WEBVTT 00:00:00.017 --> 00:00:05.017 If this works out, then Tony owes me a lot of beers. 00:00:08.797 --> 00:00:11.977 This is the Convergent Science Network podcast. 00:00:13.937 --> 00:00:17.737 Leading researchers in the domain of neuroscience, brain theory, 00:00:17.897 --> 00:00:22.277 and technology are interviewed by Paul Verschure and Tony Prescott. 00:00:22.997 --> 00:00:29.017 All right. Okay, it's Paul Verschure with Tony Prescott for the Convergence 00:00:29.017 --> 00:00:30.237 Science Network podcast. 00:00:30.457 --> 00:00:35.417 And today we're here with John Lisman, who was speaking at our summer school. 00:00:35.717 --> 00:00:39.057 And John, actually, this is a really special year for you because it's exactly 00:00:39.057 --> 00:00:43.797 20 years ago that you published your paper with IDIARD on tetragammacycle. 00:00:44.017 --> 00:00:48.797 And 20 years later, you're as excited about it as you were 20 years ago. 00:00:49.217 --> 00:00:51.617 So why is this such a big deal? 00:00:53.257 --> 00:00:59.397 Well, ideas are nice, and it's much nicer when you know that they're true. 00:01:01.837 --> 00:01:06.677 And I think that that's what motivates me. 00:01:06.917 --> 00:01:12.737 Actually, some people often will say, oh, you know, that's such a great idea, 00:01:12.757 --> 00:01:17.457 and I think it's so beautiful, irrespective of whether it's true. 00:01:17.457 --> 00:01:21.737 And I understand that some people feel that way about their work, 00:01:21.817 --> 00:01:23.337 but I don't feel that way. 00:01:23.537 --> 00:01:27.317 I feel that it's only beautiful to me if it's true. 00:01:27.677 --> 00:01:32.377 So if it takes 20 years and it turns out to be true, I'm very happy. 00:01:32.997 --> 00:01:40.957 Okay, but now just your happiness is not necessarily convincing others that it is true. 00:01:41.297 --> 00:01:46.057 I guess there's an empirical base for that, or not? 00:01:48.617 --> 00:01:54.337 I don't know whether other people are convinced. I haven't polled them. 00:01:54.577 --> 00:01:57.817 Maybe you guys will tell me what you see as the holes. 00:01:58.537 --> 00:02:01.617 And, I mean, I could talk about some of the holes myself. 00:02:02.597 --> 00:02:07.417 But in terms of representation, that is, in terms of what you can record, 00:02:08.657 --> 00:02:14.237 it seems to me completely convincing that different information is held at different 00:02:14.237 --> 00:02:18.217 theta phases and that it's not a continuous phase spectrum, 00:02:18.537 --> 00:02:24.637 that it's a discrete phase spectrum, and the discreteness is organized by gamma. 00:02:25.237 --> 00:02:29.997 So that seems to me experimentally clear. Now what, 00:02:30.957 --> 00:02:34.477 you know, somebody could argue is that I don't even believe that this is a code 00:02:34.477 --> 00:02:42.217 until you show me that some downstream network or behavioral system is reading 00:02:42.217 --> 00:02:45.577 this, and that therefore, 00:02:46.377 --> 00:02:51.197 if you put some altered information in a particular gamma slot that the animal 00:02:51.197 --> 00:02:52.037 will behave differently. 00:02:52.097 --> 00:02:57.477 That would be a reasonable next step. And so maybe... 00:02:58.968 --> 00:03:02.568 One would really have to show that before one could claim victory. 00:03:04.068 --> 00:03:10.448 But I wouldn't go that far. Perhaps I think for listeners we need to say what we mean a bit more. 00:03:10.708 --> 00:03:13.788 Okay. Because not everyone will have watched the talk. 00:03:15.188 --> 00:03:19.328 So let me try and briefly say what I think you mean. 00:03:19.448 --> 00:03:24.248 So we have these slower brain rhythms and these faster brain rhythms. 00:03:24.548 --> 00:03:29.328 Theta is between 5 and 15. Okay. Is that right, roughly? 00:03:29.888 --> 00:03:34.808 Most people would. We don't really know exactly how to define these things. Okay. 00:03:35.228 --> 00:03:41.468 But that's not so bad. Let's take that. And then the gamma range is roughly 00:03:41.468 --> 00:03:43.108 three times that, four times that? 00:03:43.648 --> 00:03:48.688 Well, maybe from 30 to 100, let's say. Okay. So really quite broad bands. 00:03:49.688 --> 00:03:56.468 And the proposal is that these are fundamentals to processing in different areas of the brain, 00:03:56.468 --> 00:04:04.948 and that they combine together so that perhaps gamma is maybe, 00:04:05.128 --> 00:04:09.548 when you say that we're thinking in discrete ways, we're thinking in the gamma 00:04:09.548 --> 00:04:17.708 cycle peaks and that the slower theta waves is modulating what can happen in the gamma spikes. 00:04:18.168 --> 00:04:24.408 Is that right? I think the fundamental idea is that some item, 00:04:24.528 --> 00:04:27.028 whatever you want to call it, it could be in the hippocampus, 00:04:27.028 --> 00:04:32.928 a place, is represented by the set of cells that fire in a gamma cycle. 00:04:33.428 --> 00:04:36.888 So we have a window there of, let's say, five milliseconds. 00:04:37.988 --> 00:04:44.308 And now we come, and we can call that the first gamma cycle within a theta cycle. 00:04:44.488 --> 00:04:47.228 And that general concept is called nesting. 00:04:47.568 --> 00:04:55.248 That is, within a theta cycle, you have room for maybe six or seven gamma cycles, the faster rhythm. 00:04:55.768 --> 00:05:01.548 The faster oscillations are nested within a cycle of the slower oscillations. 00:05:02.956 --> 00:05:10.216 So now that I've said that one place would be represented by spatial code in the first gamma cycle, 00:05:10.516 --> 00:05:18.636 we can say, well, is there evidence that different information is represented 00:05:18.636 --> 00:05:21.656 during the second gamma cycle? 00:05:22.556 --> 00:05:27.176 And the answer is, I think now, completely clear that the answer is yes. 00:05:27.176 --> 00:05:33.356 And so what you wind up having, you know, is some sort of data formatting system 00:05:33.356 --> 00:05:38.956 such that within one theta cycle, which lasts, you know, in the order of 150 milliseconds, 00:05:39.436 --> 00:05:42.636 you get to send a multi-part message. 00:05:43.176 --> 00:05:46.096 There's seven, roughly seven parts to this message. 00:05:48.096 --> 00:05:52.456 And possibly more, actually. And conceivably more and conceivably less. 00:05:52.516 --> 00:05:56.716 So, I mean, we don't know the exact frequencies and how they might vary. 00:05:57.176 --> 00:06:03.496 So you're proposing that the oscillations is providing a computational role, 00:06:03.596 --> 00:06:07.756 for instance, of chunking or packaging up information. 00:06:08.156 --> 00:06:13.516 Maybe in distant parts of the brain, you can form things in a package where 00:06:13.516 --> 00:06:19.296 if they fall within the same theta cycle, they're in some sense part of the same package. 00:06:19.596 --> 00:06:23.396 Well, if they fall in the same gamma cycle, they're part of the same package, I would say. 00:06:23.476 --> 00:06:28.396 Right, but then there's a hierarchy here. So there's the theta forming. 00:06:28.396 --> 00:06:35.296 And so then you can form a much bigger message, like all that would be contained within one theta cycle. 00:06:35.376 --> 00:06:40.996 So what's a specific example of this? This would be an example of the whole 00:06:40.996 --> 00:06:44.296 information of a transverse path. 00:06:45.744 --> 00:06:51.144 So, a transverse path would involve multiple locations between where you are and the goal. 00:06:51.364 --> 00:07:00.264 And now we have clear examples of a whole path being played out within a theta 00:07:00.264 --> 00:07:02.724 cycle. But it's not a continuous process. 00:07:03.784 --> 00:07:08.804 It's discretized by gamma. But that's the more recent work on hippocampus. Yeah. 00:07:08.924 --> 00:07:13.324 So, and I think your original inspiration was not really linked to hippocampus. 00:07:13.324 --> 00:07:16.704 The original inspiration is much more linked to ideas about working memory, 00:07:16.944 --> 00:07:21.204 about if you want a sequencing of memory, that memory has a certain capacity. 00:07:22.604 --> 00:07:29.024 Yes, you're absolutely right. I mean, the original work was quite confusing 00:07:29.024 --> 00:07:30.524 to people, and that was fair, 00:07:30.744 --> 00:07:39.064 because the data that inspired this was data about theta and gamma observed 00:07:39.064 --> 00:07:41.684 in the field potential of the hippocampus. 00:07:41.684 --> 00:07:48.144 And then all of a sudden I started applying this to principles of working memory. 00:07:48.904 --> 00:07:53.264 Most people would not think of working memory as a hippocampal function. 00:07:53.504 --> 00:07:58.004 So there was kind of a sleight of hand in that original work. 00:07:58.684 --> 00:08:04.424 And, you know, maybe it was lucky that it turned out not to be wrong. 00:08:04.864 --> 00:08:10.844 But what was really the first empirical breakthrough that you think you found to support that idea? 00:08:12.844 --> 00:08:19.844 Well, I think the first strong suggestion that different information was probably 00:08:19.844 --> 00:08:23.804 encoded in different gamma cycles 00:08:23.804 --> 00:08:30.924 was John O'Keefe's discovery in the hippocampus of the theta phase code. 00:08:31.504 --> 00:08:39.504 So you say, well, why am I so excited about this recent work if O'Keefe had 00:08:39.504 --> 00:08:47.604 already shown known that the phase of firing of hippocampal cells within the 00:08:47.604 --> 00:08:48.944 Theta cycle was so important. 00:08:52.982 --> 00:08:58.602 What he didn't show was that the phase was discretized. 00:08:59.962 --> 00:09:03.542 And so that's what this recent Foster paper shows so beautifully. 00:09:04.442 --> 00:09:09.562 So now we know that you can't just have any old phase within theta. 00:09:09.742 --> 00:09:13.102 You can only have certain discrete phases within theta. 00:09:14.782 --> 00:09:19.262 So that, in a sense, really proves that you have to have something. 00:09:19.262 --> 00:09:23.622 And, and the, the foster proved that the discreteness arises from gamma. 00:09:24.622 --> 00:09:29.862 So that really settles the issue. You have, you know, on the order of six or 00:09:29.862 --> 00:09:35.102 seven discrete chunks of information thrown at you. 00:09:35.542 --> 00:09:38.262 So in the, in the foster case, they were back in hippocampus. 00:09:38.842 --> 00:09:42.842 Yeah. And, and you, you jumped forward then 20 years again, because this actually 00:09:42.842 --> 00:09:43.882 was published this year. 00:09:44.202 --> 00:09:48.562 Right. Right. And what you show there, so I have a rat or the rat navigates 00:09:48.562 --> 00:09:53.002 through an environment and what they've managed to do is really very precisely 00:09:53.002 --> 00:10:00.362 align the play cell response while I'm moving through this environment with 00:10:00.362 --> 00:10:02.122 the TET and the gamma cycle. 00:10:02.762 --> 00:10:07.642 Well, they weren't moving through the environment, but this was the readout 00:10:07.642 --> 00:10:10.102 while the animal was still. 00:10:10.102 --> 00:10:16.542 But yeah, so in a sense, this very precise work was made possible by the fact 00:10:16.542 --> 00:10:24.722 that they had so many cells that they were able to decode position on a very fine timescale. 00:10:24.802 --> 00:10:30.862 And it's only because they could decode position on a very fine timescale that 00:10:30.862 --> 00:10:37.902 they could find out that actually the animal is thinking about position X and stuck in position X. 00:10:37.902 --> 00:10:43.122 And then 30 milliseconds later, suddenly it's thinking about position x plus 1. 00:10:43.722 --> 00:10:50.182 That's the amazing result, that it jumps. And so that's what discrete means. 00:10:50.442 --> 00:10:56.062 It's not a continuous process. And so that's supporting this idea of a discrete 00:10:56.062 --> 00:10:58.402 phase code, of a theta-gamma code. 00:10:58.722 --> 00:11:04.362 I mean, would you, so when you say discrete codes as opposed to continuous codes, 00:11:04.362 --> 00:11:09.602 That makes me think of a standard computer CPU, which obviously goes in clock 00:11:09.602 --> 00:11:12.022 cycles, which are determined by an oscillator. 00:11:13.002 --> 00:11:16.182 Are you thinking this is the brain's clock cycle? 00:11:18.062 --> 00:11:27.482 Well, let's talk about the concept of a word in computer framework, right? Right. 00:11:27.742 --> 00:11:31.222 So what is exactly a word in computer? 00:11:31.222 --> 00:11:42.002 Well, I mean, it's a certain number of bits, and the bits are ordered, right, and, well, 00:11:43.502 --> 00:11:50.122 automatically we have to say that, and they're ordered in time sometimes, right? 00:11:53.099 --> 00:11:56.119 So, what are the analogies to the theta-gamma code? 00:11:56.979 --> 00:12:05.199 Well, one huge non-analogy is the information content. 00:12:05.539 --> 00:12:11.739 So, this elementary unit in the brain, in the computer, is just zero or one. 00:12:12.719 --> 00:12:19.239 But what is the information content within one gamma cycle? 00:12:19.619 --> 00:12:22.839 Well, it's almost infinite. it well i 00:12:22.839 --> 00:12:26.539 would say slightly different if you think about a standard cpu your 00:12:26.539 --> 00:12:29.559 clock cycle is how many instructions you 00:12:29.559 --> 00:12:32.279 can do one instruction during a clock cycle and then 00:12:32.279 --> 00:12:35.519 you can do millions of those within a second in a modern cpu 00:12:35.519 --> 00:12:41.039 but the instruction set also changes so that a modern cpu has a much bigger 00:12:41.039 --> 00:12:48.459 instruction set so rather than just say uh adding one one together you can now 00:12:48.459 --> 00:12:52.919 do some very sophisticated numerical computations in an instruction. 00:12:53.239 --> 00:12:55.119 So that makes it faster. 00:12:56.679 --> 00:13:01.419 And if there is any analogy at all to what's happening in the brain, 00:13:01.579 --> 00:13:04.419 and I just want to push you on this to see if you think there is, 00:13:04.559 --> 00:13:10.119 then the clock cycle in the CPU is really saying that, 00:13:10.759 --> 00:13:14.359 the way we're going to work is that we're going to do instructions one after 00:13:14.359 --> 00:13:19.199 another sequentially, and we're going to do them at this speed and within that 00:13:19.199 --> 00:13:23.279 each time step we can just do one instruction, 00:13:23.479 --> 00:13:26.859 maybe apply one operator to something that's in memory. 00:13:27.459 --> 00:13:33.679 And in the brain, presumably what you would, if you're going to agree with that analogy, 00:13:33.819 --> 00:13:36.919 which you might not want to, you would want to say that in certain parts of 00:13:36.919 --> 00:13:41.559 the brain at certain times there appears to be something like a clock cycle 00:13:41.559 --> 00:13:45.079 which is organizing computation into these sort of discrete steps. 00:13:46.339 --> 00:13:47.859 Okay, well, you know, let me. 00:13:49.399 --> 00:13:53.059 Talk about computation for a minute, and then you can tell me whether you see an analogy. 00:13:53.959 --> 00:13:59.339 So first of all, let's look at what happens within a theta cycle. 00:13:59.519 --> 00:14:04.859 Well, we have the first gamma cycle, and we have an ordered set. 00:14:05.099 --> 00:14:09.059 So already this buys you an interesting analogy. 00:14:09.139 --> 00:14:13.299 You have an ordered set over 60 or 70 or 80 milliseconds. 00:14:13.339 --> 00:14:17.139 You have this gamma cycle, the second, the third, and fourth. 00:14:17.139 --> 00:14:21.779 And so you get order out of that one. Without theta, you wouldn't have order. 00:14:22.839 --> 00:14:26.439 Now let's look at what happens in a gamma cycle. 00:14:26.699 --> 00:14:31.259 So first of all, as I was saying before, there's an enormous amount of information, 00:14:32.179 --> 00:14:33.759 represented within a gamma cycle. 00:14:33.859 --> 00:14:38.859 This is the brain's way of doing things, is that I'm going to coordinate the 00:14:38.859 --> 00:14:44.099 firing of the million neurons in the dentate gyrus during one gamma cycle, 00:14:44.199 --> 00:14:46.959 and some percentage of them, let's say one 00:14:46.959 --> 00:14:50.599 percent are going to be active and that's going to represent the 00:14:50.599 --> 00:14:54.899 information that i'm representing in that gamma cycle and so 00:14:54.899 --> 00:15:02.979 that's you know how that's you know probably a million bits of information because 00:15:02.979 --> 00:15:06.619 you've got the spatial pattern you know so many different spatial patterns that 00:15:06.619 --> 00:15:14.519 you can form given a million cells um so this is you know just the opposite of a digital computer, 00:15:14.579 --> 00:15:19.119 which maybe in one step is zero or one, here we have such a high-dimensional. 00:15:21.196 --> 00:15:28.896 So the next question is, you know, what kind of computation can you do within a gamma cycle? 00:15:29.076 --> 00:15:32.996 Is it just the cells of fire that represent X? No. 00:15:34.736 --> 00:15:40.416 In a paper that Paul and Cesar and I published, 00:15:40.896 --> 00:15:48.476 we showed that actually it's not unreasonable within that gamma cycle to do 00:15:48.476 --> 00:15:54.796 what could be called pattern completion or could be called an attractor operation. Why is that? 00:15:55.796 --> 00:16:04.396 Because, let's say that the network receives input about a pattern, 00:16:04.496 --> 00:16:09.016 and that pattern is slightly corrupted and incomplete. 00:16:10.176 --> 00:16:16.596 So, let's say 10,000 cells should fire, but of the 10,000 cells that should 00:16:16.596 --> 00:16:22.216 fire, in fact, you know, only 5,000 are triggered. 00:16:22.216 --> 00:16:29.296 Well, it only takes two or three milliseconds for those cells through their connections, 00:16:29.896 --> 00:16:37.096 with other cells in the network and specific synaptic weights that connect these 00:16:37.096 --> 00:16:41.956 axons selectively to the cells that are part of that 10,000 pool. 00:16:43.076 --> 00:16:50.136 And as a result, two or three milliseconds later, bang, the cells that didn't fire. 00:16:52.109 --> 00:16:56.949 Will fire. I mean, the cells that are part of the pattern will fire, 00:16:57.089 --> 00:16:59.289 even though they weren't fired by the external input. 00:16:59.489 --> 00:17:07.789 So here's an example of pattern completion occurring within two or three milliseconds, i.e. 00:17:07.829 --> 00:17:14.889 Within the 20 or 30 milliseconds that is allotted to each gamma cycle. 00:17:15.009 --> 00:17:20.569 So yes, you can do computation within a gamma cycle. 00:17:22.249 --> 00:17:25.849 So, okay, now we can step back from this and you can give me your opinion. 00:17:27.829 --> 00:17:31.449 How would you compare this to the way a computer works? 00:17:31.809 --> 00:17:37.569 Well, I think the analogy might be with something like a graphical processor, 00:17:37.649 --> 00:17:43.909 which can do lots of parallel operations, for instance, on an image in one clock cycle. 00:17:43.909 --> 00:17:47.969 So that the the only thing 00:17:47.969 --> 00:17:50.749 that i'm really trying to say is is does the brain have a clock cycle 00:17:50.749 --> 00:17:55.769 because as you say some people will will will argue definitely not and people 00:17:55.769 --> 00:17:59.329 will talk about if they want to use a computer computer analogy they will talk 00:17:59.329 --> 00:18:04.229 about event-based codes so it's a continuous time system and then you get spikes 00:18:04.229 --> 00:18:10.529 and we do all the processing on those spikes and it's actually, 00:18:10.669 --> 00:18:14.589 you're losing information when you say, let's have time steps. 00:18:15.049 --> 00:18:17.649 Well, let me tell you one more piece of information and get your opinion. 00:18:18.249 --> 00:18:22.009 So there's been a big debate about the anti. 00:18:24.189 --> 00:18:27.189 There are a group of people who are sort of anti-oscillations. 00:18:28.229 --> 00:18:31.689 And one of the things that they've pointed out, which is absolutely true, 00:18:32.789 --> 00:18:37.389 is that the periods of oscillations are not clock-like. 00:18:37.389 --> 00:18:42.169 The periods, if you look at, let's say, even during a theta cycle, 00:18:43.089 --> 00:18:46.769 and you look from one gamma cycle to the next and ask the question, 00:18:46.949 --> 00:18:50.769 you know, how regular is the gamma period? 00:18:51.729 --> 00:18:54.229 And the answer is, not very regular. 00:18:55.469 --> 00:19:03.849 And it's not only empirically true, but given current models about how gamma is generated, 00:19:03.849 --> 00:19:10.189 this sort of ping model that's just a negative feedback between the principal 00:19:10.189 --> 00:19:11.729 cells, inhibitory cells, 00:19:12.009 --> 00:19:17.869 then feeding back onto the principal cells, those models also do not predict 00:19:17.869 --> 00:19:24.129 that this is going to be a regular oscillation. So. 00:19:26.219 --> 00:19:30.159 If you use the word clock, you'd have to say, oh my God, this is a terrible 00:19:30.159 --> 00:19:35.459 clock because gamma period is highly fluctuating. 00:19:35.579 --> 00:19:40.639 So a downstream, one kind of operation that maybe in a computer you could do, 00:19:40.799 --> 00:19:45.279 but in the brain, according to the theta gamma code, you can't do, 00:19:45.279 --> 00:19:52.979 do is to say, given that I have seen the output of cells during one gamma cycle, 00:19:53.199 --> 00:19:59.619 I can predict that exactly 25 milliseconds later, I'll get another big bang, right? 00:19:59.699 --> 00:20:01.959 That is not true. You cannot do that. 00:20:03.079 --> 00:20:06.579 But maybe there's also another aspect to this that makes 00:20:06.579 --> 00:20:10.159 a computer metaphor maybe less helpful because 00:20:10.159 --> 00:20:13.059 in a computer it you 00:20:13.059 --> 00:20:15.939 know it largely operates also on 00:20:15.939 --> 00:20:22.499 a segregation of memory and program or memory and processing right these are 00:20:22.499 --> 00:20:26.919 strongly separated and to keep the content and the operation synchronized you 00:20:26.919 --> 00:20:30.419 have to clock the whole thing well but if you look at this hippocampal process 00:20:30.419 --> 00:20:35.499 it's not so obvious that let's say the memory and the process are actually differentiated. 00:20:35.679 --> 00:20:40.299 And maybe the problem you will have much more if you are a hippocampus or a 00:20:40.299 --> 00:20:45.939 brain is to impose some sort of order on these continuous streams of events 00:20:45.939 --> 00:20:47.699 that you're bombarded with. 00:20:47.839 --> 00:20:50.839 So maybe more this event-based interpretation that also Tony mentioned earlier. 00:20:50.979 --> 00:20:58.019 So another way maybe to think about the teta-gamma cycle is that first you try 00:20:58.019 --> 00:21:04.019 to get to some sort of a temporal segmentation of these continuous input streams through Teta, 00:21:04.139 --> 00:21:08.699 but you rely very much on these local competitive processes that give rise to 00:21:08.699 --> 00:21:11.619 Gamma, so the excitatory-inhibitory interactions that are fairly local, 00:21:11.819 --> 00:21:17.799 to then fill in within those short segments that Teta gives you what the most, 00:21:17.879 --> 00:21:21.499 let's say, dominant features are in that local process. 00:21:21.839 --> 00:21:24.899 Because the other thing we shouldn't forget about, it's not that Teta Gamma 00:21:24.899 --> 00:21:29.959 as we now see it in the hippocampus is all there is to it, because actually 00:21:29.959 --> 00:21:35.299 gamma is something that is playing out in lots of local volumes in a completely 00:21:35.299 --> 00:21:36.679 desynchronized fashion, 00:21:36.839 --> 00:21:40.739 because it is just the range of an interneuron, you could say, 00:21:40.759 --> 00:21:44.979 that dictates what the sizes of a population of excitatory cells that play a 00:21:44.979 --> 00:21:46.879 role in generating gamma. 00:21:47.039 --> 00:21:50.339 So if you go, let's say, a few millimeters further down, you have a very different 00:21:50.339 --> 00:21:52.659 kind of, let's say, gamma encoding taking place. 00:21:52.679 --> 00:21:57.099 It has nothing to do anymore with the first one, but still they're all being 00:21:57.099 --> 00:22:01.679 then, if you want, aligned in time through the slow cycle of theta. 00:22:02.339 --> 00:22:06.599 And then if you talk about the clock, maybe the clock is much more at the level 00:22:06.599 --> 00:22:08.599 of the theta cycle than gamma. 00:22:08.659 --> 00:22:11.239 Gamma is much more asynchronous and local, but the real clock, 00:22:11.319 --> 00:22:15.959 if you want to talk about the clock, is then the system that dominates the theta cycle. 00:22:16.099 --> 00:22:18.999 So how regular is theta compared to gamma? 00:22:21.430 --> 00:22:27.330 I'm not really sure, but my guess would be that theta is maybe even more irregular. 00:22:27.410 --> 00:22:35.890 In fact, you know, we don't really even know how to distinguish theta in the cortex from alpha. 00:22:36.110 --> 00:22:42.550 And this is all very confusing right now. In other words, where do these names come from? 00:22:42.550 --> 00:22:45.710 And the answer is they they 00:22:45.710 --> 00:22:49.150 came originally from the 00:22:49.150 --> 00:22:52.050 eeg field and they had 00:22:52.050 --> 00:22:55.570 zero intellectual basis that is 00:22:55.570 --> 00:23:01.230 it was just arbitrary they said okay let's just define this range by one greek 00:23:01.230 --> 00:23:07.490 letter and this range by another greek letter and so you know we've just all 00:23:07.490 --> 00:23:11.850 been very confused by you whether to make a big deal out of the difference between 00:23:11.850 --> 00:23:14.610 8 Hz in the hippocampus, 7 Hz in the hippocampus, 00:23:15.350 --> 00:23:20.350 and maybe that's the same kind of process as 10 Hz in cortex. I mean, who's to say? 00:23:21.530 --> 00:23:29.330 But we have maybe a handle on that, right? Because we know the origins of these oscillations, right? 00:23:29.390 --> 00:23:35.970 So gamma is seen to arise out of a local circuit where excitation inhibition interacts, right? 00:23:36.030 --> 00:23:38.710 I think there's consensus on that, whether it's cortex or hippocampus. 00:23:38.710 --> 00:23:40.650 Fair enough. This is seen as a source of gamma. 00:23:40.830 --> 00:23:45.050 Okay. That means then that rhythmicity and regularity will depend on the local 00:23:45.050 --> 00:23:46.750 properties of that circuit. 00:23:47.050 --> 00:23:51.050 Right. Theta and hippocampus is seen as arising from the septum. 00:23:51.510 --> 00:23:55.530 Okay. It's really as a driver of a slow, of this slow oscillatory response. 00:23:55.710 --> 00:24:00.230 For the cortex, it might be the thalamus that dictates, or it will be the thalamus 00:24:00.230 --> 00:24:02.270 that dictates the slow rhythm. 00:24:02.830 --> 00:24:07.070 Right. So, if you now look at the sources of theta and gamma. 00:24:07.070 --> 00:24:12.530 Well, as you said, I think you were on the right track when you said maybe what's 00:24:12.530 --> 00:24:18.710 really important in terms of interactions between brain regions is the slow oscillation. 00:24:19.110 --> 00:24:23.390 And that does maybe have to be synchronized between areas that communicate. 00:24:24.170 --> 00:24:27.170 Maybe the gamma's not so important. It's all local. 00:24:27.530 --> 00:24:34.270 I think that the notion of a clock, although in a computer it's going to be a metronome. 00:24:34.290 --> 00:24:40.010 It doesn't have to be, because the key word for me that John said was discrete. 00:24:40.270 --> 00:24:45.510 So you have a cycle, and it may be varying in its timing, but as long as the 00:24:45.510 --> 00:24:51.450 different parts of the system that need to understand each other are on the same point in the cycle, 00:24:52.030 --> 00:24:56.450 then the order that you're imposing by saying, I'm only going to. 00:24:57.746 --> 00:25:00.846 Treat things within a gamma cycle as being together 00:25:00.846 --> 00:25:03.646 and i'm going to take these bunch of events that happen with the 00:25:03.646 --> 00:25:06.366 gamma cycle and treat them as one event in some 00:25:06.366 --> 00:25:09.346 way and process that so that seems 00:25:09.346 --> 00:25:16.846 to me the strong claim you're making that the the brain buys some uh some usefulness 00:25:16.846 --> 00:25:22.246 from this binding of everything together into discrete chunks and that this 00:25:22.246 --> 00:25:27.086 is maybe makes more sense than just having all these continuous processes. 00:25:27.326 --> 00:25:31.546 Because if I'm a bit of the brain over here, listening to this bit of the brain 00:25:31.546 --> 00:25:35.906 over here, I'll have to listen continuously and integrate everything that's 00:25:35.906 --> 00:25:39.066 happening in order to know what's going on. 00:25:39.186 --> 00:25:43.086 But you're saying, I get a series of snapshots with the gamma cycle, 00:25:43.286 --> 00:25:47.786 or maybe with the theta cycle, and I can just pay attention to what's in that snapshot. 00:25:48.606 --> 00:25:52.066 I don't like that metaphor either. 00:25:52.066 --> 00:25:54.886 That no i i like it but i want to i want to 00:25:54.886 --> 00:25:58.346 i want to propose that we use another word okay instead 00:25:58.346 --> 00:26:01.866 of snapshot movie right and why 00:26:01.866 --> 00:26:04.526 do i like that because let's say 00:26:04.526 --> 00:26:09.466 in the experiments that we've been discussing today in various forms the johnson 00:26:09.466 --> 00:26:16.566 reddish experiment where the animal in a sense gets a view of what happened 00:26:16.566 --> 00:26:22.766 if it goes down one path uh of the maze so what do I mean by view? 00:26:23.006 --> 00:26:32.226 Well, what the data show is that each point along the path is represented, and that is a movie. 00:26:32.686 --> 00:26:38.406 That's not a snapshot. It's a sequence of snapshots. It's a sequence of snapshots, which is a movie. 00:26:39.986 --> 00:26:46.426 And I mean, that's really impressive in terms of decision -making, 00:26:46.566 --> 00:26:50.686 because you're getting that movie, you know, within less than 100 milliseconds, 00:26:51.366 --> 00:26:53.906 and that's very valuable information. 00:26:54.406 --> 00:26:58.446 And I was discussing, Paul, I was discussing before with Tony one thing I really, 00:26:58.506 --> 00:27:04.786 really like about now imagining how the basal ganglia evaluates this movie. 00:27:05.956 --> 00:27:12.756 Um the movie can be rich the movie can say you know i i turned last time i turned left, 00:27:13.516 --> 00:27:20.716 and uh you know the first thing i came to was you know some breadcrumbs and that was okay, 00:27:21.296 --> 00:27:27.856 and then i went on and uh and i smelled some cat urine and i thought for sure i was going to get eaten, 00:27:27.996 --> 00:27:36.696 but finally I got to the reward site and there was some really yummy milk. 00:27:37.036 --> 00:27:41.756 Okay, I mean, so you have costs and benefits about the left choice. 00:27:42.136 --> 00:27:49.896 Well, all of that information is passed within less than 100 milliseconds to the basal ganglia. 00:27:50.496 --> 00:27:58.616 And I think that current current models of the basal ganglia would allow you to actually integrate, 00:27:58.936 --> 00:28:02.936 to evaluate each of those steps within, let's say, one gamma cycle, 00:28:03.136 --> 00:28:07.936 get the value or the costs and the benefits of. 00:28:09.305 --> 00:28:15.585 And wind up at the end with some sort of integrated number which told you the 00:28:15.585 --> 00:28:17.285 sum of the costs and the benefits. 00:28:18.585 --> 00:28:23.845 And that's wonderful. I mean, and we know that that's the way we want to run 00:28:23.845 --> 00:28:28.265 a railroad is not just to look at what's at the end of the path, 00:28:28.905 --> 00:28:35.565 but to know that, you know, what we're going to risk if we take that path along the way. 00:28:35.905 --> 00:28:41.725 It was sort of the point there is that Another element I think that you have 00:28:41.725 --> 00:28:45.485 not emphasized yet, even though you're dealing with it, 00:28:45.905 --> 00:28:49.385 what's maybe more important is that also if you are a brain, 00:28:49.525 --> 00:28:55.585 which we all are in some way, you have to progressively throw away more information 00:28:55.585 --> 00:28:58.025 and distill things down to what really matters. 00:28:58.025 --> 00:29:01.425 And also the way you think about 00:29:01.425 --> 00:29:09.545 the gamma cycle, it's also a progressive deletion of noise, if you want. 00:29:09.625 --> 00:29:14.205 Because also what you're saying, in the gamma cycle, I can generate, 00:29:14.285 --> 00:29:19.785 what, seven plus or minus responses within a theta cycle. 00:29:20.765 --> 00:29:24.745 But these are all the result of a local winner-take-all. So what I'm going to 00:29:24.745 --> 00:29:29.885 report in this tata cycle is the stuff that really stands out for me among possibly 00:29:29.885 --> 00:29:31.805 hundreds of possible responses. 00:29:32.685 --> 00:29:36.565 So if I'm in this maze, like we take the Johnson and Reddish experiment, 00:29:36.785 --> 00:29:39.785 I'm at the T crossing in the maze. 00:29:41.645 --> 00:29:46.965 I imagine I'm going to the left. So now I have a sweep through my hippocampal 00:29:46.965 --> 00:29:48.585 representation, right? 00:29:48.585 --> 00:29:57.305 Within this gamma cycle, I have 100 milliseconds, and now what pops out are 00:29:57.305 --> 00:30:01.165 actually the most relevant spots I might visit there. It's not all possible spots. 00:30:01.385 --> 00:30:05.305 So it's all this incremental selection that might be an important role there, 00:30:05.325 --> 00:30:09.185 and not only how the information is processed further downstream. Right. 00:30:10.296 --> 00:30:13.536 Okay, well, now we get into a really complicated area of episodic memory. 00:30:13.896 --> 00:30:17.996 So, I mean, strictly speaking, you know, what you would recall, 00:30:18.216 --> 00:30:25.716 if we really think in terms of episodic memory, is one of the times that I went down that path. 00:30:25.716 --> 00:30:34.776 If you take the point of view that the hippocampus is really the episodic store 00:30:34.776 --> 00:30:41.696 and that what Johnson and Reddish is seeing is a recall of one previous traversal, 00:30:41.696 --> 00:30:46.016 not the statistical properties of all previous traversals of that path, 00:30:46.116 --> 00:30:50.736 but one of them, then we get into this whole area of, 00:30:51.176 --> 00:30:54.596 you know, is this the way you want to railroad? road? 00:30:54.656 --> 00:31:01.196 Do you want to, you know, do you want to use your hippocampus to remember individual events? 00:31:01.336 --> 00:31:04.396 And then if so, which ones, which exact ones do you recall? 00:31:04.596 --> 00:31:08.736 Or do you want your hippocampus to be statistical? 00:31:09.316 --> 00:31:15.596 I take the point of view that, you know, the great thing about the hippocampus 00:31:15.596 --> 00:31:21.936 is that it recalls an individual, at least it hopes to recall an individual trial. 00:31:22.056 --> 00:31:27.156 And that has pluses and minuses. It says, you know, that time on July 4th when 00:31:27.156 --> 00:31:31.576 I went down this path and, you know, I thought there was going to be a cat, 00:31:31.836 --> 00:31:37.516 but, you know, there was actually a buzzing going on and there wasn't a cat. 00:31:37.676 --> 00:31:41.376 So this time, if there's not a buzzing, if there's a buzzing, 00:31:41.456 --> 00:31:44.056 I'll take the risk because maybe there's not a correlation. 00:31:44.256 --> 00:31:48.576 I mean, all that detail could be important in making your next choice, 00:31:48.736 --> 00:31:51.496 but that's not statistical. 00:31:51.896 --> 00:31:55.776 And I mean, I love the idea that people have shown that actually. 00:31:57.945 --> 00:32:04.185 You know, if you use your episodic memory, you will often make worse choices. 00:32:05.625 --> 00:32:10.405 So my favorite example, and I've heard that these experiments are a little controversial, 00:32:13.565 --> 00:32:19.885 they actually, they did the following. They showed people two houses, 00:32:19.985 --> 00:32:25.425 and they went through each house. And this was done probably on a computer. 00:32:25.565 --> 00:32:28.665 And they said, oh, here's the kitchen. It's a wonderful, modernized kitchen, 00:32:28.745 --> 00:32:33.265 but here's the bath. It's not really ever been modernized. And here's the bedroom. 00:32:33.445 --> 00:32:36.545 You know, it's really large, but here's the living room. It has a great view. 00:32:36.785 --> 00:32:40.765 And the costs and benefits of the two houses were very different. 00:32:41.465 --> 00:32:46.105 And in fact, they had arranged it so that they thought that most people would 00:32:46.105 --> 00:32:48.565 pick house one as the better house. 00:32:48.825 --> 00:32:54.525 Okay, now they had two groups. After showing them these two houses, 00:32:54.645 --> 00:32:58.005 one group was told, please do this arithmetic problem. 00:32:58.925 --> 00:33:06.085 And the other group was told, you know, think about this house. 00:33:07.405 --> 00:33:10.605 Then they asked them, which house is better? 00:33:11.865 --> 00:33:16.225 And, you know, one can argue about whether it's possible to decide objectively 00:33:16.225 --> 00:33:22.185 that one house is better. But anyway, the people who did not think about it were better. 00:33:22.765 --> 00:33:26.765 Now, why would this be? Well, there's actually a very interesting and, 00:33:26.805 --> 00:33:27.945 I think, useful explanation. 00:33:28.845 --> 00:33:36.505 Probably what happens when you're using your thinking is that you get stuck in silly examples. 00:33:36.605 --> 00:33:41.445 You say, you know, that kitchen reminded me of my least favorite aunt. 00:33:42.025 --> 00:33:46.825 And, you know, she, damn it, you know, I went to her house and she gave me this 00:33:46.825 --> 00:33:49.225 porridge that was just disgusting. 00:33:49.845 --> 00:33:52.845 And you're spending all your time thinking about this aunt, so you actually 00:33:52.845 --> 00:33:54.505 aren't thinking about all the other rooms. 00:33:56.252 --> 00:33:59.552 And so it turns out that your basal ganglia which is sort of more statistical 00:33:59.552 --> 00:34:04.792 actually winds up doing a good job because even you know your episodic memory 00:34:04.792 --> 00:34:08.552 isn't sampling very well anyway. 00:34:10.632 --> 00:34:15.952 Look i i think i think there's a problem here uh i'm sorry to to bring the bad 00:34:15.952 --> 00:34:20.672 news but earlier you said and you referred to a paper we published together 00:34:20.672 --> 00:34:24.872 that that memory is is based on attractor dynamics. 00:34:25.832 --> 00:34:29.212 And by virtue of that, you can do pattern completion, for instance. 00:34:30.432 --> 00:34:36.352 But an attractor dynamic has, by necessity, statistical properties because it 00:34:36.352 --> 00:34:39.952 will pull input states towards some average state. 00:34:40.732 --> 00:34:45.672 So in that sense, to then declare that attractor state as a factual, 00:34:45.712 --> 00:34:51.972 accurate representation of a single sampling, I think would be a really long shot. 00:34:52.132 --> 00:34:55.152 How are you going to defend your attractor from drifting and that's one of the 00:34:55.152 --> 00:35:01.952 things we showed we explained we explained also in that paper that the drift that people saw, 00:35:02.492 --> 00:35:06.832 in the kind of say a three memory when you change the environment so now we 00:35:06.832 --> 00:35:10.192 are actually visiting the houses we were slowly changing them that there is 00:35:10.192 --> 00:35:14.812 a drift in that memory and that drift is also an expression of a statistical 00:35:14.812 --> 00:35:18.572 property that you are averaging over experience right so. 00:35:20.432 --> 00:35:25.712 Isn't the idea of an attractor dynamic based memory actually forcing you to 00:35:25.712 --> 00:35:31.692 also commit to a more statistical interpretation of episodic memory well let 00:35:31.692 --> 00:35:39.872 me try to see if i can argue the other way um i mean you know. 00:35:41.792 --> 00:35:45.892 Why is the attractor relevant when you you know you think about the kitchen 00:35:45.892 --> 00:35:51.812 in that house you You were shown, and it was because it was really very similar 00:35:51.812 --> 00:35:54.192 to your aunt's kitchen. So there's the attractor. 00:35:54.712 --> 00:35:56.812 So we've utilized the attractor. 00:35:58.752 --> 00:36:05.192 But if it was just generalized kitchen, right? 00:36:05.252 --> 00:36:09.892 You're saying it's really a kitchen attractor. It's not my aunt's kitchen attractor. 00:36:10.772 --> 00:36:16.492 Then it would bring up associations with all kitchens, which might be very positive on average. 00:36:16.712 --> 00:36:20.912 But in fact what happened was that it. 00:36:22.449 --> 00:36:28.129 Was not generalizable that kitchen just got you stuck in your aunt's kitchen 00:36:28.129 --> 00:36:33.849 i think we're possibly looking at a false dichotomy here because you what you 00:36:33.849 --> 00:36:34.989 want for your episodic memory, 00:36:35.629 --> 00:36:42.689 is to do pattern completion so you want to be able to fill out what's happening 00:36:42.689 --> 00:36:47.009 now based on things which have happened in the past which are relevant which 00:36:47.009 --> 00:36:49.049 may help you interpret that situation. 00:36:49.369 --> 00:36:53.309 So you're in that kitchen, you're reminded of your aunt's kitchen. 00:36:53.589 --> 00:36:57.009 And that may also make you think that, well, maybe this kitchen has got some 00:36:57.009 --> 00:37:00.809 of the properties of that other kitchen I was in, and that might be useful for 00:37:00.809 --> 00:37:01.669 understanding that scene. 00:37:02.809 --> 00:37:08.629 So the pattern completion is going to partly depend on many possible past events 00:37:08.629 --> 00:37:11.109 which are relevant to interpreting this one. 00:37:12.069 --> 00:37:15.369 And then the other thing you want to do is pattern separation. 00:37:15.689 --> 00:37:18.569 So you want to, and this is you want 00:37:18.569 --> 00:37:22.249 to be able to say but this isn't my aunt's kitchen so although i 00:37:22.249 --> 00:37:25.069 remember it as being similar it's not 00:37:25.069 --> 00:37:28.429 the same thing so your episodic memory has to be able to distinguish 00:37:28.429 --> 00:37:31.229 this kitchen from previous ones which is what you're 00:37:31.229 --> 00:37:34.109 you're saying is that we needed to do both these things 00:37:34.109 --> 00:37:36.929 and the fact that i think it can do 00:37:36.929 --> 00:37:39.789 both but it doesn't do either of them perfectly is the 00:37:39.789 --> 00:37:43.769 thing we want to try and explain about episodic memory that sounds 00:37:43.769 --> 00:37:46.809 like a very nice compromise i'll go i'll go 00:37:46.809 --> 00:37:50.149 for that good at 00:37:50.149 --> 00:37:53.249 least john will not demolish the studio now that's great 00:37:53.249 --> 00:37:58.529 but so now so so now we have an idea about this episodic memory in the hippocampus 00:37:58.529 --> 00:38:03.369 and we want to use that as a way to understand coding in the brain right and 00:38:03.369 --> 00:38:07.489 what we see there is that this episodic memory that's playing out on on the 00:38:07.489 --> 00:38:11.909 slow oscillation of theta in the gamma cycle is um. 00:38:13.049 --> 00:38:18.669 Is telling us, let's say, what's the most relevant aspects of the processing 00:38:18.669 --> 00:38:22.049 going on within this piece of volume of the hippocampus. 00:38:22.229 --> 00:38:26.649 But now if you go back to the data of Foster that you talked about earlier, 00:38:27.169 --> 00:38:33.649 where they have, with very great temporal precision, unpacked the response in the gamma cycle, 00:38:34.969 --> 00:38:38.609 in some sense it showed, again, a further complexification classification because 00:38:38.609 --> 00:38:43.329 it's not that on every single gamma cycle there's just a single spike. 00:38:43.709 --> 00:38:47.889 You have multiple spikes riding within the gamma cycle. 00:38:48.089 --> 00:38:52.389 So can you imagine that the nesting would go further than only theta gamma, 00:38:52.609 --> 00:38:55.289 but it also go, let's say, low and high gamma? 00:38:58.810 --> 00:39:03.610 Well, first of all, you know, we don't really understand anything about the 00:39:03.610 --> 00:39:05.970 differentiations between low and high gamma. 00:39:06.110 --> 00:39:12.350 And there's one really, you know, accepted place where you see these simultaneously, 00:39:12.410 --> 00:39:14.890 or at least overlapping, 00:39:15.150 --> 00:39:21.930 and that's in the high gamma that comes into CA1 from the cortex and the low 00:39:21.930 --> 00:39:27.250 gamma that comes into CA1 from CA3. 00:39:28.270 --> 00:39:35.590 And it's really just unclear to me what this all means. I don't think anybody knows. 00:39:35.730 --> 00:39:41.930 One surprising thing is you might say, let's look at the output and see who 00:39:41.930 --> 00:39:43.750 the output is listening to? 00:39:44.650 --> 00:39:50.470 And the answer is neither. The output just has its own gamma, 00:39:50.670 --> 00:39:52.650 not tied to either of them. 00:39:53.870 --> 00:40:00.830 So, I mean, one very simple way of looking at it, which I favor, you know, is that, 00:40:01.730 --> 00:40:06.250 you know, inputs are coming in and the cell is integrating them on its own time 00:40:06.250 --> 00:40:11.010 constant and deciding how to output the information on its own. 00:40:12.730 --> 00:40:18.810 That's not tremendously satisfactory, but that seems to be the best description, 00:40:19.610 --> 00:40:22.530 that I can come up with of what has actually been found. 00:40:22.890 --> 00:40:26.390 But that seems reasonable, right? Because let's say I'm in CA3, 00:40:26.910 --> 00:40:31.530 I'm packing up information in my gamma cycle based on a local competitive process. 00:40:31.670 --> 00:40:37.470 Now, what I'm telling CA1 about, the next station upstream or downstream from me. 00:40:37.790 --> 00:40:41.750 Well, that would be the subiculum. The downstream from CA1 is the subiculum. 00:40:41.890 --> 00:40:45.310 I was just talking about CA3, right? No, I was talking about CA1. 00:40:45.510 --> 00:40:48.710 Okay, right. So downstream of CA1. No, but I want to start at CA3 because I 00:40:48.710 --> 00:40:52.650 want to say, basically CA3 makes a preselection and said, okay, 00:40:52.730 --> 00:40:54.490 these are the few items that I care about. 00:40:55.330 --> 00:40:57.150 So it has filled out a lot of noise. 00:40:58.030 --> 00:41:02.050 Now CA1 will do something similar because it now gets this whole barrage of 00:41:02.050 --> 00:41:04.250 inputs from CA3 still and it has to make a selection. 00:41:04.430 --> 00:41:09.090 So again, it selects within its own gamma cycle using the same competitive process 00:41:09.090 --> 00:41:10.950 to tell the subiculum what it really cares about. 00:41:11.790 --> 00:41:15.850 So it's really also a hierarchy then of competitive processes. 00:41:16.590 --> 00:41:20.950 So this is how you think about it. I don't know. I think this is just going to... 00:41:21.470 --> 00:41:26.810 I personally would think that, you know, a lot of effort should now be put into 00:41:26.810 --> 00:41:31.450 understanding CA1 because it's just so fascinating and what's happening and 00:41:31.450 --> 00:41:35.010 we have so much data and something, some principle is going to emerge, 00:41:35.230 --> 00:41:40.430 but I'm not, I honestly have no idea what it is, but I actually think I'd like to spend some time. 00:41:41.310 --> 00:41:42.690 Seeing if I can figure it out. 00:41:43.893 --> 00:41:51.373 So then early in the week, we also had quite some information presented to us 00:41:51.373 --> 00:41:52.713 by Alfred Moser on the grid cells. 00:41:52.953 --> 00:41:58.173 And the grid cell story actually for the last 10 years looks so clean and nice 00:41:58.173 --> 00:42:02.073 because here we had these cells in the entorhinal cortex input to the hippocampus, 00:42:02.693 --> 00:42:06.773 have a very nice grid-like response to space driven by velocity. 00:42:06.773 --> 00:42:09.673 Velocity and this might then 00:42:09.673 --> 00:42:12.493 be a great substrate for place cells and hippocampus to 00:42:12.493 --> 00:42:15.393 learn to learn about space right now that 00:42:15.393 --> 00:42:18.793 story has become more complicated because it looks like place cells 00:42:18.793 --> 00:42:22.913 can develop place fields in hippocampus even if you have no grid cells available 00:42:22.913 --> 00:42:27.673 to you so that the place cells are way more promiscuous in that sense than initially 00:42:27.673 --> 00:42:34.093 anticipated so that might give us some space to also speculate about what these 00:42:34.093 --> 00:42:37.653 grid cells really are for if they're not really, 00:42:38.393 --> 00:42:40.753 the only source of spatial information. 00:42:41.493 --> 00:42:44.453 So it was up to you. How else would you abuse the grid cells? 00:42:45.773 --> 00:42:50.373 Well, we very much favor a new view of grid cells. 00:42:51.073 --> 00:42:56.273 And the question is, you know, what functions are there out there that you have 00:42:56.273 --> 00:43:00.353 to account for and which ones do you want to assign to grid cells. 00:43:00.453 --> 00:43:05.013 So classically, what people have said is, well, you have sensory information 00:43:05.013 --> 00:43:07.293 about the outside world, you know, here I am. 00:43:08.073 --> 00:43:13.353 People have said, path integration, I integrate velocity, so if I knew where 00:43:13.353 --> 00:43:16.293 I was, I integrate my velocity, I know where I am now. 00:43:18.913 --> 00:43:23.833 And then this interesting phenomenon, which we've talked about earlier, 00:43:23.833 --> 00:43:27.313 on the basis of Johnson and Reddish, which is I'm not moving, 00:43:28.033 --> 00:43:33.773 but my mind is going to move down one of the paths. And we call that mind travel. 00:43:35.173 --> 00:43:39.013 So in the end, we have these three functions. 00:43:39.513 --> 00:43:43.913 How are we going to build a system that does all three? That's what we've got to do. 00:43:45.053 --> 00:43:53.273 And by my way of thinking, the grid cells are really well suited for doing this mind travel. 00:43:53.973 --> 00:43:58.033 So they move you in an imaginary way through space. 00:43:58.933 --> 00:44:03.973 They can integrate, but what they integrate is not your actual velocity, 00:44:04.053 --> 00:44:10.213 but some artificial velocity which part of the brain generates saying, 00:44:10.393 --> 00:44:15.193 I'm curious about what happens if I move down in this direction with this velocity. 00:44:15.973 --> 00:44:19.333 And then the grid cell system can give you the answer. 00:44:20.593 --> 00:44:27.013 It's a coordinate system that says, well, given this velocity in this direction, 00:44:27.093 --> 00:44:32.273 this is where you will be, and it gives you that path. 00:44:32.893 --> 00:44:39.353 The beauty then is that it can force that movement through space on the hippocampus. 00:44:40.833 --> 00:44:43.753 What can the hippocampus contribute to that? 00:44:44.633 --> 00:44:50.413 It can contribute associations that happened with those places. 00:44:50.793 --> 00:44:56.713 So if there was reward or a cat associated with some place that you mind travel 00:44:56.713 --> 00:44:59.993 to, that's incredibly valuable information. 00:44:59.993 --> 00:45:06.153 So I don't think that the grid cells know about cats or food, 00:45:06.313 --> 00:45:09.673 but they do know about the coordinate systems of space. 00:45:09.673 --> 00:45:17.413 So you move through space and you tell the hippocampus, here's how we're moving through space. 00:45:17.613 --> 00:45:23.013 And then the hippocampus says, given that you've made me move through space, 00:45:23.213 --> 00:45:27.553 this is what I found associated with that. 00:45:27.553 --> 00:45:31.673 That so then we 00:45:31.673 --> 00:45:34.873 would say well great if if if that's 00:45:34.873 --> 00:45:38.393 what grid cells are good for don't we 00:45:38.393 --> 00:45:41.733 also have to do path integration well we 00:45:41.733 --> 00:45:44.553 there we can make some suggestions that it 00:45:44.553 --> 00:45:47.333 might occur earlier in the structure and there's 00:45:47.333 --> 00:45:50.493 some interesting candidates about where path integration might occur 00:45:50.493 --> 00:45:53.753 so that's the integration of real velocity so um 00:45:53.753 --> 00:45:56.693 i have a bit of a worry about this story 00:45:56.693 --> 00:45:59.813 because you know edward moser presented this 00:45:59.813 --> 00:46:03.193 beautiful data set and his interpretation 00:46:03.193 --> 00:46:06.253 or my interpretation of his interpretation is that 00:46:06.253 --> 00:46:10.333 we have these modules at different levels of 00:46:10.333 --> 00:46:13.673 granularity which are giving us this wonderful 00:46:13.673 --> 00:46:21.013 metric map of space so that when you come into this room that metric is laid 00:46:21.013 --> 00:46:27.313 out over the space so that every point in space now has a unique code assigned 00:46:27.313 --> 00:46:30.893 to it across these different modules of grid cells. 00:46:31.753 --> 00:46:36.693 And it's anchored possibly at the door or some other point. All the grid cells 00:46:36.693 --> 00:46:39.633 always seem to have an anchor at the place that that is introduced. 00:46:40.913 --> 00:46:43.333 Now I've got this metric map of space. 00:46:45.775 --> 00:46:52.995 I can navigate in it. But your suggestion seems to be that you're going to take 00:46:52.995 --> 00:46:55.455 my map of space and you're now going to start moving it around. 00:46:55.675 --> 00:47:01.515 So I'm no longer on firm ground knowing what my coordinate frame is because 00:47:01.515 --> 00:47:03.455 my coordinate frame is shifting. 00:47:04.115 --> 00:47:07.475 You're moving it around. Or have I misinterpreted what you're saying? 00:47:08.855 --> 00:47:14.015 Maybe misinterpreted because when you described Edward Moser's proposal, 00:47:14.015 --> 00:47:18.275 I found myself nodding, not nodding to sleep, but nodding in agreement. 00:47:19.055 --> 00:47:26.195 And so the answer is, I think that I would agree with everything you said. 00:47:26.235 --> 00:47:30.215 It's a coordinate system which allows navigation. 00:47:31.295 --> 00:47:35.635 The question is, you know, what do you use that coordinate system for? 00:47:35.795 --> 00:47:40.615 And the nice thing about it, it's a pre-wired coordinate system. 00:47:40.855 --> 00:47:43.535 So you can use it in any environment. you just have to 00:47:43.535 --> 00:47:46.955 attach it yeah to a given environment and 00:47:46.955 --> 00:47:49.835 now you have the means for saying well if 00:47:49.835 --> 00:47:52.575 i want look this way you know where will 00:47:52.575 --> 00:47:55.895 i wind up or if i go this way where will i wind up which 00:47:55.895 --> 00:47:58.675 is what i'm saying right so i don't think we're i 00:47:58.675 --> 00:48:04.295 don't think we're at odds uh the only um no 00:48:04.295 --> 00:48:07.195 i i guess i don't see them i don't see them as yeah as country 00:48:07.195 --> 00:48:10.215 okay i guess you're you're putting the ability to 00:48:10.215 --> 00:48:13.175 move within the map in the grid cells which are themselves the 00:48:13.175 --> 00:48:17.135 coordinate frame map and i worry a little bit whether you can do that whether 00:48:17.135 --> 00:48:21.235 the the grid cells just have to form the coordinate frame and then some other 00:48:21.235 --> 00:48:25.855 system can represent movement within it okay so i think you did say one thing 00:48:25.855 --> 00:48:32.335 that really i'm forced to to come back to and admit that's a a problem and suggest a solution. 00:48:33.115 --> 00:48:37.695 And that is, if we use the grid cell system for. 00:48:40.795 --> 00:48:47.455 These imaginary movements, then how do you reset it at the end of a theta cycle 00:48:47.455 --> 00:48:49.695 so you're back to current position? 00:48:51.455 --> 00:48:56.915 Because we've sort of mucked with this. Right? And my, 00:48:58.075 --> 00:49:05.755 proposal is that that we do have another integrator which is integrating true 00:49:05.755 --> 00:49:12.315 velocity and which is, in a sense, doing what everybody thought the grid cell system would do. 00:49:13.935 --> 00:49:18.815 So, you know, you could say, well, I've made life more complicated now because 00:49:18.815 --> 00:49:23.135 now I have a whole other system which is keeping track of actual position, 00:49:23.375 --> 00:49:27.455 another grid-like system that's keeping track of actual position, 00:49:27.655 --> 00:49:36.335 and that can, since it knows where I am, can reset the grid cell system to current 00:49:36.335 --> 00:49:41.015 position after I muck with it and make it do this imaginary movement. 00:49:41.835 --> 00:49:45.855 You were absolutely right. You have to, after you do the imaginary movement, 00:49:46.015 --> 00:49:47.655 you have to move back to current position. 00:49:47.955 --> 00:49:50.075 So, somebody must know current position. 00:49:51.095 --> 00:49:58.375 But, I mean, it's not, I mean, this complexity is demanded by the fact that 00:49:58.375 --> 00:50:02.395 we see these imagined movements through space. 00:50:02.595 --> 00:50:04.875 So, somebody's got to be doing this. 00:50:05.075 --> 00:50:10.595 So, if you take another model, they also have to solve of these dual requirements. 00:50:11.755 --> 00:50:16.835 I think both of you are much too pessimistic about this because on the one hand. 00:50:18.415 --> 00:50:21.155 Why would I even need a parallel system for that? 00:50:21.335 --> 00:50:24.795 Because I just need to replace my velocity vector, right? 00:50:25.055 --> 00:50:31.315 Either the velocity vector comes from my vestibular system or optic flow, 00:50:31.495 --> 00:50:33.495 whatever, that tells about physical movement in space. 00:50:33.755 --> 00:50:36.235 And with that, I'm driving my grid cell security response. 00:50:36.755 --> 00:50:40.515 This signal is arriving driving in the entorhinal cortex over the thalamus. 00:50:40.575 --> 00:50:45.395 So I have a number of stages in this pathway where I can basically hijack the 00:50:45.395 --> 00:50:47.875 signal. I can pump in anything else I want. 00:50:48.095 --> 00:50:50.135 I drive my grid cells around again. 00:50:50.535 --> 00:50:54.615 Now you have to, and the point is, so far the data, at least on rats, 00:50:54.715 --> 00:51:00.495 has shown that if they might travel, like in the sweeps we've seen Johnson and 00:51:00.495 --> 00:51:05.035 Reddish or Pfeiffer and Foster, the animal's standing still. 00:51:05.275 --> 00:51:10.595 The animal's not moving, right? I'm not aware of any data where the animal is 00:51:10.595 --> 00:51:13.195 actually doing this kind of mind travel as it is moving. 00:51:13.375 --> 00:51:18.695 So that means as I now start to move, I take over this whole channel again to 00:51:18.695 --> 00:51:20.535 pipe a velocity signal into my grids. 00:51:20.535 --> 00:51:24.135 Let me stop you, Paul, because I'm afraid I have to tell you that the phase 00:51:24.135 --> 00:51:28.235 precession, the interpretation of the phase precession that I think everybody 00:51:28.235 --> 00:51:32.155 would accept at this point is that while the animal is moving, it is looking ahead. 00:51:32.155 --> 00:51:39.015 But not in this extreme way, as is shown in, let's say, the Johnson and Reddish 00:51:39.015 --> 00:51:42.875 sweeps or the Pfeiffer and Foster paper. But it nevertheless is. 00:51:43.155 --> 00:51:47.155 And I mean, the interesting thing is now when we come to experiments, right? So…. 00:51:48.260 --> 00:51:53.880 Why do I think I'm on the right track? Because this looking ahead that occurs 00:51:53.880 --> 00:51:59.080 while the animal is moving, which is reflected in this phenomenon called phase precession, 00:51:59.260 --> 00:52:03.920 that does disappear when you get rid of the grid cells or when you get rid of 00:52:03.920 --> 00:52:05.060 the medial entorhinal cortex. 00:52:05.440 --> 00:52:09.480 So there is some experimental basis for this, but the play cells survive. 00:52:11.980 --> 00:52:15.860 So we're beginning to get experimental support for this dissociation. 00:52:15.950 --> 00:52:21.590 In fact, it's this sort of dissociation that was the demise of the previous theories. 00:52:24.010 --> 00:52:29.110 We're starting to get a lot of data which now begins to differentiate between different models. 00:52:29.670 --> 00:52:33.050 This model that we're putting together now, I think, is at least consistent 00:52:33.050 --> 00:52:35.770 with all the existing data. It may fall apart when new data comes, 00:52:35.930 --> 00:52:37.270 but this is where we're at. 00:52:37.730 --> 00:52:42.150 I think I'm more happy with your model if I'm understanding it correctly. 00:52:42.150 --> 00:52:45.070 Which we have this metric map 00:52:45.070 --> 00:52:47.930 which is anchored to the room and all you're 00:52:47.930 --> 00:52:51.010 saying is that the grid cell activity in 00:52:51.010 --> 00:52:54.210 the grid cell can represent the fact i'm moving around in 00:52:54.210 --> 00:52:57.890 an imaginary way in the room and that different patterns 00:52:57.890 --> 00:53:01.070 of grid cells are firing to represent those different 00:53:01.070 --> 00:53:04.430 points i'm visiting that will then feed through 00:53:04.430 --> 00:53:12.670 to ca3 ca1 to represent the the non-spatial features of those locations in space 00:53:12.670 --> 00:53:17.030 what it would look like if i was standing there looking around and yeah i'm 00:53:17.030 --> 00:53:21.410 fine with that because that my metric map is anchored and. 00:53:22.490 --> 00:53:27.770 When i when i open my eyes and look then yeah i'll be my my grid cell map i'll 00:53:27.770 --> 00:53:31.790 be back at the point my grid cell map where i really should be so i'm not sure 00:53:31.790 --> 00:53:35.530 there is a problem that you're describing i i thought you were co-opting the 00:53:35.530 --> 00:53:40.030 grid cells to do something other than have a stable, wet graph of space. 00:53:40.350 --> 00:53:43.390 No, I was not. It's only the velocity factor that's changing, actually. 00:53:43.690 --> 00:53:51.710 Right. And, you know, McNaughton has tried to generate models in which all you 00:53:51.710 --> 00:53:57.450 do is you have one sort of grid cell system which is responding both to real 00:53:57.450 --> 00:54:00.930 velocity and to artificial velocity. 00:54:02.150 --> 00:54:03.450 So the idea is that. 00:54:05.302 --> 00:54:09.622 I know where I am because I've integrated real velocity, but now I'm going to 00:54:09.622 --> 00:54:14.122 add an additional velocity and move imaginary through space. 00:54:14.562 --> 00:54:19.882 But, and now what he does is has a speculation about how you can subtract out 00:54:19.882 --> 00:54:23.162 that artificial velocity and 00:54:23.162 --> 00:54:26.642 it says make it a negative, and so that puts you back to where you are. 00:54:27.882 --> 00:54:32.002 So here you could deal with two kinds of velocity in the same network. 00:54:32.182 --> 00:54:41.182 And my only difficulty with that is that it's computationally difficult to have 00:54:41.182 --> 00:54:46.202 two velocity terms and then subtract out the effect of only one of them. 00:54:46.622 --> 00:54:49.302 It's not impossible, but it's not elegant. 00:54:50.362 --> 00:54:55.982 And therefore, I favor the view that you just keep it elegant and have one guy 00:54:55.982 --> 00:55:02.462 who's dealing with true velocity and another guy who's dealing with the artificial velocity, 00:55:02.462 --> 00:55:06.762 but who can be reset after all is said and done by the guy who's keeping track 00:55:06.762 --> 00:55:09.162 of real velocity. It seems more elegant to me. 00:55:09.442 --> 00:55:15.542 But also what you would be saying is that in my look ahead with this mind travel. 00:55:17.002 --> 00:55:20.262 I have a finite capacity to look ahead. 00:55:20.942 --> 00:55:23.362 I can only do this for how many teta cycles? 00:55:24.222 --> 00:55:30.002 It's one full teta cycle with the seven odd gamma responses I have within that? 00:55:30.202 --> 00:55:35.322 Yes. So if you want to now look ahead, let's say if you are in an environment 00:55:35.322 --> 00:55:37.162 that you know very well, like Woods Hole, 00:55:37.382 --> 00:55:41.262 you might be able to look ahead basically to any kind of street, 00:55:41.442 --> 00:55:46.542 any possible trajectory, with much higher precision and much further depth in 00:55:46.542 --> 00:55:49.042 space than you can do in a novel city like here, Barcelona. 00:55:49.942 --> 00:55:54.762 So how then can I change the spatial scale and my resolution on which I can 00:55:54.762 --> 00:55:59.882 look ahead if everything has to happen within a single tether cycle? 00:56:01.296 --> 00:56:10.036 Well, we strongly suspect, based on some evidence, that the whole system is just replicated, 00:56:10.896 --> 00:56:18.416 along the long axis of the hippocampus at larger and larger scales. 00:56:19.596 --> 00:56:24.316 So the dorsal part of the hippocampus, which is the one that's almost exclusively 00:56:24.316 --> 00:56:29.076 studied these days, keeps track of very small distances. I mean, 00:56:29.096 --> 00:56:33.096 we're talking about distances, you know, of a foot or so. 00:56:34.956 --> 00:56:40.156 In contrast, the ventral hippocampus, when you look at the size of the place 00:56:40.156 --> 00:56:43.676 fields, when you look at the phase precession, you know, is looking over meters. 00:56:43.936 --> 00:56:46.116 I'm changing units here, but that's okay. 00:56:46.936 --> 00:56:51.496 I'm in Europe. Yeah, don't worry about it. Confusion is expected. 00:56:53.616 --> 00:56:56.556 So you know and and you know we 00:56:56.556 --> 00:56:59.456 were talking with david reddish the other day about you know why do we 00:56:59.456 --> 00:57:02.436 know so little about the ventral hippocampus is because it's 00:57:02.436 --> 00:57:09.336 so hard to study it's hard to get an electrode in there um and you know it's 00:57:09.336 --> 00:57:14.376 so easy to generate lots and lots of data because you can get in the dorsal 00:57:14.376 --> 00:57:18.236 hippocampus you not only get an electrode in there easily but you get a zillion 00:57:18.236 --> 00:57:19.516 electrodes in there easily. 00:57:20.396 --> 00:57:24.496 So you get really rich data sets, and it just isn't true. 00:57:24.676 --> 00:57:31.456 But, you know, at this point, you know, some data is available about the ventral 00:57:31.456 --> 00:57:35.796 hippocampus, and then I think we'll have, you know, more answers to your question at that point. 00:57:35.936 --> 00:57:38.836 But the available data does show that the scale changes. 00:57:39.656 --> 00:57:44.836 So then, do you see this as a unique feature of hippocampus and toralocortex, 00:57:44.976 --> 00:57:46.296 this ability for mind travel? 00:57:46.296 --> 00:57:52.076 Travel it's unique or can cortex also mind travel hmm that's a very interesting 00:57:52.076 --> 00:57:55.636 question maybe this is a good point to end i have no idea about that. 00:57:57.876 --> 00:58:02.036 No no we're not going to end here john so um 00:58:02.036 --> 00:58:05.256 so it's okay let's let's let's 00:58:05.256 --> 00:58:07.956 see let's see what the future brings us but now so this 00:58:07.956 --> 00:58:10.696 is amazing this is a great moment you're going to all buy his 00:58:10.696 --> 00:58:14.116 beers later on 20 years of teta gamma coding right 00:58:14.116 --> 00:58:17.736 and we still could not talk it out of your head even not in this this podcast 00:58:17.736 --> 00:58:22.836 interview i think you're still going strong on that thank you having having 00:58:22.836 --> 00:58:29.396 you know fought this battle now for so long and still standing and smiling what 00:58:29.396 --> 00:58:32.556 is john's law that we should adhere to to study and understand the brain, 00:58:33.436 --> 00:58:39.476 to study the rest of the brain everything what's john's law what what assures 00:58:39.476 --> 00:58:40.896 progress and understanding the brain. 00:58:41.576 --> 00:58:44.296 Wow. You're on the right of the wall here. John's law. Right, 00:58:44.296 --> 00:58:46.516 John's law. You just have to define it. 00:58:48.288 --> 00:58:52.308 Yeah, that it's not as complicated as you think. 00:58:52.448 --> 00:58:57.788 That is to say that, you know, if we think it through carefully, 00:58:58.108 --> 00:59:01.988 you know, we'll see organization principles fall in. 00:59:02.028 --> 00:59:10.068 So I definitely think that people have underestimated how quickly we will come 00:59:10.068 --> 00:59:11.028 to understand the brain. 00:59:11.028 --> 00:59:14.908 And I think that understanding the brain is a lot like solving, 00:59:15.188 --> 00:59:18.568 you know, other kinds of puzzles. 00:59:18.928 --> 00:59:21.768 And when you look at the process, let's say of a jigsaw puzzle, 00:59:21.948 --> 00:59:26.628 you know, at first you don't see the patterns, you know, you don't know how 00:59:26.628 --> 00:59:31.488 to recognize certain, you know, uniformities of colors and patterns. 00:59:31.488 --> 00:59:35.368 And you also don't have many constraints but 00:59:35.368 --> 00:59:38.068 everything changes you know when you get near 00:59:38.068 --> 00:59:41.108 the end you understand the rules you have constraints 00:59:41.108 --> 00:59:47.648 from other um you know pieces of the puzzle and everything just goes very fast 00:59:47.648 --> 00:59:52.948 as you get near the end and i think we you know we are getting to that point 00:59:52.948 --> 01:00:00.668 okay so don't get confused by complexification right so now the The other thing is, John, 01:00:00.788 --> 01:00:06.168 as you know, Tony likes traveling and he likes hamburgers and Captain Kidd. 01:00:06.588 --> 01:00:10.368 So five years from now, he wants you to take him to Captain Kidd in Woods Hole. 01:00:10.708 --> 01:00:15.028 And then he will confront you with a prediction you're going to make today that 01:00:15.028 --> 01:00:16.888 you will have proven right or wrong by then. 01:00:17.208 --> 01:00:22.268 So what's the most important prediction you would like to commit yourself to 01:00:22.268 --> 01:00:26.388 that needs to be verified by that time? Oh, well, that's easy. 01:00:27.068 --> 01:00:36.768 Because, you know, even older than the theta-gamma idea is the molecular basis of memory, 01:00:36.948 --> 01:00:43.148 KAM kinase, that now we have refined that model to say that it's the complex 01:00:43.148 --> 01:00:47.248 of KAMK2 with the NMDA channel at the synapse. 01:00:48.540 --> 01:00:53.100 That is the actual molecular memory. And in my own lab, 01:00:53.160 --> 01:00:59.880 we've done the critical test to see whether we could erase LTP by attacking 01:00:59.880 --> 01:01:04.080 the CAMK2-NMDA complex, and we could. 01:01:04.800 --> 01:01:11.960 So that was great. And so now we have set up upon the next phase of our work, 01:01:11.980 --> 01:01:14.780 which is to erase a behavioral memory. 01:01:15.760 --> 01:01:22.800 So we have everything set up. We have a nice learning task of spatial memory. 01:01:23.700 --> 01:01:27.960 We have a virus which contains a dominant negative form of CAMK2. 01:01:29.300 --> 01:01:35.780 And so what we're doing is to put the virus into CA1 and see if we can make the animal forget. 01:01:36.580 --> 01:01:44.120 And we've even arranged this to be a spectacular form of virus which does just 01:01:44.120 --> 01:01:46.520 what we want. It turns out to be an HSV virus. 01:01:47.080 --> 01:01:51.880 What's so spectacular about it is that it only expresses for two days. 01:01:53.320 --> 01:01:57.900 And so what we can do is teach the animal something, inject the virus, 01:01:58.160 --> 01:02:00.500 let it do its thing, but then it's gone. 01:02:02.240 --> 01:02:08.920 Now, if the memory is gone, when we measure 10 days later, all those negative, 01:02:08.940 --> 01:02:13.780 nasty reviewers, they can't say, well, the virus is just mucking things up, 01:02:13.820 --> 01:02:15.920 because then we can say, no, the virus is gone. 01:02:16.940 --> 01:02:23.740 And we really, and the only reason that, you know, the memory is not there anymore 01:02:23.740 --> 01:02:25.140 is because we've erased it. 01:02:25.520 --> 01:02:29.040 That's what we want to say. We want to say we were able to erase it. 01:02:29.800 --> 01:02:37.880 Cool. And so we think we have the, so if this works out, then Tony owes me a lot of beers. 01:02:38.740 --> 01:02:43.080 Great. John Lissman, thank you for this conversation. Okay. Thank you. Thank you. 01:02:44.080 --> 01:02:49.200 Music. 01:02:49.045 --> 01:02:54.365 The CSN podcast was produced by the Convergent Science Network of Biometrics 01:02:54.365 --> 01:03:01.105 and Biohybrid Systems, a project funded by the European Sevens Research Framework Programme. 01:03:02.385 --> 01:03:07.645 For more interviews, recorded lectures, or upcoming conferences in the field 01:03:07.645 --> 01:03:13.905 of biometrics and biohybrid systems, go to csnnetwork.eu. 01:03:14.205 --> 01:03:16.065 And thank you for listening. 01:03:20.125 --> 01:03:24.825 That's great, on the anniversary of your paper. Yes, let's see. How old do I have to get? 01:03:28.305 --> 01:03:32.505 Before I prove the KMK2 hypothesis, if it has to be in years of five, 01:03:33.965 --> 01:03:35.205 I'm not going to live that long. 01:03:36.585 --> 01:03:40.285 Well, that sounds like a fantastic experiment. Thank you for generating a prediction, 01:03:41.105 --> 01:03:43.165 on a topic that we hadn't discussed at all. 01:03:46.625 --> 01:03:50.285 That was very helpful, John. It was fun. Yeah, it was very good.