prediction is supported by large-scale functional neural networks
Q&A
8
Question to Dr Embersonx
Tu11 César Caballero-Gaudes
Basque Center on Cognition, Spain
Mapping cortical functional activity and connectivity in the infant brain using fNIRS: New approaches, methodological considerations and influence of bilingualism
Q&A
2
Question to Dr Caballero-Gaudesx
Tu12 Heather Bortfeld
University of California, Merced, USA
Differential activation of primary and supplementary motor cortex across timing tasks
Q&A
2
Question to Dr Bortfeldx
Tu13 Franck Amyot
Walter Reed National Military Medical Center, USA
Assessment of cerebrovascular dysfunction after traumatic brain injury with fMRI and fNIRS
Q&A
1
Question to Dr Amyotx
Tu14 Martina Mancini
Oregon Health & Science University, USA
Prefrontal cortex activity and gait in Parkinson’s disease with cholinergic and dopaminergic therapy
Q&A
2
Question to Dr Mancinix
TuP2 Panel Discussion Neurodevelopment and Aging    Moderators Felipe Orihuela-Espina & Ippeita Dan
Q&A
0
Question to the Neurodevelopment and Aging Panelx

Return to main chat page https://fnirs.org/conferences/fnirs-datablitz-2020-chat/

Subscribe
Notify of
61 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments

Dan
Dan (@dan)
13 days ago
Q&A Clare Elwell" Read more »

Interesting finding. What is the cause of developmental difference of the brain responses between British and Gambians? Is it biological or social?

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  Dan
13 days ago

we are looking at which risk factors have the most “toxic” effect on infant brain development particularly malnutrition.
Our data is showing that Gambian infants in the upper growth quartile have the similar responses to UK infants. This isn’t about “biology” but about the impact of environment – probably including maternal nutrition

inês almeida
inês almeida
13 days ago
Q&A Clare Elwell" Read more »

thanks for the interesting talk, and projects! :)
do you thank that cultural differences / social (non-adverse) practices may explain some of the results, in particular the different habituation rates? or do you atribute these to adversity alone?
what controls could be used?

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  inês almeida
13 days ago

we’re interested in whether the differences are due to delayed responses which may be recovered at older ages, .ie. a maturational lag – the longitudinal aspect is proving useful for this
It’s also possible that the environment of the Gambian infants leads to them developing different adaptive responses to novelty

Lloyd-FoxSarah
Lloyd-FoxSarah (@lloyd-foxsarah)
Reply to  inês almeida
13 days ago

In addition to Clare’s reply I would add that while we look to understand longitudinal responses in the UK (as this type of paradigm hasn’t been used before in NIRS), primarily we are exploring responses within the cohort in The Gambia rather than comparing across sites. So within the sample of N = 200+ we want to explore whether poverty associated adversity factors may explain these findings, but also, as you suggest, whether social/family caregiving practices may explain accelerated specialisation of habituation responses in some infants compared with others.

Xin Zhou
Xin Zhou
13 days ago
Q&A Clare Elwell" Read more »

Very interesting projects! Dr Elwell, do you apply short channels for children at 3 years of age when they are actively doing tasks?

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  Xin Zhou
13 days ago

good question – the current data is based on 30mm spacing with the Brite24 Artinis system. No short channels, as yet

EmbersonLauren
EmbersonLauren (@embersonlauren)
13 days ago
Q&A Clare Elwell" Read more »

thanks for this fascinating data Clare– I’m wondering about your FC developmental pattern towards more inter than intra-hemispheric connectivity. My understanding from the resting state fMRI literature is that the pattern is the opposite, at least for older infants. Can you comment? Is this because your infants are so young? Is this the pattern you see in the UK as well as Gambia?

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  EmbersonLauren
13 days ago

Hi Lauren, it could be the age of the infants. I know Chiara (who is leading this work) is looking at the UK infants…let me remind myself of that data and get back to you!

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  EmbersonLauren
13 days ago

looking at Chiara’s UK data there seems to be more intrahemispheric connectivity in the Gambian than UK sample at 1 month but it is worth noting that the sample sizes are different (n=125 for Gambia vs n=42 for UK). Less UK infants had good data at 1 and 5 months
Chiara knows this data well – she’s on maternity leave at the moment!

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  Clare Elwell
13 days ago

so fascinating. thanks for following up on this. I look forward to seeing where this direction of the project goes. of course, it is important to consider that infants in the Gambia could have a very different pattern of neural development with regards to FC or anything else and we shouldn’t expect it to map neatly onto what we’ve seen in Western samples. we have a story about why inter hemispheric is better to see first and than intra later but it’s more a description of what has been seen than anything else. thanks for all of your inspiring work in this population and in general.

Anne Gallagher
Anne Gallagher
13 days ago
Q&A Clare Elwell" Read more »

Beautiful work Clare! I have 4 basic questions regarding the portable recordings in toddlers: how long does the installation take with this portable device? How many channels do you have? How long are you able to record on average? Do you have to replace the caps during the recordings? Thanks

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  Anne Gallagher
13 days ago

Hi Anne, I’m going to ask Paola Pinti to comment on this as she performed the studies!

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  Clare Elwell
13 days ago

you can contact Paola directly at p.pinti@ucl.ac.uk

PintiPaola
PintiPaola (@pintipaola)
Reply to  Anne Gallagher
13 days ago

Hi Anne,

sorry for the delay, I was having issues with this Q&A page. Seems to be working now!

To answer your questions:

  • With our probe configuration, we had 11 channels per hemisphere, so 22 in total, with one Brite24
  • Toddlers were quite compliant. We used the social/non-social task developed by Sarah Lloyd fox and used in many studies, including the Bright Project. That task was developed for infants and lasts ~7mins. Some older kids complained that it was boring, but all of them completed the whole 7mins. Some of them didn’t want to take the cap off at the end, so that’s promising. Some of others of course just wanted to finish the experiment asap. I expect we could record more than 6-7mins. We are about to open the ToddlerLab here in London, where we will have an immersive VR CAVE. Hopefully such tasks will be more engaging and we can record for more time.
  • The task we used was a computer based task, so kids were just sitting in front of the monitor. Almost none of them touched the cap and the cap didn’t move, so no, we didn’t have to re-place the cap. I think that the fact of having a wearable system helped in this regard. The system was lightweight and well tolerated, we didn’t have the usual long and heavy fiber bundles that pull the head back. So the cap was stable + it was more comfortable for kids.

If you have any further question, feel free to email me p.pinti@bbk.ac.uk

Xin Zhou
Xin Zhou
13 days ago
Q&A John Spencer" Read more »

Nice response pattern in the r-IFG for the increased working memory demands! How about fNIRS responses in the left IFG?

SpencerJohn
SpencerJohn (@spencerjohn)
Reply to  Xin Zhou
13 days ago

So far, we’re not seeing much in l-IFG, however, we are currently re-running the GLM to tweak the HRF shape…perhaps l-IFG will show up in this new analysis.

Dan
Dan (@dan)
13 days ago
Q&A John Spencer" Read more »

Interesting findings! Do you thins hemodynamic patterns differences are bound to brain regions or cognitive tasks?

SpencerJohn
SpencerJohn (@spencerjohn)
Reply to  Dan
13 days ago

We’re definitely seeing differences in the HRF over regions. I’m seeing this in other studies as well. Certainly, the HRF is modulated by task as well. Nice and complicated :)

Beth Smith
Beth Smith
13 days ago
Q&A Clare Elwell" Read more »

Hi Clare, such interesting work! Re: Gambian infants, I assume they are generally at risk for stunting of growth/poor neurodevelopmental outcomes, is that correct? If so, in regard to differential head growth, are you also measuring nutrition and sleep/wake cycles? Thank you, Beth

Clare Elwell
Clare Elwell (@clare)
Editor
Reply to  Beth Smith
13 days ago

Hi Beth, yes they are at risk of wasting. We are measuring a range of markers others including feeding and sleep diaries, and have biological samples also to look at nutritional markers, e.g. iron status. All of those data are going into the model we are creating.

Dan
Dan (@dan)
13 days ago
Q&A Lauren Emberson" Read more »

Impressive approach! My wild idea is to extend the length of the task and then you see the network better. What do you think?

BlancoBorja
BlancoBorja (@blancoborja)
13 days ago
Q&A Lauren Emberson" Read more »

Thank you Lauren, nice talk. What do HbR responses look like in the results you presented?

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  BlancoBorja
13 days ago

Great question. As you often see with infants, we generally find much noisier responses in deoxy and don’t have significant responses in any of our deoxy results unfortunately.

Abigail Fiske
Abigail Fiske (@abigailfiske)
Reply to  EmbersonLauren
13 days ago

If possible, are you able to point me in the direction of some literature on the noisier deoxy response in infants? I think this might be useful in understanding some interesting results we have recently found with 10mo.

Edouard Delaire
Edouard Delaire (@edelaire)
13 days ago
Q&A Lauren Emberson" Read more »

Thanks for the talk. How can we be certain that we are only removing task signals and that we are able to recover good resting state data when looking at the residual of the GLM ?

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  Edouard Delaire
13 days ago

Great question. First, the GLM by definition picks up all of the significant task-evoked responses. We also look at the residuals from the GLM and see that there is no task related response (i.e., no significant responses in any channel). Ai-Aldroos et al 2012 PNAS does a very nice job of showing how this approach takes care of the task evoked responses. But we’re not interested in looking at the resting state connectivity but the task-evoked connectivity and these can be very different and shouldn’t change with task conditions like we find.

Heather Bortfeld
Heather Bortfeld
13 days ago
Q&A Lauren Emberson" Read more »

Lovely, Lauren. I’d love to hear more about the photgrametry approach you mentioned. Will check out the paper you mentioned in any event. Thanks!

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  Heather Bortfeld
13 days ago

Thanks. Let’s talk more offline. It’s a great method for infants and only required about 2-3 minutes and you can even use an iPhone to do it (we use a GoPro). So it’s cheap, easy, fast.. babies can move and play with a toy while the video is taken so we have a higher degree of success with this for all infants.

Lloyd-FoxSarah
Lloyd-FoxSarah (@lloyd-foxsarah)
Reply to  EmbersonLauren
13 days ago

Lauren, I would love to know more about the technique you use too. Please do follow up offline with me too. Thanks!

Sarah Lloyd-Fox
Sarah Lloyd-Fox
13 days ago
Q&A Lauren Emberson" Read more »

A great talk, really interesting study! I love the connectivity findings in the frontal-parietal area during predictive events. Is this work that you plan on doing in the first month of life as well, and would you hypothesise finding different connectivity at this age? Finally, quick question, which system are you using with your 70 channels?

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  Sarah Lloyd-Fox
13 days ago

We’re using a Shimadzu LABNIRS. We haven’t planned on doing this that young but I think could you find this kind of effect very early in infancy but probably with a slightly easier sequence. There is good work showing this kind of effect in EEG at 3-4 month olds but I think they probably can’t learn the 4 element sequence that young but we have data that even neonates can modulate occipital responses based on predictability so I suspect this higher-level network would be involved too. I’d love to talk more!

LIN CAI
LIN CAI
13 days ago
Q&A Lauren Emberson" Read more »

Hi Lauren, thank you for your great talk. I am just wondering whether you used the mean FC value averaged across all channels in frontal or parietal regions when you calculated the frontoparietal FC ?

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  LIN CAI
13 days ago

We first averaged all the responses across the channels and then did the connectivity calculation (correlation). Doing it the other way doesn’t yield good connectivity results and we’ve found that across a number of datasets now.

Helga Miguel
Helga Miguel
13 days ago
Q&A John Spencer" Read more »

Very interesting work! How young were you able to digitize toddlers? How many points did you use?

SpencerJohn
SpencerJohn (@spencerjohn)
Reply to  Helga Miguel
13 days ago

6mo were the youngest in the group. We had 32 channels in India and 48 in the UK.

Ellie Smith
Ellie Smith
13 days ago
Q&A Lauren Emberson" Read more »

Really great talk – thank you so much.

Looking at your new paradigm the time periods you are displaying your stimuli for are much longer than in your previous fNIRS paradigms. Do you think the shorter presentation times (~1-1.5s followed by 4-9s baseline) in your previous paradigms provide enough time for the haemodynamic response to develop and return to baseline before the following trial? I’m exploring using a similar paradigm and am unsure about paradigm timings.

Thanks in advance.

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  Ellie Smith
13 days ago

Great question. Yes this is a much longer time of both presentation and baseline. When you have longer stimulation times you need longer periods for rest because you get a stronger response (it’s block vs. trial). So I think the previous timing is good too for baseline as long as you don’t have a long period of stimulation before it. But longer baselines are better if you can do them. The problem that we have is that babies stop complying when the baselines are too long.

Anne Gallagher
Anne Gallagher
13 days ago
Q&A César Caballero-Gaudes" Read more »

Thanks for the great talk! Why do you think you get null results between awake vs. asleep data?

BlancoBorja
BlancoBorja (@blancoborja)
Reply to  Anne Gallagher
13 days ago

Thank you Anne, all the participants in the studies Cesar presented were tested during natural sleep, we observed null results in our resting-state study in the comparison between monolingual and bilingual infants.

LIN CAI
LIN CAI
13 days ago
Q&A César Caballero-Gaudes" Read more »

Thanks for you great talk. How did you determine the coordinates on the scalp for each channel in your infant study? Maybe by 3D digitizer?

César Caballero-Gaudes
César Caballero-Gaudes (@ccaballero)
Reply to  LIN CAI
13 days ago

We did not use 3d digitizer, but we localized the position of nasion, inion and preauricular points were used as external head landmarks, and we employed caps of two different sizes (i.e., 40 and 42) to adapt to individual head perimeter. For the sensitivity profiles, we then use Toast++ with an average 6-month-old infant template (Richards et al., 2016).

BlancoBorja
BlancoBorja (@blancoborja)
Reply to  LIN CAI
13 days ago

Hello Lin, a 3D digitizer was not available for these studies. Channel coordinates were estimated based on the sensitivity profile of our setup based on a 6-month-old infant head model.

LIN CAI
LIN CAI
13 days ago
Q&A John Spencer" Read more »

Thanks for your wonderful talk. I am curious about whether you used the same probe geometry to examine the developmental changes of Visual WM during the first 2 years?

SpencerJohn
SpencerJohn (@spencerjohn)
Reply to  LIN CAI
13 days ago

Same geometry, but we scale the source-detector distance by head circumference (detailed in a 2014 paper we published in NeuroImage).

MesquitaRickson
MesquitaRickson (@mesquitarickson)
13 days ago
Q&A Lauren Emberson" Read more »

Hi Lauren,
Very interesting talk (and congrats on the new position at UBC!).
Concerning the background connectivity, have you ever looked at the correlation between the background connectivity strength and the functional task response?

Last edited 13 days ago by MesquitaRickson
EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  MesquitaRickson
13 days ago

Great question. We haven’t but note that the condition difference is opposite. We get greater task responses in the unpredictable sequences and greater connectivity in the predictable sequences.

EmbersonLauren
EmbersonLauren (@embersonlauren)
Reply to  EmbersonLauren
13 days ago

But it could be really interesting to see whether the relative task-based responses across conditions related to the relative strength of the connectivity in individual infants. I would hypothesize that they would be related: strong connectivity modulation, larger differences in task responses.

MesquitaRickson
MesquitaRickson (@mesquitarickson)
Reply to  EmbersonLauren
13 days ago

I would suspect that the more “energy” the brain spends to perform a task the less is seen in background activity, as you saw for the predictable/unpredictable sequences. Did you use the fNIRS response in the FIR model? This is very interesting!

Felix Scholkmann
Felix Scholkmann (@felix)
Admin
13 days ago
Q&A Heather Bortfeld" Read more »

Interesting results. Did you also have a look at the HHb data? Did you see the same spatial pattern with HHb compared to O2Hb?

Heather Bortfeld
Heather Bortfeld (@heater)
Editor
Reply to  Felix Scholkmann
13 days ago

Yes, apologies for not including those data. But indeed, the HHb follows the HbO as it should, which was good to see. Thanks!

Felix Scholkmann
Felix Scholkmann (@felix)
Admin
Reply to  Heather Bortfeld
13 days ago

Great! ;)

César Caballero-Gaudes
César Caballero-Gaudes (@ccaballero)
Reply to  Heather Bortfeld
13 days ago

Indeed, it is great!!

Dan
Dan (@dan)
13 days ago
Q&A Heather Bortfeld" Read more »

Very interesting paradigm! Is syncopation task simply more difficult or is there any qualitative difference? How is it affected on cortical activation?

Heather Bortfeld
Heather Bortfeld (@heater)
Editor
Reply to  Dan
13 days ago

That’s a great question, and is one we’re interesting in pursuing. Syncopation certainly involves far more “cognitive” resources, such as working memory and executive function. The question is whether that’s simply quantitative in nature (harder, but the same), or rather represents qualitative differences. Much more work to be done to tease that out.

Andrew Lapointe
Andrew Lapointe
13 days ago
Q&A Franck Amyot" Read more »

Very interesting study.

When you looked at your data, were there any interactions (between MRI & NIRS) that allowed you to identify group differences that could not be identified using a single modality?

AmyotFranck
AmyotFranck (@amyotfranck)
Reply to  Andrew Lapointe
13 days ago

I didn’t find any interaction. Both modalities give CVR measure on the cortex. We wanted to validate CVR measure with fNIRS for our clinical trial on TBI drug therapy and monitoring vascular injury during the acute phase.

Dan
Dan (@dan)
13 days ago
Q&A Martina Mancini" Read more »

Interesting talk! Is the difference related to task switching?

Martina Mancini
Martina Mancini (@mancinim)
Reply to  Dan
13 days ago

thank you! Not sure if it is related to task switching, it seems PFC activity is decreasing with levodopa compared to no-medication (increased automaticity) while this effects is annulled with the presence of both levodopa and donepezil. Interestingly walking performance is better with the combination of the two drugs. Happy to answer to more questions!

Rodrigo
Rodrigo
13 days ago
Q&A Martina Mancini" Read more »

Can you please comment on what you would expect to find in the primary motor cortex with cholinergic medication?

Martina Mancini
Martina Mancini (@mancinim)
Reply to  Rodrigo
13 days ago

Great question. Hopefully we’ll find that out soon!