Tim Laumann “Brain activity is not just for thinking”

Transcript

• 00:05Already.
• 00:08Well, I guess they don't
• 00:09get that part, alright.
• 00:10Also feel a little warm.
• 00:12This thing was sitting on
• 00:14my front and showing up.
• 00:18Well, so I I wanted to talk to the
• 00:23about some just some mostly just
• 00:26concepts and ideas more so than
• 00:29a lot of my own empirical work.
• 00:32In particular just as a way of I guess
• 00:36explaining or framing a little bit
• 00:38in my point of view on how I think
• 00:41about understanding resting state
• 00:43activity and this this thing that
• 00:45I've been studying for so long now.
• 00:47And I I feel a little bit like the
• 00:51brain here, but it also represents
• 00:53a little bit of the idea that I
• 00:56think I'm going to try to get at.
• 00:59So the big question.
• 01:01Is what is spontaneous activity?
• 01:04What is this thing that we're
• 01:06trying to study and that we've
• 01:08made so much in over the years?
• 01:10Can we, can we connect it to anything about
• 01:13how we understand our experiences and
• 01:16also how we understand how the brain works?
• 01:19And I'm probably out over my skis,
• 01:23so to speak,
• 01:23on some of these things given that
• 01:25it's not stuff that I, you know,
• 01:26a lot of work that I wasn't involved in,
• 01:28but I think it's helpful to think
• 01:30about some of the basic neuroscience.
• 01:32Mechanisms to try to understand
• 01:35and frame how we interpret this.
• 01:38So I'm going to go back a long way.
• 01:41This is on Burger,
• 01:43who is a psychiatrist who actually
• 01:46invented the EEG back in 1924.
• 01:49And he was the first, you know,
• 01:52always bad to the first,
• 01:54but I'll say one of the first
• 01:56in true people to really notice
• 01:59this intrinsic electrical signal,
• 02:01that there's a spontaneous activity that
• 02:03you can measure in the brain if you
• 02:05use some device to record its behavior.
• 02:08And you see,
• 02:09there's a cool trace that he
• 02:12has in this monograph.
• 02:13And what's really amazing is monograph.
• 02:16I love this quote from 1929 says is it
• 02:20possible to demonstrate the influence of
• 02:23intellectual work upon the human EEG,
• 02:26and several gram insofar
• 02:27has been awarded here.
• 02:28Of course.
• 02:29Of course,
• 02:30one should not at first entertain
• 02:32too high hopes with regard
• 02:33to this because mental work,
• 02:35as I've explained elsewhere,
• 02:37adding only a small increment.
• 02:40The cortical worker which is
• 02:42going on continuously and not only
• 02:44in the waking state.
• 02:45So I think it's very presented
• 02:47statement that he made now nine
• 02:50years ago almost and really actually
• 02:53reflects a lot of how I think about it,
• 02:56which is maybe not unrelated to the fact.
• 02:59This is quoted from actually Mark
• 03:02request paper who has said a lot
• 03:04of influence on on how we think
• 03:07about these things.
• 03:08So we can also see this spontaneous activity,
• 03:10not just in EEG or electrical recordings.
• 03:14People have done this.
• 03:17In the optical imaging,
• 03:19this is calcium imaging now in
• 03:21anesthetized cat in darkness.
• 03:24And you can see a video on the left
• 03:26hopefully of you see sort of a miss
• 03:29and then there's a kind of organized
• 03:31pattern that comes out of the mist.
• 03:34There's there's several really
• 03:35neat papers about this from
• 03:37the 90s and early 2000s.
• 03:39And what it was really demonstrating
• 03:42showing was that even in anesthetized
• 03:45animal with no stimulation.
• 03:47Coming in in darkness that these
• 03:50patterns of functionally corresponding
• 03:53signals were coming out of the
• 03:56mist and their correspondence to.
• 03:58What the evoked signals look like when
• 04:01the animal was away can be exposed
• 04:04to orientation columns in this case.
• 04:07So there's this correspondence
• 04:09between function that's a vote
• 04:11and function that is spontaneous,
• 04:14or access to connectivity
• 04:17that is spontaneous.
• 04:19As we all know, we see the same
• 04:21kind of thing in bold signals.
• 04:23Now we're talking obviously
• 04:25different spatial,
• 04:26temporal scales across these different
• 04:28modalities and there's lots of
• 04:30complexity to consider with that,
• 04:32but it's it's this very
• 04:35similar observation that.
• 04:37At rest.
• 04:39In an MRI scanner,
• 04:41we see fluctuating activity that
• 04:43we all know correspond to these
• 04:46functional systems that we've
• 04:48been well described over and
• 04:50over again in different ways,
• 04:52but convergent ways.
• 04:55So this fact that we're seeing
• 04:58this spontaneous activity that
• 05:00has some functional representation
• 05:02really suggests that it has some
• 05:06significant Physiology some and it's
• 05:09an important physiological significance.
• 05:12But one of these things about OK.
• 05:15So one view might be that these
• 05:20fluctuations actually reflect mind wandering.
• 05:23There are unconstrained cognition that
• 05:25there is a stream of consciousness that,
• 05:28you know,
• 05:29we all subjectively experience that that.
• 05:34Working very comfortable.
• 05:35We spent our whole lives enjoying
• 05:38this subjectivity that we perceive
• 05:40things we can think about,
• 05:42things we can think about the future.
• 05:44We beat ourselves up about stuff that
• 05:48we've done the that, you know, the 1st.
• 05:51And we imagine that we said,
• 05:53you know, I think about my breakfast,
• 05:55I think about what I'm going to have
• 05:56for dinner. I think about problems.
• 05:59I think you fantasize about other things.
• 06:03Umm. You.
• 06:08The color is.
• 06:12And. It's really compelling to
• 06:14think about this because this
• 06:16is really all of our experience
• 06:18doesn't how we experience our lives.
• 06:20This is the dominant way we understand
• 06:23reality is through our waking,
• 06:25conscious experience and subjective state.
• 06:29And it's sort of easy to imagine that
• 06:31that must correspond to the way what
• 06:34we're seeing in these bold activity
• 06:36patterns and you're laying in the
• 06:38scanner and they fluctuate and this
• 06:41relates in some way to that experience.
• 06:44And actually the the human
• 06:47neuroimaging literature encourages
• 06:50this idea not unreasonably.
• 06:53And that we've we have an amazing
• 06:56history of decades now of task
• 06:59devote literature that shows that
• 07:01when we impose different task states
• 07:04we see changes in both signal that
• 07:07are functionally localizable and we
• 07:09believe that that means that reflects
• 07:12cognitive operations associated with.
• 07:14This past manipulations.
• 07:16And you know it's was done really 1st
• 07:19and and had studies at a whole brain
• 07:22level and then obviously functional imaging.
• 07:28So in this context,
• 07:29and that is in the task of both contexts,
• 07:33we actually view the spontaneous
• 07:35activity as as background noise.
• 07:38I mean that's how it was always
• 07:40considered until people really
• 07:41started focusing in on on its own.
• 07:43I give you a stimulus and I see
• 07:46a response and then there are
• 07:48deviations around that response.
• 07:50So I have to keep showing you
• 07:51these stimulus over and over.
• 07:52And it's what's this irritating thing in
• 07:54the background that keeps making it hard
• 07:56for me to see that thing I care about.
• 07:58And that covers all of this.
• 08:00Spontaneous activity is actually just noise.
• 08:03It's in the way.
• 08:04And so we want to average it
• 08:06away and that's what we do.
• 08:07We still do.
• 08:10But now let's, you know,
• 08:11talk a little bit about the properties
• 08:14of that spontaneous activity
• 08:15when we look at it on its own,
• 08:17outside the context of a task.
• 08:21And so I, you know,
• 08:22I wouldn't miss not to mention these things.
• 08:25We have to talk about these things.
• 08:28As you all know,
• 08:29this is unavoidable the,
• 08:31the,
• 08:31the possible sources of variability
• 08:35that that might not be of super
• 08:38great interest from understanding
• 08:40the brain and how it operates,
• 08:44but they're actually huge in,
• 08:45in this particular measurement.
• 08:47So there are scanner artifacts that
• 08:49could be going on in that background.
• 08:51Thermal noise from the,
• 08:53the way the measurement is
• 08:54made with the instrument,
• 08:56there's head movements that that
• 08:58affect the signals that we're seeing.
• 09:00There is also an interesting
• 09:03Physiology like changes in PCO 2,
• 09:05respiration,
• 09:05other things that are actually
• 09:07quite interesting in the but
• 09:09they might not directly reflect
• 09:11the neural activity that we're
• 09:13we're kind of interested in.
• 09:17There's also this possibility
• 09:20that maybe we're going to.
• 09:25See changes in connectivity or in in
• 09:28this case. So I skipped sort of a lot
• 09:30of background because I assume a lot of
• 09:32folks are really aware of this stuff.
• 09:34But here we have correlation matrices from.
• 09:4010 minutes of data from day-to-day.
• 09:42And what you'll observe when you collect
• 09:4410 minutes of data on individual is just
• 09:47for us pulled back over to 414 months that
• 09:50there is variability in that response.
• 09:53So what's the source of that variability?
• 09:56That's something about what he's
• 09:57thinking on day one versus the other day.
• 09:59Is there something that happened
• 10:01that day that might be different?
• 10:03Well, maybe.
• 10:04And actually you can kind of put
• 10:07in this game of doing sampling at
• 10:10A at a smaller scale.
• 10:13So there's now a large literature on this
• 10:17where you might look at a within a run,
• 10:22fluctuations of this spontaneous
• 10:23activity or changes in spontaneous
• 10:26activity over even shorter timescales,
• 10:29minutes, 2 minutes.
• 10:31And what you'll observe are these huge
• 10:34fluctuations in the connectivity and the
• 10:36strength of the connectivity that we measure.
• 10:42But without going too far into it,
• 10:46I'm for the moment going into this,
• 10:48dismiss all of that as being
• 10:51really a statistical phenomenon.
• 10:53Primarily that when you're
• 10:55making your measurement,
• 10:57an estimate of something like
• 10:59functional connectivity or the
• 11:01pattern of spontaneous activity
• 11:03which is from which we're arriving,
• 11:06that it depends on how much data
• 11:08you're getting and the proper
• 11:10underlying statistical properties.
• 11:12Of the time series that we're measuring,
• 11:16what that variance looks like.
• 11:18And so it turns out that with that pull
• 11:21drag data and even at shorter time scales,
• 11:25the properties of variability,
• 11:27the fact that we see those fluctuations
• 11:31is perfectly almost perfect.
• 11:34It's very well explained by this
• 11:38important statistical principle,
• 11:40sampling variability, which is that.
• 11:42Words decreases as sample size increases,
• 11:45or other words saying variance
• 11:47increases and sample size decreases.
• 11:49So you make a smaller high measurement each.
• 11:53You're going to get a worse estimate
• 11:54of the thing that we're measuring,
• 11:56and it's going to look more and
• 11:59more uncertain, more fluctuating.
• 12:03OK. Studying aside those concerns,
• 12:08sampling variability.
• 12:09Artifacts, Physiology.
• 12:11The remaining stuff is neural activity,
• 12:15right?
• 12:15It must be about cognition then, right?
• 12:18Where we've got rid of all the stuff
• 12:20that we we worry about and we're
• 12:22still seeing this amazing pattern.
• 12:24OK, is this cutting this down issue?
• 12:26OK, so more arguments against
• 12:29this being about cognition.
• 12:32What happens if we change the level of
• 12:35consciousness or the capacity to thing?
• 12:38Well,
• 12:40resting state networks as we just
• 12:43saw actually enjoying this talk,
• 12:44they're they're not actually fundamentally
• 12:48altered by different states of consciousness.
• 12:50And I say in in terms of the
• 12:53topography in light sleep in humans,
• 12:55it's been shown you can see the same
• 12:58pattern and the attention network
• 13:00and the default network during
• 13:02sleep as you would during wake.
• 13:05And this is, you know,
• 13:06not like they're dreaming sleep.
• 13:08This is sleeping, sleeping.
• 13:10So they're not cognizing in some way.
• 13:13But we're still seeing this,
• 13:14so that's interesting.
• 13:17Similarly,
• 13:18you can actually anesthetize
• 13:22humans and actually retain as we
• 13:25saw in the rats that they have.
• 13:31Also have similar patterns
• 13:33of resting state networks.
• 13:35Now of course if you go and
• 13:37complete that and anesthesia,
• 13:39you eliminate all that,
• 13:41your renewal, managing all of
• 13:42that negativity at that point.
• 13:44But we still see these patterns and these
• 13:48these folks we think are not thinking.
• 13:51OK, what about the opposite?
• 13:52What happens if we deliberately
• 13:54change what people are thinking?
• 13:58So we can again a huge literature on this.
• 14:03We can impose task task States and then
• 14:07analyze data we get in terms of that
• 14:10background activity as opposed in terms of
• 14:13the evoked activity and see what are the
• 14:15changes we observe in the network structure,
• 14:18in the in the connectivity
• 14:20structure during tasks.
• 14:21And the reality is you do see changes
• 14:26that are measurable and significant.
• 14:29However. They're extremely similar, actually.
• 14:33What you see the difference.
• 14:36You can make a difference if you
• 14:37control a little effect,
• 14:38but the gross organization
• 14:41is highly structured and only
• 14:44modestly perturbed by those tasks.
• 14:47This is what the task on the left
• 14:50that's rest on the right it's task.
• 14:52I would you might be able to
• 14:54point out some of the effects,
• 14:56but the the the overall
• 14:59organization is is very similar.
• 15:02And in fact,
• 15:03this has been quantified very
• 15:05nicely in this paper from Katarina.
• 15:09That shows that those the the amount that
• 15:12the task explains the variability across all
• 15:15of these measurements in this population,
• 15:19these are now individuals in
• 15:20the midnight scanning club,
• 15:21women stand 10 times,
• 15:23they've had 10 task runs done.
• 15:26They've had they've connectivity
• 15:28extracted from each one of those stands.
• 15:31The variability across all of that data
• 15:34that's explained by the task resection
• 15:36versus not most of the variance is.
• 15:39Either common structure
• 15:40across all the individuals,
• 15:42or about individual differences in
• 15:45the individuals network structure
• 15:48that's constant across that
• 15:50across paths for that individual.
• 15:56OK. So there's significant evidence
• 15:59against spontaneous activity being
• 16:01related to ongoing technician?
• 16:04A apparent large scale dynamic
• 16:06content is frequently misattribution
• 16:08of spurious non neural sources.
• 16:11Resting state networks fluctuations
• 16:13are present despite variable
• 16:15states of consciousness.
• 16:16It's minimally perturbed by tasting.
• 16:19It's largely stable across long periods
• 16:21of time within and across subjects,
• 16:24and I didn't really show the
• 16:26the precise evidence for this,
• 16:28but there is some reason to think
• 16:31that also arousal effects can observe.
• 16:3410 account some of the observed
• 16:36changes in responsiveness.
• 16:41OK, having said all that,
• 16:42let me pose an alternative view
• 16:44about how to think about what
• 16:46we're saying if it's not admission.
• 16:51And I will bring up, says the reminder,
• 16:55I think a very salient observation,
• 16:58which is the brain has a constant high energy
• 17:01and it represents 2% of the body weight.
• 17:05Possibly only 20% of the
• 17:07energy consumed at all times.
• 17:10And regional increases in absolute blood
• 17:12flow associated with imaging signals,
• 17:14as measured with the path,
• 17:16are rarely more than five to 10% of
• 17:19the resting blood flow in the brain,
• 17:21even during the most arousing and
• 17:24perceptual and vigorous activity.
• 17:26So again, there is a lot of stuff
• 17:29going on even when we're not
• 17:32asking the land to do very much.
• 17:35OK, well, why?
• 17:37So this is a very simplified way of saying
• 17:43that we don't just need brain activity to.
• 17:47Process information and to develop cognition.
• 17:51We have the need for brain activity
• 17:55to build the brain itself.
• 17:58And to maintain itself through time,
• 18:00I might argue that these are more important
• 18:05properties for brain activity itself.
• 18:08Then those instantaneous uses of
• 18:12neurons for information processing.
• 18:15And at this point,
• 18:16I do want to introduce just
• 18:18a kind of jargon term,
• 18:19but I think it's a helpful way to think
• 18:21about this distinction that there are
• 18:24online mechanisms that the computations
• 18:26that we typically associate with cognition.
• 18:30Things that would instantiate perception,
• 18:34motor behavior,
• 18:36thinking, prospection,
• 18:38rumination.
• 18:38And then there are these other things
• 18:41that we'll call offline mechanisms.
• 18:44That are related to neural
• 18:46activity dependent processes.
• 18:48Kids neural activity dependent.
• 18:50These are as well that you do not generate
• 18:54immediate behavioral outputs and they
• 18:56usually occur after an experience.
• 18:58OK,
• 18:58now I I did not come up with these terms.
• 19:02These are from a long literature.
• 19:06Really amazing stuff over the last
• 19:08decades of folks through studies as well.
• 19:11Looks like very pasaki in particular
• 19:14as well articulated,
• 19:15this point of view and offline
• 19:19payment mechanisms include
• 19:21things like memory consolidation.
• 19:25Generating representations after the fact.
• 19:28And then another important thing,
• 19:30restoring excitatory inhibitory
• 19:32balance and synaptic scaling,
• 19:35what I might call homeostatic mechanisms
• 19:38that the brains needs to instantiate.
• 19:41OK. So I will label this back
• 19:43on this simple model here.
• 19:45We have to build the brain.
• 19:47We have to maintain a brain.
• 19:48Those are offline processes
• 19:50using the brain that's online.
• 19:52OK in the moment.
• 19:55Another helpful ideas about comes
• 19:58from David Marr about learning
• 20:01machines and the brains learning
• 20:03machine that we all know and this.
• 20:07It's simply articulated as
• 20:08a learning machine.
• 20:10Requires 2 alternating phases of order.
• 20:12Would be able to get new information
• 20:15and store that information and
• 20:17consolidate it incorporated integrated.
• 20:20There's a learning phase in which
• 20:22the machine is connected to the
• 20:24inputs and is getting information,
• 20:26and then there's restorative
• 20:27phase in which the machine is
• 20:29disconnected from the inputs.
• 20:31Connection between elements or rebalance.
• 20:33Not going to get into details of
• 20:35that for those who you know, if you.
• 20:37So, uh, I'm machine learning model.
• 20:40This is the principle behind which a
• 20:42lot of these things have operated.
• 20:44You get information, you get some
• 20:47discrepancy, you update the content.
• 20:52Um, so. Now we're going to do some
• 20:56neuroscience basic mechanisms, again the
• 21:00first first few days of neuroscience,
• 21:03but I want to bring it up again.
• 21:05So heavy and plasticity is the concept
• 21:08of course that things that fire together
• 21:12wire together which is most closely
• 21:14associated with concepts of long term
• 21:16potentiation or long term depression.
• 21:18And we all know this mechanism and it
• 21:21requires a synchronous activity in the
• 21:24presynaptic and the postsynaptic synapse.
• 21:26And it leads to changes in those synapses
• 21:31that strengthen the relationship between
• 21:33those synapses for future activity.
• 21:36These are based on cellular mechanisms
• 21:39that transport receptors in the membrane
• 21:41and they engage on a scale of seconds.
• 21:44You know, we we get information and and
• 21:46it and if there's this kind of synchrony,
• 21:49you might actually be development that.
• 21:51But there's an important point,
• 21:53heavy and plasticity alone would not work.
• 21:56We're sustaining a system and this
• 22:00has been demonstrated to you.
• 22:03If you're not careful and you had a
• 22:05process where you had a heavy impressive
• 22:08plasticity mechanism and then you
• 22:09stopped other activity from occurring,
• 22:12you actually would get a positive
• 22:15gain feedback and your your.
• 22:20Activity would continue to grow.
• 22:22In the opposite case, long term depression,
• 22:25it would continue to decrease and
• 22:27it would be good to run away firing
• 22:30in in your synapse.
• 22:32We had no mechanism for controlling that,
• 22:34for regulating the extent of
• 22:36having plasticity that would get
• 22:39included in our knowledge.
• 22:41So we have all these processes actually
• 22:44that we as opposed to heavy and plasticity
• 22:47which is about building up these connections,
• 22:51homeostatic plasticity is about
• 22:53sustaining and incorporating them
• 22:55within the current system that we have.
• 22:58And it acts to counter that happened.
• 23:03One of the main concepts there is the
• 23:06process that will maintain the mean firewing.
• 23:09So I'll just illustrate the bottom here.
• 23:13We have,
• 23:14you know,
• 23:14the case where we've done potentiation,
• 23:16the synapse has now it's got all these
• 23:19extra receptors and it's exciting to
• 23:21do the thing it's going to do next.
• 23:23But actually there is a consolidated process.
• 23:26There's a synaptic scaling process that
• 23:28occurs after the fact that tries to
• 23:31make sure that neuron doesn't have now.
• 23:33Much more excitatory activity than
• 23:35it had before,
• 23:36but it rebalances that activity
• 23:39so that it is not being overused,
• 23:43but it's still reflecting that information
• 23:46that was stored at that synapse level.
• 23:49So now we have a synapse that is.
• 23:53Uh,
• 23:53more precise.
• 23:54There's another mechanism here,
• 23:56again very similar concept.
• 23:58It's just excitatory inhibitory balancing,
• 24:01but we don't want firing rates
• 24:03getting out of control.
• 24:04So if there's an increase in an
• 24:07excitatory happy and mechanism,
• 24:09there will be rescaling within
• 24:11the system to generating for
• 24:13higher inhibitory inputs.
• 24:16So that that doesn't get out of control.
• 24:18OK.
• 24:21Now getting back to this idea of
• 24:24different mechanisms that might relate
• 24:26to the different phases of learning.
• 24:29And here I just wanted to introduce
• 24:32quickly the idea of sharp wave ripples.
• 24:35These are observed particularly
• 24:39in the hippocampus,
• 24:41and they have a very characteristic
• 24:43pattern of frequency content.
• 24:45And what's important to note is that
• 24:48they're present particularly when an
• 24:50animal is still or asleep and what
• 24:52it's doing something walking around,
• 24:54you don't see these things nearly as much.
• 24:57And so the the sleep wake cycle
• 25:00in the same animal is.
• 25:03It's the most well studied example of
• 25:05this kind of toothpaste principle.
• 25:08So in this case,
• 25:10we have in the middle the animal walking
• 25:13around in a cage on either side over here,
• 25:17here and here.
• 25:18That's when the the animals
• 25:20resting and the dots,
• 25:22the blue dots are just reflecting the
• 25:24frequency of the sharp wave ripples.
• 25:26So during rest we're getting
• 25:28all these sharp wave ripples.
• 25:30When it's awake,
• 25:31we're not seeing those things and these
• 25:35are believed to reflect consolidative.
• 25:38Plasticity properties that are
• 25:41encoding information that's being
• 25:44learned during during the main's past.
• 25:48OK,
• 25:52OK, OK, let's.
• 25:55The so.
• 26:00One important thing so we have these,
• 26:02we can have these different phases we've
• 26:05been awake where we go from getting
• 26:07information from the environment,
• 26:09going into sleep and having different
• 26:12mechanisms play out that are integrating
• 26:15that information so that we can maintain
• 26:18our our what we've learned through time.
• 26:22But we actually might postulate
• 26:24that this kind of online,
• 26:25offline discrepancy is not
• 26:27just seen in the most extreme.
• 26:30Sleep wake distinction.
• 26:32But it's actually a existing at all times.
• 26:38And even while you're away,
• 26:41that there is a kind of balance between
• 26:44online and offline marketing online and
• 26:47offline mechanisms throughout the brain,
• 26:49even while we're away.
• 26:52And this idea can be illustrated in
• 26:56the kind of competitive way in which
• 27:00during that on online activity.
• 27:03So in this case, we give an we this,
• 27:06this is a study from ITO colleagues
• 27:08where they've looked at human primate
• 27:11data and human effort marine data with
• 27:13the exact same kind of analysis and
• 27:16then they can show that during the task
• 27:20that background activity is suppressed.
• 27:23So there's this when the animal's
• 27:26not doing something,
• 27:27you actually get larger fluctuations
• 27:30in spontaneous activity.
• 27:32When it's doing the task locally,
• 27:34you're getting a suppression
• 27:36of that activity,
• 27:37and so there's this competition between
• 27:40the what the local neurons are doing for
• 27:44actual information processing and what
• 27:46they're doing to maintain themselves
• 27:49to incorporate that information.
• 27:52And we see the same principle
• 27:54in human fMRI data.
• 27:55During the task periods in the background,
• 27:58activity increases.
• 28:03And this may explain actually
• 28:05some of the effect that we saw
• 28:07earlier with the task force's rest,
• 28:09same idea that that some of this
• 28:11is just about stimulus crunching,
• 28:14that it may not be so much that we're
• 28:17seeing new coherence between areas,
• 28:18though that might be part of it,
• 28:21but we're also suppressing background
• 28:24spontaneous activity and that's
• 28:26part of what's being reflected in
• 28:29those differences between tasks.
• 28:32OK. So this last thing is just the,
• 28:35I think one of the starkest examples
• 28:37of this idea that when we're thinking,
• 28:39when we're seeing spontaneous
• 28:42activity and functional connectivity.
• 28:44I think it's really more helpful to
• 28:46think about it in terms of plasticity
• 28:48as opposed in terms of partnership.
• 28:50So in this experiment we had folks,
• 28:55very generous volunteers who were
• 28:57and not have injuries,
• 28:59but we're willing to put a cast on
• 29:02their arm for two weeks at a time.
• 29:04And we scanned them every day
• 29:06for two weeks beforehand.
• 29:07Then we scan them every day
• 29:09while they were casted for trees,
• 29:11and then we scan them every
• 29:12day and afterwards.
• 29:13And we only have 3 subjects.
• 29:17Not everybody is running to do the study,
• 29:20but what's what we found was really
• 29:24remarkable that during the casting period,
• 29:27there was dramatic changes in
• 29:29the functional connectivity,
• 29:31but the key point here.
• 29:34So you know you see the classic left
• 29:36right motor cortex from the top right.
• 29:39We put the casting on the dominant
• 29:41arm and it essentially goes away,
• 29:43but it takes a few days for it to happen.
• 29:47It's not instantaneous, OK?
• 29:49And then it comes back then.
• 29:54Though I don't know if we go
• 29:56long enough for Omar here.
• 29:58Took a little extra time.
• 30:02I think it wasn't going to be enough sleep.
• 30:05But here's an important control that
• 30:08wearing a cast during the scanning
• 30:10itself does not cause the change.
• 30:13So if you take the cast and just put it
• 30:15on the person and put them in the steamer,
• 30:17you don't see the big effect.
• 30:19The effect is because you've had the cast
• 30:23on for a period of time and you started
• 30:27to engage those plasticity mechanisms.
• 30:30That, that, that are now being reflected
• 30:33in this spontaneous activity that
• 30:36we're observing in the motor cortex,
• 30:39specific to the region that
• 30:41we're affecting with the past,
• 30:42the rest of the brain is not change.
• 30:45It's specific to that region
• 30:47that's being changed.
• 30:51And actually there is a really amazing
• 30:54phenomenon we observed in this what
• 30:57we call these pulses of bold activity
• 31:00that were all that were increased
• 31:03in frequency during the cast.
• 31:05In particular at what exactly
• 31:09this correspond to?
• 31:11Actually don't yet know if they have
• 31:13an obvious like electrical correlate,
• 31:15but we see this is in the Volt signal.
• 31:17We've seen these little pulses of
• 31:20activity that were particularly
• 31:22concentrated during the task and I
• 31:25invited an analogy with some of these
• 31:28other mechanisms that I was talking about.
• 31:31But we have to have to
• 31:33demonstrate that obviously.
• 31:34So in conclusion,
• 31:35the discussion is is not to mean
• 31:38a complete or a clear conclusive
• 31:42explanation of what spontaneous
• 31:44activity is about and it's also
• 31:47not going to exclude thinking
• 31:49something's breaking.
• 31:50Obviously we are thinking and
• 31:54we observe things,
• 31:57but I think it's helpful to frame
• 31:59our interpretations in this way.
• 32:01As opposed to unconstrained partnership.
• 32:04And actually it it makes me think about
• 32:07every study I looked at differently
• 32:09and it changes the way I think about
• 32:12how I might design A future study.
• 32:15And I really loved the Emily
• 32:18Jacobs talk yesterday.
• 32:20It makes me think of how I would,
• 32:21how I would interpret that that what was
• 32:24going on with the dramatic changes in
• 32:28functional connectivity over the cycle.
• 32:30I was wondering,
• 32:32well,
• 32:32maybe maybe this reflects
• 32:35changes in plasticity that occur
• 32:38over the financial cycle,
• 32:40but it it,
• 32:41it engages us to try to come up
• 32:44with experiments.
• 32:45We're interested in studies by
• 32:47things that that relate to these
• 32:50kinds of mechanisms.
• 32:51Things like training arousal separate
• 32:54sensory deprivation as opposed to
• 32:56manipulations of conflict conflict.
• 33:01And obviously we have to do some things
• 33:04to really prove this hypothesis,
• 33:06as it is just the hypothesis about
• 33:09what's going on at this point.
• 33:11I'm not well. Thank you all.
• 33:19Question on it. Wonderful.
• 33:24Thank you.
• 33:28Continuously.
• 33:37So I'll have always.
• 33:40Distracts on the range. And.
• 33:46That's about 5%.
• 33:50On the right side just for that
• 33:52would be something that would
• 33:54and on the left side of the 9%,
• 33:57that's not predictable.
• 34:0710 percent, 90%. Obviously.
• 34:1610%. That's not completely different.
• 34:24Hands up native.
• 34:28Also most of the bands.
• 34:33When you were saying that.
• 34:37Passive.
• 34:39Good news, quencher. Yeah.
• 34:50Support. The idea that you have it.
• 34:56Yeah, no, I so. One the first
• 34:58thing one might think is that well
• 35:01with if this is really all about
• 35:03stimulus frenching that it would,
• 35:06it might be all kind of One
• 35:09Direction that you go from seeing
• 35:12higher correlations to seeing lower
• 35:14correlations in the region that
• 35:16you're a vote activity is you know
• 35:19in the same region you'd see the
• 35:21above activity and that's mostly
• 35:24true but there are exceptions to that
• 35:28and it might have something to do.
• 35:30With a difference between, say,
• 35:33what the default mode network is
• 35:35doing during rest versus what
• 35:37some other system is doing during
• 35:39rest and in terms of are they,
• 35:41which is it online or offline and
• 35:44how would we interpret what these
• 35:46different systems are doing at it under
• 35:49a different a given task state and
• 35:51that makes it more complicated to see.
• 35:54Well, I'm just going to expect one
• 35:56behavior across the whole brain.
• 35:58It might be that it could go
• 35:59in this direction.
• 36:07Right, right.
• 36:11Yeah. So what I mean is when you're
• 36:14when you're in the rest state,
• 36:16people speculate that something
• 36:18that there might be in the sense
• 36:21on the online component of what
• 36:23default mode regions are doing is,
• 36:26is actually happening when you're at rest.
• 36:30As opposed to when I asked you to
• 36:32do an attention test, the online
• 36:35component is during the attention test.
• 36:38So that would lead to a difference in
• 36:40the way you would interpret it. But.