Consciousness Live! with Claire Sergent (S4 E16)

0:00 okay how long have you been doing that oh well we just went live i know people are starting to ask me questions right when we go live but since you've been yeah i'm not a pro me neither obviously uh since i should have given you like the which i did not do so i apologize for that but uh let me just answer your question let me defer that for a second but uh joining me today extra special class uh guest claire sergeant sorry i tried to pronounce your name right but i'll just

0:33 stick with that uh so claire can you just say who you are and what you do uh for those who might not know you yes so hello and thank you very much for inviting me um so i'm claire sergeant in french and i'm i'm now a professor at paris university university de paris um and i've well i've studied biology a long time ago and then and i was always fascinated by the

1:04 subject of mine in general and i discovered that cognitive sciences existed so i conducted um a phd with uh standalone um and uh and and i got my phd in 2005 and then i went to different places at ucl um so university college london and then back in paris uh working with also clinical people like leonel nakash and and then in the end universite de

1:39 paris where i am and all that time i was uh always interested in the general topics of attention perception and consciousness very cool wow okay so there's a lot there now i i'm sorry that i cut you off and did this sort of cold open on you but you were asking me right before we went on how long i've been doing the uh this podcaster series or thing it just passed the three year mark so june 28th 2018 is when i had my first guest who

2:13 was hakwan lao and i used to do a podcast before this with pete mandick another philosopher and i really enjoyed it but i found that what i enjoyed the most was just having the conversation part and then not the taking it afterwards and editing it and adding the sound bites he did all that stuff and so i just kind of uh picked up where we left off and just said i just want to talk to people so okay cool so yeah thank you uh and mostly it's cause i'm nosy i just wanna know what people are up to and um and i do wanna understand uh the views that i don't agree with um from from the point of view of the

2:45 people who defend them so i think that's important for generally but but i want to ask you just a general question before we get to your particular work and um if you don't mind about uh so your interest in attention and consciousness you described your trajectory he ended up doing a phd with uh stan i want to ask you about all that stuff but i'm also curious about if the uh assc played any role um during that journey and and whether you would care to comment on what you think the value or role of that society is for people interested in the things that we're interested in because let me just say that i'm interested in

3:18 the the way that consciousness science is as i see it poised to become sort of a work a day area of neuroscience much as i view what what attention did in the 90s till now so and i think the ssc has a large role to play there um and so i just wonder if you agree or what your thoughts about that if you care to share them you could say no by the way um no i think i definitely agree with that okay so so you mean both from a more personal perspective i mean it's made

3:50 it's it's it really uh made me realize that i belong to a family that had the same language as me the same um with whom i could exchange uh with no uh no need for much translation in what we're working on or censorship right you don't have to watch what you say you could die that's true yeah and as he said and and the other part is probably i i don't know if i'm a good judge but probably it's uh an important association for making

4:25 this um uh this topic of investigation honorable scientific sound scientific and uh yeah i guess it played an important role in that uh and that's yeah did you feel uh just as a so more personally from your journey did you feel when you first started becoming interested in science that consciousness was a legitimate topic of study for science or did you become convinced of that sort of on the way or yeah i'm curious about how you saw it

4:58 at the beginning okay um well okay it depends where you place the beginning ah you are a philosopher i see i am well no a fake one we're all fake ones okay the phrases no um well um i i noticed i mean um i had no specific preconceptions of whether there were uh topics that were

5:33 out of bounds for science and and others that would not uh i just realized uh when i joined before my phd i joined i joined john driver's lab uh studying attention and and hemming neglect patients at ucl and uh i did realize that uh around that time before the 2000 uh the the the topic uh was yeah was considered as maybe not

6:10 entirely scientific and it was better to just simply speak about attention right and and that would do but in the end when you study hemi neglect you study people that uh that do process some information unconsciously right so right so it's just you're just avoiding the topic really by saying it in a way that makes it respectable yeah exactly and then and then um just i mean just looking at the the the experiments

6:43 that were being done and the results that uh that emerged they i mean that was just obvious that uh we were tackling the issue right directly and and raising and on the way raising important points that we had not foreseen before so uh i think i mean the question of it whether it's legitimate or not uh is just waved away by evidence by the work that was being made

7:15 so and then and then that was also comfortable to be uh in the shadow of uh standard han because uh like nobody would say to standard and uh that's not legitimate what you're doing yeah right exactly but then so it helps to have someone that's respected that's also taking the thing seriously yeah the bottom line is okay everyone is able to recognize uh the value of a conclusion when it's

7:48 presented correctly so i think that's the bottom line yeah i like that way of putting it i think that's a nice way of putting it that the the proof is in the pudding or whatever like the work that's being done it justifies the existence of the field we don't need a special plead so i like that way of putting it because i do think of i mean i know this is sort of just sort of abstract but i part of the reason why i like these conversations something about the next generation of you know the ones that are coming up behind us so to speak um and and i want them to feel like there's a place and i know that when i was a graduate student

8:20 part of what drew me to this was there was the assc there was a place i could submit papers and go and talk to people that cared about the same stuff sort of the same things that you were just saying and in other fields there wasn't i didn't notice that in like philosophy of language or just straight philosophy of mine so i i think that's important that there is this identity that we have that we know that people that of course now whether um that survives is the question i guess uh because because i guess some people worry but anyway um well let's let's take yeah some people are worried that

8:52 uh the field is i mean you know i guess i won't name those people but we could deduce who they were but some people are worried that the field is sort of not taken seriously by people outside of it because of the influence maybe the disproportionate influence that certain theories have which imply certain absurd results maybe that ryan would like smite my me or something like that uh so i don't want to i'm not trying to have a debate about that i understand that there are some people who are worried that that from the outside it looks like it's like politics and big names but not real like worried

9:24 um about science okay but you disagree with that i mean but you're you're i think the global work i can see your point but i i'm maybe more optimistic like um i mean first i think it's we so much enjoy and and we can feel it we so much enjoyed these gatherings right uh that we'll find a way to make them live and and you'll notice that i mean i mean by the the very force of the production we have

9:58 uh that's what makes us legitimate but yeah i i see what you mean there are and it's also hard though because like grants and right exactly things like that and labs too though like when you tell a student that comes to you and says i want to study consciousness where do i go and you say europe i guess i mean because we're in the u.s there are people like out now saying yeah studying consciousness yeah there's an u.s europe uh issue probably there as well yes yeah yeah exactly so hopefully things will change but uh that

10:30 so it's good i think that people see what we're doing that we're out and we're open we're not like secretive about it anyway so i appreciate your openness and thank you for sharing but um i want to ask you all about the global workspace theory because as far as i understand your work you take yourself to be defending that view i think maybe you even think maybe it's true or something can i just ask like what is your do you think it's just a working hypothesis or do you think it's something more the global workspace hypothesis oh well it's it has to be a working hypothesis i mean i'm sort of thinking do you think it's

11:03 true like do you think there's evidence that so that you convinces you to think think that is true or more than yes but there are yes but there are as in every model there are there are many things there some of which are untrue and and the and the paper i've just uh uh uh published uh recently is actually uh is actually showing that some of the aspects of the global workspace were not true and i'm very very happy to be doing that because

11:36 i think uh the role of a model is uh to reduce yeah well not only i mean i think not only obviously to be falsifiable is is a very interesting feature for a model otherwise it's not very useful but also i think we really do need to have um a path to put our feet on like right i assume this is it it is like this

12:08 and then if it is like this i'll go that way and that will direct not only give me falsifiable predictions but also direct my thinking right so i think well maybe it's only personal but i cannot make these long strides if i don't have a model to to just relax my mind on something that's like i assume it's settled

12:41 okay and if it's settled then then this oh it can't be so you see and uh and and if uh yeah i like to say that to my students because they go well and if one day we have to to put the global workspace in the rubbish bin i'll say thank you global workspace and so long and that's yeah so maybe i'll have a little tear but i don't know i mean you see what i mean it's really uh and it's really important to emphasize

13:13 also to the young generation that you're not fighting for a model it's it's just a uh like a bootstrap to your mind a limited mind that okay it guides your thinking to what you can test yeah exactly and um but it's true that uh i i think also from from my history but also from being convinced by by the evidence it's true that i've i've mostly worked and i tend to

13:47 uh to cast uh this framework of the global workspace uh on my interpretation of things uh perhaps can i just back up for a second that the paper you mentioned is this the uh the uh bifurcation dynamics paper or is that a different one yes yes okay so i want to ask you about that so i yeah cause i'm glad that you mentioned that it partially it shows something partially wasn't carried out um that you would have expected if the gold workspace theory were true um but so but i want to so and but there's some part of it that was sort of

14:20 borne out so it's an interesting result an interesting experiment and a really cool paper so i want to get to that but i want to ask about what some of the things you were just saying because it sounds like what you're saying is that you don't think that or at least the kind of science that you would pursue isn't like model free or like just data data data data like sometimes you meet scientists and they say look i don't do theories i just collect data um and it sort of sounds like what you're saying is like well you know you can't make sense of what they're doing really if that's what they say they're doing i not and i would agree with that so i wonder if that's how you would if you'd go that far

14:52 um so okay so you mean that i so can you repeat your question yeah so that um you were saying that without a model you don't really know how to guide your thinking or like how to generate hypotheses or predictions or so that but sometimes you meet people who say that they aren't doing that that the uh what they're doing is just collecting data and it's model free and it's not going by yeah it's sort of theory neutral or something like that and i've always been somewhat skeptical

15:24 of that that there is such a thing maybe the theorist themselves is not thinking about the theory but they're being guided in some sense by what they find they should be testing what's predictable what isn't right i mean or so i thought that's what you were sort of saying but maybe i was misunderstanding that's true on the other hand um uh on your on the other hand you could find also very valuable work uh that is um that is not having such a complex uh construction behind it right just

16:00 like uh uh it i mean and maybe people that have this disposition would say i'm i'm i'm without theory uh just uh probing uh this made me curious and i wonder this or that without a big model behind that but then i guess once you want to interpret your data you're already making a mini model yeah exactly i would say once you're even collecting today yes yeah once you design the experiment you

16:33 already have a model sort of i would not go as far as saying uh that uh if you're conducting thing experiments without a clear idea of the models you have behind that is useless i would not say yeah no i didn't meant useless but i meant uh that it's not technically true that you can play without some kind of yeah yeah you are yeah you always have friars even even though you're not aware of them yeah exactly yeah right that's a better way of putting it exactly you have priors even if you're not explicitly

17:05 aware of them um because i think that's what's interesting the last couple of years explicitly i think i mean as a philosopher you know i i like to see that i feel like in the assc in the community there there's been a kind of more awareness about theory questions questions about theory how do we test specific theories instead of just confirming evidence what falsified and just generally questions about philosophy of science which is good so i i'm very appreciative of this um meta philosophy philosophical strand um yeah so i now okay i want to ask

17:39 about the global work space more specifically so there's two things i want to ask you so first of all i know you said that these are models and i'm really glad to hear you say that you'd be happy to see it in the dustbin or you wouldn't be happy but you would as a scientist that's you would predict or expect that eventually you may have to face that possibility um but do you think of the global workspace as like a psychologically real thing which you know involves things like working memory some people have questioned whether working memory is is even psychologically real a natural kind or so i i don't know exactly what to think

18:13 i think these are hard questions but i'm just curious about what what you're is it just a model or do you think we have reasons to map parts of this model onto actual psychologically real features of the mind um so you mean so you mean psychological features or neurological features yeah both both okay because okay um so well because you might you might i mean the gold workspace started to cut you off right when you're about to say something but the goal is that global workspace might be

18:44 something that describes a possible psychology but not human psychology maybe we don't we don't work that way maybe a.i would but we don't so there's a question about does our mind actually work that way and then secondly if our mind is implemented by the brain does the brain do it too so i think they're related obviously but okay so well on the first point um my impression is that um although the the global workspace uh that was like uh proposed by bars um is an idea

19:18 that that's that's really strongly uh anchored into psychological research and about you he uh the functions of a human mind right um as well as some um some uh uh engineering approach into how you might be implementing these functions that use ai for for thinking that's the way i see it

19:52 but if you read bars it's it's very much psychological research of uh and i know you've interviewed him uh recently yeah i haven't had the chance to see the interview but uh yeah but i mean he was inspired by he says that he was explicitly inspired by like the work from ai and so he was hanging out with some cognitive scientists and boxology and all that but uh but but i guess so i so i don't know so do you think that in

20:24 so obviously i mean let me back up five steps what do you see as the main difference between like the stan dehaan version of global neuronal workspace theory and the sort of quote-unquote traditional bars version are they different theories or is the end just an implementation step or do you think of them as distinct saying different things and and how would you explain what they are saying yes okay so i i might be mistaken but uh i see i see two um two two points of relationship

21:03 uh well the first one is is i i really do see the global neuronal workspace put forth by standalone and and jean-pierre as a proposition for a neuronal uh uh implementation of the main uh of bars main idea okay that's okay that you have to uh uh so that that you have to approach the problem from a functional point of view right

21:34 and that in order to um account for the specific functional properties of consciousness that we can observe we need to assume that there is some sharing of information in a functional network so i think um standards and jean-pierre model is just trying to propose an implementation of that um but then that leads them and then and in their wake other

22:08 researchers like me uh to make perhaps uh even more specific predictions about dynamics etc uh more wet predictions i would say i don't know uh yeah and the other aspect that i would like to dive in uh is that uh when you read the bars it's much richer than just uh the proposition of the global workspace but because bars proposes a hierarchy uh that goes up to the self

22:44 that has not been touched upon at all in the global neuronal or not that i know of in the global neuronal workspace so i think we're both inspired from the core global uh workspace um proposition by bars but there are there's a big mind to be digging in as well i think uh in his propositions yeah that's i think you're exactly right about that in fact when i talked with bernie about that he seemed to think

23:17 that global broadcasting was like a necessary condition for experience or consciousness but he didn't think that all by itself global broadcasting accounted for um for a consciousness because he says there has to be other systems that access that stream like the self system which then you start to say well that sounds like a different theory uh no there's no higher order theories but um it's so yeah i think that's that you're right that that's uh one of the key differences is that the way that you guys have developed the theory is that the broadcasting itself

23:49 seems to be necessary and sufficient um so that that there is well but that was i guess should be my next question is do you see it as sufficient all by itself uh simply broadcasting or is there some kind of accessing of the information which is important on your way of thinking about it yeah i guess that's an excellent question because uh and i guess that's uh that's the main criticism to this idea that you can broadcast in uh in a computer you can broadcast in a

24:22 in in various uh systems and if uh broadcast is sufficient then okay then uh we're backed well i guess that's that that would go down to the same issue that iit would have like it's a property of communication that can be uh implemented implemented right in any like a connected network so uh why would it be so special and i and i think um yeah you're you're putting the finger on on something

24:57 that's exactly right what's instrumental is that you assume broadcasting in an already functional network in a in a functional uh manner right uh so you're but then then i need to specify what i mean and then that's where i'm like uh i would have uh difficulty i guess i mean actually i actually that

25:29 that's where i uh i would like to go now is uh which is what are the specific nodes of this global workspace that account for the functional properties of conscious uh access i don't think with uh i think it's time now to um to understand that especially in the

26:02 context of this new step that we have where we uh we probe conscious processing in the absence of uh report in the absence of tasks right the question comes back uh in a new form and we it means we really have to tackle uh what we mean by the the the key functional nodes for this workspace right i mean so there's a lot there i mean one

26:35 just i can't help because i there's you're saying so many interesting things i had a plan and now i was like i'm sorry and it's just too interesting um but it's all there's so many nodes and connections so when you say conscious access i mean one thing i wonder about is what you particularly mean by that and whether you're using that word like as a code word to signal allegiance to like not ned block's interpretation consciousness or something like that or or what you mean by conscious access so why don't you just talk about

27:06 phenomenal consciousness and are reporting about it or something as opposed to conscious access so yeah i am curious about it because that word shows up a lot in your in your work and i just want to understand why you chose that phrase i guess that's true uh i think um [Music] probably initially it was just a way to um avoid the debate actually yeah what's wrong with that we want um do that it's also when you're writing an article

27:37 you don't uh you want to present uh uh results but not get into an argument uh with your words about things that are maybe outside of the scope of the article so i guess that's the basic reason why i use it uh it's true but but uh but that's cool that you asked the question where's the phenomenal consciousness there um my my impression my intuition is that um it it comes with the package like that uh that's uh

28:14 when so if if it is true that um conscious perception let's talk about perception because uh i mean that's what is easy to study and that's what i'm saying but uh conscious i think when uh if if it's true that conscious perception is uh uh comes with uh the the broadcasting of sensory informations to a functional network

28:49 then how can you imagine that it doesn't get to the valuation systems and to the emotional systems so to me it's uh uh to me that's something that was not uh touched upon by people working on the global workspace because they were interested in in complex tasks and not asking about these as uh emotional or value aspects of the of the processing but i mean the major the one of the main

29:26 uh functional roles of of your brain is to evaluate uh whether what you're perceiving is good or good for you or not at that moment right and uh and and that's that's that also is conveyed by by emotions and the evaluation system and interestingly these this focus on uh limbic system and valuation systems comes from uh decision science right i mean the very science that is interested in

30:00 reasoning uh a rational decision uh tells you well yeah but decisions are made from a point of view and that's where where like yeah the self is getting in in in the in the basic sense like is it good for me or not so for the moment it's a big uh so it sounds like what you're saying a bit vague but uh but i i i do think that uh broadcasting uh especially to the limbic system

30:34 uh might might be a natural like bed for uh uh phenomenal consciousness and yeah so this is broadcasting outside the prefrontal cortex to the to to like hippocampal systems are you saying like limbic system is a term i i think people i don't know are unhappy with nowadays i don't know but uh i'm not neuroscientist but do you mean like the dentist hippocampal structure and the amygdala and those sorts of things broadcasting from the cortex

31:06 to there or from somewhere else to there or like from everywhere to what so when you say broadcasting to there i'm wondering where the themes were from well you see that's that's where you uh that that's where i hit my limits because i i i i definitely would like to um become more knowledgeable on this system so i i'm only now also be because i'm i'm getting away from uh the tasks in my experiment uh i i'm i'm keen on looking more into um

31:43 like how yeah how the limbic system uh gets lighted up or not by a stimulus and whether uh you you get these phenomena of uh broadcasting or not uh but okay but my uh to uh to the limit given my limited knowledge on the topic uh it seems to me that the the frontal cortex is very much linked with the limbic system and

32:15 evaluation system in general it depends on which part of the frontal cortex i think yeah or maybe event ventral central me and vitro medial but dorsal like the frontal pole i don't think it's a lot of direction yeah but that's interesting then uh because for example um it seems but we we still need to investigate that but it seems that uh when we contrasted uh becoming conscious of the same information with and without a task

32:46 uh you have some part of the frontal cortex that cease to be involved but others um that needs to be uh still it still needs to be confirmed but uh others seem to seem to still uh be involved yeah so because i mean i'm getting ahead of ourselves here or and that's fine but um because i still i'm trying to understand you're you're talking about all your interesting work that you're doing and i want to talk about it too but i'm still trying to

33:18 like understand the way you think about the global workspace theory and and what it says and um because i never really have honestly understood what it means when something's i mean there's this word you guys use mobilize into the workspace and i'm like i'm not sure what that means um neurologically speaking like there are some neurons that fire then they're connected to some other neurons that fire so what has been mobilized into the workspace is the pattern of like what what i mean so there's no thing that has been moved into the

33:51 workspace at the neural level or do you think that there is a like something that can be interpreted as a representation or something that literally is like broadcasted so i so that's what i one of the things i guess i don't really understand like at the nuts and bolts levels like what does it mean to broadcast because for example my background is like i did some wet work and i worked in animals in grad school uh but looking at very briefly looking at um long-range connections between the density gyrus and the cortex and we're looking for synchrony to sort of confirm ltp and things of that nature but there the

34:24 sense of broadcasting makes sense to me because what you have is a bunch of neurons that are firing they get entrained into a theta rhythm and then they sort of all start firing in the theta rhythm and then you can make sense of what it means for that information to be carried on the theta cycle or whatever but is that the sort of thing that you mean when you talk about broadcasting or not you but i mean the global workspace theorists because it doesn't see i don't see the i don't see that kind of nuts and that's sort of like what's happening in the spike chain level when you move something into the workspace and does this question

34:56 makes is this not a good question or the wrong question uh yeah i think it's an excellent question i mean you we need to specify what what is meant by a broadcast yeah so um and what you're like do you agree with like the edvard moser and the like you know theta broadcast and trade like that whole kind of picture of broadcasting or are you guys working with a different notion that's just to be specified in terms of the mechanism at the neuronal level um well i think first i think that you you can have these um a kind of communication of or synchrony

35:33 so i think i do think that uh synchrony is uh is the language is the communication mode across uh across regions right uh but i guess i guess this can exist also for unconscious processing right exactly like unconscious percepts are bound together and synchrony probably has something to do with that yes although although there are some indications that um that uh probably becoming conscious of a

36:08 stimulus or at least having attention uh honest stimulus helps link different aspects right different levels of representation of this which which i think is true right um but uh but then i think um yeah i think it's a cool question because uh that's uh that's in in these details that you we can examine the robustness of the proposition or not um so what so uh

36:46 what i i think this global broadcast uh mean is uh is that um there there are some maybe the attentional system that uh some systems that represent a a bottleneck and that at this level there will be competitions between the influence of various sensory representations and that

37:21 uh the the winning representation and that's very barr's uh view that's very much barr's view uh the winning representation the one that if if you want maybe it's better to say that influences the target the target regions most will will then uh be gated to uh other possibilities so i maybe it's a bit of a career future but let's assume that you have uh competing

37:55 uh visual representations they are getting to uh the the attentional system uh let's say uh in the parietal cortex and one is for some reason for the task or for yourself more interesting than the other or wins the wins the the the competition um and winning the competition is also i guess um a coupling between uh attentional regions and sensory

38:29 regions then the idea is that this opens the gate to uh several other um systems including language systems even if it's a visual uh stimulus that is not directly connected to language areas this uh this opens up a a communication channel to other um other functional systems that will make the processing of this information more

39:03 more memorable and more uh flexible uh in in and stabilize uh stabilize the the system around this representation but obviously this representation is not going to be the same at the sensory level and at the other levels but the uh the influence uh we can trace back the influence of this specific representation to the the the functional system that is responding

39:36 uh-huh so would it be wrong to say that to summarize what you said is that being in the global workspace means there are representations in these various systems that causally trace back to this representation in the cortex because yeah or is that is it more than that because what i'm thinking is that there's two pictures one picture is like it's information not states right not it states in terms of the brain so let's let's imagine that there's six neurons and representing that there's something out there involves these three neurons

40:08 firing in pattern n so these three neurons firing in pattern n is is the representation of some property in the external environment aligned whatever okay let's whatever so now that thing needs to get in the global workspace this representation we would say goes in the global workspace so what does that mean that the information goes in the global workspace so this state stays here these neurons firing in that pattern stay there but then the information is copied or something or is it rather that somehow those systems get connected to

40:41 this state and that state influences what they do but there's not a representation up there which is doing something which this thing caused and that's doing the work so those seem to me to be sort of two different ways of specifying what you mean and i've heard people say both actually it seems like that's true i would rather i mean your second proposition seems much more to make much more sense uh neurophysiologically right right and that's why i just i'm not trying this isn't an ambush or anything i'm

41:14 just trying to get clear on the theory because that does seem like yeah one is like very computational there's information and one is very like biological like there's something so there's a big difference i think in the way people are thinking about what the global workspace theory is but i'm not sure yeah but that's that's a good point i'm not sure um yeah you're a philosopher you know where to put your finger so that's it it's a nerve [Laughter] yeah i don't mean to i'm just trying to understand sorry i didn't yeah yeah you know i mean what some people tell me

41:49 when i ask these questions is i don't care i just like the data here's some data let's talk about the experiment and there's a and you could tell me that and i would move on but you know uh no no i i like to hit that nerve okay cool yeah but but yeah but the problem is i'm i'm not sure i mean actually i i do realize that i'm not sure that um the the first uh proposition like the representation is sent to another area uh makes

42:22 physiological sense i don't think so i mean uh different areas do process information they don't just acknowledge or re-represent but right that's my view and um yeah and i think right so i mean do you think that part of the problem here is is that we don't really know what all the spiking and neurons and things like that uh what what they're i mean we say we talk a lot about them doing stuff but we don't really i'd say you know what i tell them i teach intro to neuroscience

42:54 um for non-science majors at my college it's like a general ed class and and one of you know joe you try to tell them about neural representations what what's the only the best thing we can say is like synchrony and then intensity of spike firing trains like you know you press really hard you press lightly and then you get some rate code like and then you kind of say okay now take that and extrapolate to everything it's like how do you how do you do that well i don't think we're really at the point where we like can say all those what that stuff means or even a spike trains or someone's in the chat asking about neurotubules and

43:26 it's like i don't know like how like what do you think about like glial cells and spike like in the neuroscience of 100 years or glial cells going to be the star and we're going to be like no glial cells are like is it still going to be spike trains like i just don't like if you look at the long range of science it's it's it's interesting to think about a thousand years from now whether they will laugh at us like the way we laugh at aristotle which is something i'm constantly worried about yes oh no they won't be i mean but why not but uh we would anyway be the ones who

43:57 let them to where they are sure yeah we'll say that um yeah but uh yeah okay i see what i said okay i see what you mean like the brain representing something uh we're we're not even at it yet we we're not understanding that yet well i guess part of my question because then i say mattias is in the chat as well asking sort of a similar question what i was asking i think he's asking it better like he usually does but uh what what i'm trying to get at is is

44:30 what is whether you think we can like tell the difference right now between these two ways i put the global workspace hypothesis like so all the same stuff would be lighting up and blah blah blah but one interpretation is there's information which is being sent now the information is over there and it's being there's a representation now which is doing stuff the other one is no they're just as coupling and there's but once so which one is the are we looking at i don't can we tell do you think uh oh okay i'm well spontaneously i would say number two no hesitation but

45:04 i'm saying that realizing that i always uh use the wordings of number one okay it's true but if i think physiologically uh i think uh yeah the number two it makes a lot of sense it's it's not sending some information it's just having uh yeah having some communication so uh well like influencing the the the way uh representations are changing in

45:36 in area b uh thanks to so so like conceptually it would be almost more like channels of water that are flowing which you could divert into different paths to move things around rather than something computational something um what you would just because it sounds like when you describe something more mechanistic it's just there's pathways that are opening up and things are flowing now in this way but not really moving from one place to another it's just now there's this path that exists between that thing here and that thing there but not a new

46:08 representation of this thing over there which is what mattis is asking he's saying do you believe that sensory representations are re-represented and i think what you're saying is not just simply re-represented that they cause things up there which may be our representations but not re-representations yeah i think they cause things upset up there and they uh they block other things up there i think the the competition aspect uh is is occurring as well um and um yeah and i think

46:44 from from the representation um like um how the brain represents things um topic i think we're we're moving up uh nicely now with the with current uh uh experiments um like beautiful experiments obviously in the hippocampus looking at much more complex patterns than the spares coding observations that were made in the

47:16 individual cortex and where we and and also recent studies that show that representations notably in higher levels such as prefrontal cortex do evolve during the course of a task right from uh from something that is directly readable as like probably resembling the the the sensory input and then it seems that the role of this area is just to cast it uh to another plane which is the

47:49 decision plane so i think um i think i would look at this literature to uh to shape uh my understanding of what is this broadcast and i think what is broadcast is a strength of influencing uh in various areas but obviously when you send to area b it's doing its job it's not re-representing it

48:22 it would not mean anything i think right so i mean part of the reason why i think this matters and and people and matthias asking the chat sort of question where i was leading hopefully too was because it matters for this current like uh craze or fad of decoding um which i people i think people are getting i love multivariate pattern analysis and actually one of the things that's so cool about your new paper tons of pickleball is the mvp on erp which i hadn't seen before so i don't know if that's new or how new it is but i've done a little bit

48:54 of ee erp work but never i didn't even dream of m like multivariate pattern that's like a cool idea so anyway i want to ask you about that um but not right now i know i'm being very annoying i'm sorry um but the reason that this matters though is because there's a lot of like you know enthusiasm for testing predictions of these theories and some people say decoding is a way to go and some people have said global workspace theories are committed to decoding from content of conscious experience from prefrontal cortex so one thought i had was like the way you're putting things

49:26 now it's you still could get a kind of decoding if you knew what these causal relationships were but it wouldn't be a decoding of the content of conscious experience from the prefrontal cortex it would be a decoding of the indirect measure of the thing with down there and which is conscious or something like that on this view right so i guess that's another way of asking the question of which state is the conscious experience states on the view the ones that are caused or the ones that's doing the causing both both yes because it's the dynamical

49:58 relationship that puts you in the state where um your brain is mostly influenced by that and your your psychological function and and perhaps your behavior is mostly influenced by uh disinformation right so that's often uh often we're tempted to um [Music] to actually get away from the workspace

50:33 idea by saying well everything is sent to the prefrontal cortex and that's exactly uh the the kind of thing that that bars wanted to avoid with the idea of global workspace that's i mean that's a negation of global work space to say things are sent to the to the to what uh homunculus prefrontal homunculus that's exactly what what bars actually tried to uh get around because that because that's that's a problematic way of seeing

51:06 things functionally so um so are you saying that these two states are really one complex state are you saying that there are two mental states two states of the mind which are conscious or that make a contribution to the conscious or do you think this is going to be what you're talking about i think it's a single state yeah so the broadcasting the content that is broadcasted and the receiving of the contents that is broadcasted are one continuous state slash process

51:39 on this view well uh probably uh then now that you force me to specify i would say you're not you're not sending a content you're creating another content that is uh that's that's connected to the initial one and and also and what's interesting also is i think that i think that there are episodes in the stream of uh of consciousness um and and that's uh that it's not

52:14 a snapshot uh that we can identify some homogeneous uh consciousness episodes around um an initial cause that uh that will um lock into this state and then evolve until we're done with the thing and then we're we're moving on to another one but uh but i think within each uh becoming conscious of an information we we are in a dynamical state but but uh yeah i think also one

52:51 one interesting aspect that's emerged with the proposition of neuronal implementation of the workspace is uh is this focus on the dynamics that has led us to uh observe and then theorize that uh there is this bifurcation right there is such a thing as yes now i can tell you that it's this representation that is mostly influential in the brain

53:24 uh now i can tell you but then i i don't think that there's one moment in time where uh you're conscious i think it's an episode it's it's but but i think there is an entry point because absolutely uh phenomena yeah i mean that's so interesting because it does seem like then a lot of people have been thinking of how you might test this the ideas that you have in in the wrong way like some of the stuff that stuff that address allegedly is trying to test this theory wouldn't be testing it like you

53:58 know um could you decode content from the pfc alone versus could you decode it from the sensory cortices isn't a question that's going to really address the sort of theory that you are are talking about right now right is that is that really the upshot of this well at least yeah if you're trying to test the global workspace you you need to test the connections right i'm saying that but i'm not testing connections for example in the last paper yeah exactly well no actually in the last paper you're establishing that there are that there's this disassociation or trying to establish that there's this dissociation this bifurcation

54:30 which is an important aspect of of the theory and i promise we are going to just kidding so um yeah this is so interesting because i'm learning a lot and i really value you taking your time to explain this to me and help me understand it because i do feel like i i this isn't the picture that i got from reading and i have made it i may have read stan's book i've read your papers um i don't really feel like i was i'm not saying you weren't saying this not i'm just saying i wasn't it wasn't what i was picking up so this is changing my mind about how i think about the global wakes workshops theory yeah which is good i think you're you're

55:04 actually i mean it's supposed to have this conversation you're making me say ex explicitly these things that i do realize i think and i've never said explicitly actually time for a book from you clara that's what it sounds like to me [Laughter] but that's really cool so um someone is asking like uh uh queerto is that how you say your name i don't know um in the chat how you became convinced of the global working space uh what so i and you did mention that you sort of thought there was evidence

55:36 that personally that you thought that this was like of the theories of consciousness maybe the one that's the best we have right now would be a way to put it or closest to what you think is capturing something important so i i wonder like i know is it is it could i venture the guess that's is it the general idea that you know when we do the contrast of method that we find this this this dissociation late activity from people when they see it early activity when they say they don't see it that's in common with when they so is is it that general line of reasoning which has convinced you or is

56:08 it something else um change blindness or what what what overall what just everything being presented or is there specific things which kind of led you to go yeah that's that's the thing that's leading me um i think okay so uh i think there are some um theoretical aspects that really convinced me and there are some um and there are some um empirical results that really convinced me

56:42 uh from the theoretical part um i'm i'm absolutely convinced uh at least by the approach uh and and it made so much sense to me when i read the bars book the approach to say okay consciousness is a is a word of human language that that is aiming at a functional property and let's describe this functional property and then get

57:16 back to what's making this uh it's it's i think it's very compelling so i'm very much into trying to uh not not break this chain of understanding right from from the neural uh underpinnings you're trying to find and the the functional properties you're trying to describe but it's true that it has led us to focus

57:48 too much on um tasks and it's cool that people that view things differently uh had tell us okay but then you're just describing task processing but still this does not deter me from saying um i'm not buying your your theory if you're telling me uh uh if you're not showing me the link between your um your neurophysiological description and

58:21 a functional property that appears in psychological terms that's like that's a personal conviction yeah so can i just stop you for a second real fast so because at the beginning of this and i want to get the empirical reasons that convince you a second but the beginning of this you mentioned that part of the idea was that you know that that the gold workspace makes sense because consciousness is all interconnected in this way and you know it's not just isolated perception and then emotion over here it's like emotional perception and perceptual emotion and like it's all mix

58:53 and workspaces sharing information makes sense of that idea is that part of the comment like the common sense appeal that you were that you're talking about right now or is that something else the intuitive appeal i mean the idea that that it captures this idea that there's this vast interconnection of experience which you would expect no i i'm not sure i think uh that it's uh it's mirrors or reflects uh

59:28 phenomenology i i don't i don't like the uh the the the way of just making analogies like for example when when i was i was never really convinced uh by some um arguments about uh linking different features of of stimulus you know how how do you create a perceptive of oh no you can't see it you can't see it

1:00:04 you can see your glasses though [Laughter] that's okay yeah how how do you create a the concept link uh black with the shape of my uh et cetera so that's an old psychological uh concept and i was never convinced by uh just the argument of saying look we see that there can be synchrony between different modules and and we're trying to psychologically we are trying to understand

1:00:36 how these two properties are linked synchrony looks like a link so that's the basis and that's just an analogy analogy and uh i think we we need to that's a good that's a good start but explaining is not just finding analogies i think so yeah but then because the reason why i ask is because so suppose like i could agree with you i mean i'm so much skeptical of this idea but i could i could agree with you that we want to do something functional like what you

1:01:09 were saying we don't want this continuity we lose the continuity with the way science is done functionalizing things looking for the mechanisms to implement those functions etc but then why there to global workspace as opposed to there to recurrent processing which i take it is also a functional notion or higher order representation which is also sometimes or any other functional notion um why global workspace as the functional notion that was the yeah yeah yeah that's what i thought the connection stuff one might be addressing which is why i asked that

1:01:40 question like why why if you agree with you at that starting point what leads you from there specifically to the global workspace version of a functional theory of consciousness yes okay so i um i think my um my reasoning would be different uh if we're discussing local recurrence uh versus discussing uh higher order theory okay so for local recurrence

1:02:15 it has some functional role but it seems to draw the line between unconscious and conscious um a bit too low for me so from what i see and and and and this theory raises from beautiful neurophysiological work about um about the the role of recurrent processing into building up uh percepts in well or uh representations

1:02:52 that uh that really really match uh the reported process so that that's true uh still i think these uh there are some there are some building evidence that uh these uh recurrencies can occur uh even when people deny having conscious uh uh being conscious of the stimulus and i think that's the the most problematic part of uh this uh uh

1:03:26 proposition uh which has led uh some people to say that when uh neglect patient that is shown a face uh in in his or her neglected field um and denies seeing the face well if we see that the face is still lighting up uh uh fusiform face area then uh this this person is phenomenally conscious of the stimulus then that's where the breach to uh functionally is uh is

1:04:01 to me and and and i can't um i think then then you're speaking about uh another type of consciousness that i would uh spontaneously speak about if someone i mean to to me the person having the conscious um percept uh has the final authority on what is consciously perceived right so that's i think that's very problematic i i think there there are a lot of evidence that you can have local recurrence

1:04:34 uh and unconscious processing but that's just i think the level where you're you draw uh the the line yeah so just really quickly to clear something up someone uh queer to the same person in the comments is asking he keeps think interpreting what you're saying as invoking some kind of meta metacognitive um state or something and it's not i mean yeah that word the metacognition is a word that's so abused i don't i don't really like it anymore uh but there's a sense in which what you're talking about is metacognitive

1:05:07 because there's access to all these systems cognitive systems that's kind of meta in a sense but it's not metacognitive in the traditional sense right that's i mean we we agree that you're not invoking some kind of metacognitive yeah necessarily but it's true that um that's something i'm not completely settled about uh it's true but at least the kind of consciousness you're talking about isn't metacognitive consciousness like you're not talking about because that's the higher order stuff if if anything right that you're aware of these states

1:05:40 or something like that or so if metacognition is involved in some sense it's not like that kind of stuff no that you're talking about yeah no i don't know i i think it's more basic in a way yeah but i i mean my my intuition is that these um these these mechanisms that uh helps the the system to lock into um a semi-stable state have allowed

1:06:16 uh the the the evolution of uh more complex than refined functions such as metacognition right you need you need to settle the system on uh on a specific um content so so as to allow the emergence of uh more complex functions right there's got to be cognition before metacognition yeah yeah that's that's well put yeah so it look because uh because i i

1:06:51 want to get clear on what you're saying about what i interpret as phenomenological overflow the ned block kind of argument that there could be phenomenally conscious states what you can't access or at least aren't accessing at the moment um and victor lama recurrent processing theorist uh is well known for defending a version of that and i think what you were just giving was an objection to that way of thinking saying it severs the connection with functionalizability that you think is useful and i'm thinking you have in mind something like the cohen and dennett

1:07:23 maybe style paper argument in terms of cognitive science i don't know if you remember that enough um so they think that you know they're basically just arguing for you're saying that if you're gonna do science it's gotta be functionalizable so if you're gonna say something that's not functionalizable then it's outside the realm of science or something like that yeah right so all right so i i wonder if on your own global workspace view you could have a version of of if it could empirically be found out that there were some like this so suppose that you were convinced of global workspace theory like you maybe are but even more

1:07:55 convinced like we had some really good we tested it we had reason i think it was actually like you know probably true or like the best model we have uh then suppose you say you identify with um phenomenal conscious access whatever word you want to use the experience whatever with uh global workspace dynamic gw to call it g um that's in general intelligent so we don't want to do that just call it you know whatever call it i don't care x so then you find x when people say that

1:08:28 they don't see the thing did you falsify the global workspace theory yes oh okay that was easy because i would have thought that you would have said no i'd be i would have become convinced that there's consciousness without um without access okay but no i mean i mean well it depends what what's your final object if your final object is to describe uh what's behind this what's behind this

1:09:03 [Music] word of human language maybe we maybe we understand this word different differently but uh uh if if my goal is to understand what uh uh what's behind this uh this word of human language uh your conscious if you if you deny that you're conscious of the ex and uh or of why and your ex whatever yeah i know i put it to abstractly i know

1:09:35 if you if you tell me and and i have no good reason to think that you're lying to me then i mean i have every reason to throw away my model instead of your report i mean no yeah that's interesting so i mean even if you had like you know in every other case the the model had made successful predictions withstood falsification and then you became generally convinced that yes this is sort of on the right

1:10:07 track and then you found out this happens okay so how many how many times yeah i know you have to really pile it up patient i i just want because it does seem to be cons i mean to be an empirical issue but what it sounds like you're saying is no it's a conceptual issue um so what i'm trying to get at is whether it's a conceptual or an empirical issue that's really what the game i was trying to play is um it's like could there be any any empirical evidence that would

1:10:38 convince someone like you that yeah there was something that you experienced that there was something that it was like for you to stay to be in but you didn't know about it um i don't know there's something behind me right now i don't know what color it is okay it doesn't seem conceptually confused to think that it's got a color before i look so i don't know so i i i i i'm not asking seven agenda i'm literally on the fence i don't know which way to go here okay uh i think uh i get your points so um i mean it's

1:11:13 it makes sense then then then we reach um another age in my in my view where uh data has brought you more than you expected uh because you've reached and and i think that's in a way what uh lama thinks uh he's at exactly yeah where data have shown you an interesting difference uh that does not show up in in language uh uh in an obvious way exactly

1:11:49 uh yeah that's what he writes and then we will come into the the the real history of our investigation uh where where we'll have to use terms that don't come from naive psychology right um but then then nobody will be able to say the flag this is uh consciousness no this is consciousness well these are so so maybe then we'll have to to adopt the yeah the p

1:12:23 and the a or um what what's the connection with the spontaneous words of consciousness right yeah so the spontaneous now i see i think i've been misunderstanding slightly what you meant when you said spontaneous because it sounds like what you're saying it's like naive psychology philosophers would call that you know like folk psychology or just like the way we think about the mind before we have to revise it by doing science or something like that and what you're saying and what i'm liking when i hear it is that to do science we got to start with the naive

1:12:56 psychology because that's the thing we're investigating that's the thing we're interested in that's where we're here so if we're led to abandon that thing by doing science then you get into a real pickle because you have to ask yourself hmm could i start by studying something that's right over there and if i know it doesn't exist and i have to or should i stick with the thing that i knew at the beginning and say okay well then what i this can't be a good theory of it and that's a real tough philosophical question so i yeah i think that's interesting yeah i think it's it it's really helpful to think about uh something that's almost done which is

1:13:31 uh explaining life right right i mean at some point we know that there's no well one single there are different levels of explaining life that are very conceptually very different all interconnected um so i think it really helps uh um looking ahead for these big topics like consciousness yeah yeah it's so interesting okay so i see uh my usual tendency to ramble on has led us to the point where we've been here for a long

1:14:05 time and we haven't got to talk about your cool work i mean i i've been really enjoying this conversation but i want to really talk about this by stable dynamics paper you have so many interesting papers so maybe somebody will come back and we'll just go through all your actual empirical work um but but i do want to get your opinion about this dynamics uh the bi-stable bifurcation sorry not vegetable bifurcation dynamics paper um so can you just give us an idea about what you were doing and how you set up the experiment what you thought you were gonna find and what you did how it worked out okay yes so um i think it dates back to um

1:14:42 one of the first week of my phd a long time ago wow i love it when those long-term projects come to fruition after so many years that's awesome so uh we were having lunch with uh standard we just started to be my supervisor and uh liz sparky uh who's a very good friend of him and she was asking him but stanley well when you when you speak about your theory you always seem to

1:15:14 uh think that there are two very different uh processing styles the unconscious and the conscious one you seem to to consider that there there is a frontier etc um and he said yes obviously and i said uh maybe we can test that uh because i was really struck by uh the uh the fact that you can use scales uh to ask people to rate their pain and that they would use it really uh

1:15:49 really consistently etc and i said oh maybe we can test uh whether uh there there is a gradient there using a scale why not try and and and that's what what was really cool with the uh being a phd student with standard and he said oh cool yeah let's do that and i'll give you all you need to do that etc and uh and that's uh okay and that's where it really started uh as a topic instead of just a common way of speaking about the

1:16:23 things can we find a separation and and i know there are other traditions that uh really tried to uh to know whether there's a separation or not in a notably in a signal detection theory where they decided well for the moment we consider that there's no separation but anyway so and we started to investigate that in the attentional blink and it seemed to uh uh it seemed that at least in the intentional blink uh when people

1:16:56 were shown uh the same uh the same um stream of stimulation but on some trials they would tell us i've seen it and on some trials they would tell us no i haven't seen it on the scale we could see like really two groups of trials although the the they were only repetition of the same stimulus right but two groups of responses sometimes they would say yeah there was nothing and sometimes they say oh yeah i've seen it quite well and then uh in the brain which we we

1:17:29 could see a big difference between these two types of trials with the same external stimulation and it echoes lots of other psychological experiments that do not directly tackle that okay but that's that was in the intentional blink and then and then um in the meantime i had started studying uh patients with disorders of consciousness and you can't do attentional blink experiments with these people and and i thought okay but

1:18:04 then and then other um colleagues had other experiments uh with uh with masking etc that seemed to indicate also that they would find bifurcations in these kind of experiments and i thought okay it's it's uh and notably uh anton delkul and sid quide etc and and i thought okay so maybe if in the end if you take your your proposition seriously then first you can find bifurcations

1:18:39 with very simple stimuli and second uh your um [Music] that means that you should uh you should find it uh with the audition auditory stimuli that you can use in patients right across modalities right that's the yeah across modalities so i thought okay uh if you take this seriously that could be a very interesting tool for probing um patients and we need to to see whether that generalizes

1:19:13 generalizes to audition generalizes to simple stimuli and and the idea behind the protocol was also that this protocol could be uh then adapted to the patients and that's what we're doing now so yeah and that seems i've seen some of that work it's really awesome by the way so okay cool thanks and so so uh that's um so that's what that was the start of this experiment um and at the around the same time

1:19:48 i was attending the assc at the paris and now to go to shia and others were making the argument that in the end if we wanted to find the neural signature of conscious uh of consciousness we needed to uh show it in the in the absence of a task in any absence of a report and i i said okay that's definitely the the right step to go from uh um

1:20:23 healthy awake uh responding volunteers to uh patients that are not going to respond to me but i still want to read out whether they are conscious right and and so i want uh healthy volunteers responding healthy volunteers not making any tasks to see the the what's common between the two and then go to the patients and that's that's how uh that's how the we we came to this protocol uh where we use

1:20:55 very simple stimuli and uh and people uh and auditory stimuli so uh and and so you're listening to a continuous noise like this and within that noise you will have ah the french vowels because they were french participants so by the way uh this signature of consciousness only applies to french people it's important to understand that no i'm joking

1:21:26 yeah i almost took him serious for a second you got me you got me i was like yeah yeah yeah you wake you up it worked so uh and so yeah they were listening to these vowels in in noise and uh and the same participant came once for a session where they told me yeah i hear a or uh e and uh and uh and how well have you heard it etc

1:21:58 and on another session they heard the same thing but instead of rating how well they heard the stimulus uh they had quiz questions etc uh they were not doing a task on the siemens and we had a full range of uh uh stimulus intensity so that we can pro could probe that really the the like the the response function of the brain to external stimulation under task or no task and this was all

1:22:31 done with eeg though so that they they weren't in a scanner or anything like that no no that was done in eg also uh uh still in having still in mind what uh a technique that we could easily implement in patients and that's yeah yeah and uh yeah and with with eeg we wanted to look at the like the response function of the brain to these uh stimuli and um and what we did in this experiment is really to specify

1:23:03 the dynamical model that we meant uh when we were talking about bifurcation because in previous papers it was a bit uh like qualitative we were saying oh you see there's a disruption in the in the in the functional response etc uh here we really wanted to specify what we meant by uh by a model uh so that we could also fit uh the data to the and and so what's particularly new in

1:23:37 this uh in this article is that uh specifying the model has led us to uh acknowledge that um in order to um demonstrate that that there is or or there is not a bifurcation in in dynamics between processing the same information unconsciously or consciously in order to prove that we we cannot only rely

1:24:08 on averaging uh averaging the signal across trials right because uh if you want to see a bifurcation so that for exactly the same stimulation sometimes you're here sometimes you're there it will show up in the variability of activity across trials so this is not a way of saying like p300 based research is like not gonna get at what you're looking at or like anything based on traditional erp stuff and numerous trials

1:24:43 well um it could for example for example uh in way back in or attentional to our initial attentional blink paper when we sorted out uh trials uh according to the response of the participants when they said that they had seen the stimulus we would see a p300 when they say they have not seen a siemens we we would see no p300 uh okay that's that's

1:25:19 seemed compelling but that's not i mean uh it could still be um an effect of averaging indeed right as you say so in the end if you want to to prove that there are two groups you have to look at individual trials and see whether indeed uh it uh it arrives spontaneously from the distribution of trials that there are two groups of trials right exactly so you have to see if that so it's the strategy that you used was you trained the uh like a uh pattern analysis

1:25:56 thingamabob on the data and see if it could sort the trials in the scene and unseen versus given the data is that what you did is that the way it worked um yes so um yeah it's it's it's um so there are two stages first we did uh multivariate pattern analysis uh but um a very classical one uh classical from the early 2000s exactly

1:26:30 okay standard uh just to uh have a summary of uh the activity that was uh um that really reflected the siemens so we're just uh doing a multivariate pattern analysis uh to summarize the activity related to the stimulus and so we classify uh stimulus was presented versus not presented okay so that's the multivariate part and now we have uh and now we have a

1:27:07 summary of the activity at each time point after the stimulus for each trial and then we fit our model and we ask the data and try by trial data we ask do you look like this and and so this would be uh the this would be x would be uh different levels of uh of the stimulation right and and here i would plot uh the activity so uh bifurcation dynamics would be okay

1:27:40 when when the the stimulus is really low uh you only have just baseline activity when it when you're at threshold you have two possible levels okay and when you're well above threshold here you only have one possible level you're conscious every time okay right uh so that's one model the other model is like this okay okay continuous okay and then we're fitting the two models we're looking at which is explaining trial trial

1:28:14 data best uh and we see that this one is explaining the data best uh in the early phase up to 300 milliseconds this one seems to appear uh after 250 milliseconds okay so that's the first observation and so is that evidence that the global workspace gets going at that time that 250 milliseconds because that seems earlier than what i've heard other people say ah okay um well that that matches what we observed uh during the intentional blanket as

1:28:46 well but i think um also it's it's also very possible that depending on the modality or the type of stimuli you're able to trigger these mechanisms earlier or later right i don't think we we can really commit to a specific uh timing um i mean like some people have tried to argue against the global workspace on the basis of like timing of consciousness basically and yeah what you're saying is we can't really do that yet no no and and i think uh both uh lucia melanie

1:29:20 and aru and and lionel nakash and i yeah and probably many others have uh argued in this direction that uh when you're when you're splitting um activity according to whether the subjects uh tell you that they were conscious of the stimulus or not uh you will see differences that can be even before the stimulus because you will capture uh whatever influenced whether you're going to become conscious

1:29:55 of the stimulus or not so you're capturing all the neural correlates of consciousness but these correlates can be precursors of concern so so that's when you're when you are approaching the problem in this way sorting out according uh uh participants report i i think it's great it it has helped us make great uh uh great uh um advances uh but uh you can't make uh a a strong argument

1:30:30 on the moment where it appears i think right okay yeah interesting it's so cool but so isn't like the surprising one of the surprising findings is that so you do have evidence for these bifurcations but not where you thought they would be or is that is it too strong to say you were surprised by where they were uh uh you mean um okay yes so uh you mean when when you have no task yeah yes so the way i understand your paper is what you did is you you did it when you're doing the task

1:31:02 you did it when you're not doing the task and then you try to subtract in method and see what's left over basically and yes yes and then you found oh it's like in these areas that maybe you're somewhat surprising from the point of view of the called workspace theory i don't know but yeah yeah that's true so if we so actually i was so excited when i saw the data uh because i i was i was uh i have to say i was um for once i was quite agnostic about what would happen in the absence of a

1:31:34 task so in the presence of the task we do see uh first these uh um negativity the the bifurcation seems to correspond like first to uh negativity and then the p300 it's an obvious one like you have a nice big positivity here uh okay no no surprise and then uh in the absence of the task i can remember when the when the student came back with the topographies uh that uh the p300 was completely

1:32:10 abolished and uh yeah no that's cool that's so cool because it's not like just it disappeared there was something else and i thought that's that's amazing i would not have put my uh how do you say in english my hand into fire or whatever yeah you wouldn't have you wouldn't have guessed that yeah i wouldn't have guessed that actually uh we still see late activity although people have nothing to do on the stimulus we still see

1:32:44 activity between 300 and 500 even 600 milliseconds after the stimulus although people have nothing to do on the stimulus right that's when they become conscious well that's for the stimuli that are way above threshold that we know that they spontaneously become conscious of the stimulus so i thought wow that's cool so that there is something like spontaneous late activity even if you really don't care about the stimulus so uh um right so that debunks the whole line of argument that spontaneous late

1:33:19 activity is solely the result of task performance reporting demands all that there's some components that isn't related to that yeah yeah i think so and uh uh but also that debunks the the idea that p300 can be a a a signature of consciousness uh i think it's the signature of uh consciousness plus articulating that to executive system for uh the execution of the task i mean i'm when i was a student but way back when p 300 was like the uh

1:33:52 the oddball yeah it's like yeah you get a patreon when you notice a difference so clearly it seems already obviously sort of associated with something else exactly exactly but that's also the history of how we constructed these uh these experiments that it became uh rather associated more generally with conscious processing but indeed uh that's the the defining of the p300 was the the oddball and the and it was a difference between two stimuli that were all kind of conscious and that were both

1:34:25 conscious so indeed uh yeah i know i always found that weird because i was like why how did this ever get like where it came from is sort of anyway so that's interesting but so now where where do you find the the the bifurcation like neural neural neural atomically where is the interesting bifurcation in this task specifically for the french people the thing is yeah for the president i'm only speaking as americans we do it different all right i just want you to know [Laughter] okay so thank you the name of my people

1:35:00 we come in peace exactly hey we only freedom fries yesterday was the fourth all right [Laughter] i'm joking right so um uh so that's eg yeah so uh uh so that that really needs to be backed up by uh by experiments in fmri and that's what we're uh trying to build up now

1:35:34 and then do some localization of eeg yes yes so that's eg source reconstruction right that's what i meant so uh source reconstruction uh in eg needs to be uh taken with a pinch of salt because uh uh you have to make um you have to make assumptions to get at a solution right it's gotten a lot better in the last 15 years or so i'd say i think yes but uh you when you when you're doing it i'm giving you an insider's information when you're doing

1:36:10 it when you're doing it you see that you can tweak parameters you i mean you see that you can feel that indeed some parameters where will influence your result in one way or another so that you can feel the fragility of what you're uh you're looking at right but okay that said uh you can i mean you can i think it can give you a rough idea uh but a very rough idea and if you guys did

1:36:43 intercranial too right i thought i read that there were some uh so um we're currently doing it but uh we're uh we haven't published it yet i see okay yeah intracranial and uh yeah and intrapreneur we already have some data and uh and uh but then i mean the the electrodes are not necessarily in all the areas we'd like to probe right and then we'd like to and then we'd like to do fmri as well by the way claire when can we do full

1:37:17 cortex ecog we're waiting for [Laughter] you want to be the first one yeah right exactly no thanks and actually uh and that's true that's the uh um a difference of the french or the european we have like death electrodes and you have eco uh surface i didn't know that oh interesting okay yeah yeah but uh right

1:37:50 so and but for the eg source for what it's worth uh the the eg source reconstruction in this study seemed to indicate that the the focus of the difference between the uh the activity the late activity uh with the task and without a task uh was mainly uh executive area and uh actually motor premature uh sm uh supplementary motor areas so it made a lot of sense

1:38:23 okay right we're not sure okay and it doesn't make a lot of does it make a sorry to interrupt you and i've seen doing that a lot lately i'm sorry but does it make a lot of sense from the phenomenological point of view that the experience that the person is having of seeing the things or whatever hearing i guess in this case the things um is uh the primarily sma and motor cortex that seems like uh it seems like uh no i mean is this a different am i confusing what thing you're talking about uh yeah okay let me specify so if you uh

1:39:00 if you look at what's different when you not only become conscious of estimates but you have a task to do on the simulators i see then you see the sms classification uh motor clarification okay but in both cases so that's what we're trying to argue in the in the paper maybe uh when you become spontaneously aware of a stimulus with no task to do on the simus what happens is the appearance of a

1:39:34 global playground so a global playground would be this phenomenon of uh locking into a stable state for a moment around a specific representation but for no specific purpose not articulating this to um motor planning notably oh i see so that's the playground so it's just for fun it's not doing any work it's not in the workspace it's just playing yes exactly yeah and and then if you add some task and you

1:40:08 articulate this uh episode uh with um the necessity to to take a decision and to uh to have a behavior related to this information then you you have a workspace but maybe the workspace is a playground with some added modules that helps you uh execute the task yeah interesting and and that global playground if you if based on what you have right now it's not in the prefrontal cortex

1:40:41 it seems like in the parietal cortex or another area yeah it's true there are although you notice that we emphasize some frontal sides yeah is that the one but that's um yeah i think i mean again i think i'm okay i i'm just curious to see what what comes out uh from fmri right with this approach uh um but

1:41:15 i mean that's that that links us to the the prefrontal debates right uh my like my guess is that i mean this enormous part of our brain if there is something like a global global playground there's no reason why uh there would be a diffusion of an information that would be stopped at the at the central circus i don't i don't see why that would happen but uh okay

1:41:48 maybe yeah so i i can see i mean maybe dorsal lateral prefrontal which has been a a big focus for global workspace maybe it it switches off we don't see it in the source reconstruction here so maybe it switches off but i mean i cannot see how uh the the frontal cortex in general uh could be like completely not interested that yeah i'm wrong it's a huge piece of real estate like it's not it doesn't care about consciousness at

1:42:19 all seems weird [Laughter] yeah so i i think that's a good point um but i i think that that's way the things that you're talking about are also interesting empirically they're cool like the techniques are awesome but conceptually they're cool but but it also just illustrates how much how much further we have to go because some of the most basic questions i feel like we're not even able to formulate right at the neurological level or like what would count as evidence and yet there is this kind of assumption that we can oh it's all decoding it's just like it's getting the friend to code into code and then we'll know and it's like it's

1:42:52 it's sometimes i wonder if this uh we got a couple steps before it doesn't mean we should give up i'm not against science obviously but i think that there's sometimes unrealistic expectations about like you know narrowing the field falsifying theories like given what we know now uh can we do that it's not clear that we're yet there but yeah but it's it it's good to give it a try absolutely yeah exactly and we what else are we gonna do right yeah exactly but i i just

1:43:25 i do worry that there's this tendency to like be like you know for example like for the prefrontal thing we're just talking about there's this tendency to say oh well fmri null results therefore no prefrontal involvement in consciousness and it seems like that that so that line of argument is just like really um [Music] we don't know enough for that to be very convincing it's a piece of evidence i'm not good i don't think we should dismiss it but i don't think we know enough to say yeah all right let's give up on the proof on cortex because null results or like that you should move us in one result so anyway i don't want to get on my hobby

1:43:58 horse but uh it does strike me that there is this kind of enthusiasm for bad arguments uh running around not here not now but uh but but uh and i'm guilty of it too you know whatever so because if you read these common papers unconsciousness like the first paragraph is like straw manning every other position and then like moving on to what you want to say not you but i mean like i do it too so because it's hard to get the nuance and the detail of all the all these positions right um yeah yeah that's uh but it's what what you're describing

1:44:32 um sounds like like the rule of the game i mean we have to play i mean but maybe we should uh yeah and uh be more honest i probably you have to get uh to the discussion to be a bit more honest i don't know yeah no i'm all for that you know and obviously it's easy for me to say i don't have to ask anyone for money uh to do my job but you know scientists they got to convince people that you know what they're doing is important and interesting and people that give you money to do it and that's like and we'll cure schizophrenia

1:45:06 yeah exactly you got to apply and that's the other thing we haven't talked about is how you are trying to apply these ideas as a like a way of you know seeing whether patients can recover from minimally conscious states or all of those sorts of things which are consciousness meatery which this seems like a really cool way of addressing those kind of questions yeah do you see any competition between this view and like the iit i because iit i i see the iit people as also being interested in bifurcations they explain them in terms of they so they might look at your data and say ah what you've shown is that

1:45:39 in the unconscious there's not integration and what you're hooking onto is a signal of inter information integration or something so do you think that you can rule out that interpretation or so oh no i don't i don't think i i can or i would like to i don't i mean i don't know i i think um [Music] there are there are lots of commonalities between um the ideas that developed by iit and uh and the ones uh

1:46:12 from the global workspace um but a major difference is uh this functional perspective that i was talking about i think i don't have that uh yes and and i'm not ready to uh take that step that they do take that uh any connected system with this specific kind of connection is going to be conscious right for me you have to connect uh things that

1:46:43 do some information processing or that do like uh have a functional role right exactly um so given what you just said i wonder maybe you don't want to comment but do you have any do are you optimistic about this stan giulio collaboration adversarial thingamajig um and by optimistic i mean do you think that the results can actually adjudicate any of these things we've been talking about or help move this discussion forward or will each

1:47:15 side just find a way to explain the results in their own way and say yeah more work needs to be done um i'm sure it's it will end up i mean the i don't think each of the theories sufficiently specified so that uh one will not be able to say okay but and i think that's fine i think i think that's uh but maybe i'm wrong maybe there there will be one saying okay yeah right i doubt it very well no

1:47:50 really i mean seriously okay i don't know but uh i mean again i almost i don't care i mean uh if that's just uh for me that's just a machine to have us go forward so but do you think it is a way to go forward though i guess is really what i mean i think that that's what i'm most excited about uh for example iit uh for all the reservation i have uh about the um

1:48:24 sometimes esoteric aspect of it uh i think has led to beautiful uh experimental results massimini's uh experiments uh showing the the effect of uh tms uh and and how a pulse of tms will propagate differently depending on your con conscious state i think that's beautiful data the interpretation can be searched and such and such uh is that is that the basis of the zip exactly is that what you're talking

1:48:55 about right the basis of the zip and zap thing that they do it's your word for tms or no i i thought well never mind i uh i didn't mean any retreat but i thought that what they would one way they want to test whether someone is conscious knows by tmsing them and then measuring the spread of it um and then seeing so so that's what i was thought you were talking about but maybe yeah yeah that's exactly that and and these experiments that have shown a different way indeed in different uh for

1:49:28 example comparing uh wakefulness and anesthesia have shown that uh yeah the propagation of a tms pulse is going to be completely different in the brain that's awake versus not awake uh i think it's a beautiful result and i'm grateful for iit or or whatever to have brought about this uh this idea you you know what i mean so yeah exactly so and that's you know because people that are there there are different views here but uh if a theory leads to something useful

1:50:02 then it can't be all bad i should hear part of what you're saying is we've got something useful out of this theory so even if it's wrong or i don't like it it's worth having it around because it's giving us something to think about a new way yeah and and in the end the theory is uh almost necessarily wrong i mean it's okay i start with this idea i mean uh now theories are all like ladders meant to be kicked away eventually when we get a better theory the higher ladder so i agree completely people get too married to their theories but so i'm glad that you because it sounds like the optimism of this conversation is wonderful because it's

1:50:34 like we don't agree that science precedes one funeral at a time like we don't just have to wait for all these old views to die off because the people who support them die out like you're saying we can be reasonable and like adjust our views to the evidence and not be irrationally committed to our theories and defend them at all costs and realize when maybe we should reject them those are all great things that i wish some more scientists would say even if they really sorry you think it's not the case at the moment that people are too too too married to their

1:51:09 theories i think that there's a cynical way of thinking where that yeah that uh i mean that that's saying that science progresses one funeral all the time is sort of the idea yeah people they spend their whole life defending a theory and then you know the reason why people stop talking about just because they kind of die and then people you know there's no one bugging us about it anymore so we kind of move yeah but it's not real progress like we haven't discovered anything it's just like that annoying person who always said you know iit is now no longer here so we did not talk about it um and that's a real cynical view i think where everyone's just in it for

1:51:41 their own you know and but but i'm not saying i really believe that but i'm saying i know people who believe that and i don't really i'm not 100 convinced it's not the case so but it is nice to hear people saying the things you're saying like no we're not that way we science we we are not just theory-minded persons we're open to revising and we have to recognize that we could be wrong and probably are wrong so it's just nice to hear you saying that someone that's not a philosopher recognizing these things that's just nice anyway i do notice the time i've kept you forever

1:52:13 i i feel like we get this conversation going another six hours but i have my swimming lesson for my six-year-old i have to go to so i didn't have to remember but i do wonder at the end claire was there anything that you really were expecting would come up that you wanted to talk about that hasn't come up the answer can be no but uh i just like to check ah um okay well first i i'd like to say i'm very grateful for this discussion oh thank you yeah i think it uh yeah i knew that it would be so great to speak with a philosopher because uh

1:52:46 uh i mean like you're you're uh how do you say that um um you help you help uh conceptions come out of my mind oh thank you yeah socrates said he was a midwife of ideas that's sort of that yeah midwife was the term i was trying to find very traditional philosophy i think yeah yeah yeah a very uh beardy midwife but still my pandemic beard survived the panda

1:53:22 yeah so so i think that's really cool uh and and that's really a fruitful thought for me uh and uh yeah and i hope we'll have other conversations like that and from the uh the evidence uh and from the the other for the other question uh do i have anything i wanted to speak about uh yeah we we didn't mention this other aspects of what i've been doing which is uh trying to probe whether um you can become

1:53:54 conscious of something retrospectively ah interesting and i think uh yeah so in a nutshell um if he uh if you're committed to the idea that becoming conscious is not occurring during the early phase of sensory processing but it's occurring when you look into this communication mode uh then you're led to uh uh the prediction that uh if if there can be some

1:54:31 unconscious sensory traces in your system and you can uh after the fact uh managed to uh connect these traces to the workspace uh then you could become conscious of a stimulus not at the moment where it was presented but and at another um arbitrary moment where you triggered this that's awesome workspace so that's part of my uh of what i'm doing uh as well and we've

1:55:03 we've got some evidence that indeed you can be presented as stimulus you were not conscious of it and then i i tell you oh did did you notice what happened here and you said oh yes i see x but x was actually uh was actually presented uh like 400 milliseconds before etc and we're we're looking at the neural um uh correlates of that and i think that's uh i think that might be uh these kind of phenomena

1:55:34 might reconcile the um uh the the the limited aspect of conscious access that we know from psychology like it seems that conscious access is just processing one thing at the time and the more uh and our impression of the richness of the environment uh because uh because then we we see a mechanism by which we can trace back two things that we could not uh handle in the moment but that we're still

1:56:07 important so this flexibility in time i think is another interesting aspect that we did not have a time to cover but that's fine well hopefully we'll have time another time when uh you can come back because i'd love to have you back and talk about that and that is that the retrodiction stuff post-diction is that post-diction stuff that you're talking about yeah so um so the specific uh work i've been doing i've been calling retro perception but it's really really close to

1:56:38 other work uh called post-addiction indeed okay yeah very cool yeah no i will i guess we i would love to talk about that that stuff's super interesting so i'm gonna have to have you back i do have to run because i am in the pool here so i appreciate your time this conversation was fantastic uh and just hang out for one second i'll say bye off the air but uh if you don't mind um uh ending the stream now oops trying to end okay we're clear