1 00:00:00,000 --> 00:00:04,320 Closing the critical period might actually be neuroprotective. 2 00:00:04,320 --> 00:00:07,920 If we remove the brakes on the plasticity, 3 00:00:07,920 --> 00:00:11,640 then yes, there is a moment for rewiring, 4 00:00:11,640 --> 00:00:15,480 but too much of a good thing could lead to disruption 5 00:00:15,480 --> 00:00:16,320 or degeneration. 6 00:00:26,320 --> 00:00:28,760 I'm John Foxe, director of the Del Monte Institute 7 00:00:28,760 --> 00:00:31,200 for Neuroscience at the University of Rochester. 8 00:00:31,200 --> 00:00:33,760 And I'd like to welcome you to another episode of Neuroscience 9 00:00:33,760 --> 00:00:35,240 Perspectives. 10 00:00:35,240 --> 00:00:37,720 I'm really excited to introduce to you today 11 00:00:37,720 --> 00:00:40,160 my guest, Dr. Takao Hensch. 12 00:00:40,160 --> 00:00:42,440 Dr. Hench is a professor of neurology 13 00:00:42,440 --> 00:00:45,400 at Harvard Medical School at Boston Children's Hospital 14 00:00:45,400 --> 00:00:47,840 and professor of molecular and cellular biology 15 00:00:47,840 --> 00:00:50,120 at Harvard Center for Brain Science. 16 00:00:50,120 --> 00:00:53,400 He leads the National Institute of Mental Health, NIMH, 17 00:00:53,400 --> 00:00:57,340 Silvio Conti Center on Mental Health Research at Harvard 18 00:00:57,340 --> 00:01:01,160 and the International Research Center for Neurointelligence. 19 00:01:01,160 --> 00:01:04,160 He has received a plethora of honors, 20 00:01:04,160 --> 00:01:07,120 including the NIH Director's Pioneer Award 21 00:01:07,120 --> 00:01:11,320 and the Mort Sackler MD Prize for Distinguished Achievement 22 00:01:11,320 --> 00:01:13,840 in Developmental Psychobiology. 23 00:01:13,840 --> 00:01:17,040 His research examines how early life experience shapes 24 00:01:17,040 --> 00:01:20,120 brain function, critical periods of brain development, 25 00:01:20,120 --> 00:01:23,160 and whether plasticity can be opened or targeted 26 00:01:23,160 --> 00:01:26,840 to treat neurodevelopmental and neurodegenerative diseases. 27 00:01:26,840 --> 00:01:28,040 So thank you for joining us. 28 00:01:28,040 --> 00:01:32,040 And I'm just absolutely delighted to have you here, Takawa, 29 00:01:32,040 --> 00:01:35,040 to have a chat here in Neuroscience Perspectives. 30 00:01:35,040 --> 00:01:36,120 Welcome to Rochester. 31 00:01:36,120 --> 00:01:36,960 Thank you for having me. 32 00:01:36,960 --> 00:01:38,360 I know you had a late flight in, 33 00:01:38,360 --> 00:01:43,360 so we really appreciate you getting here and joining us. 34 00:01:44,040 --> 00:01:46,280 Let's get started with your research. 35 00:01:46,280 --> 00:01:48,680 What we're gonna definitely want to do is come back 36 00:01:48,680 --> 00:01:50,320 and understand your journey into science. 37 00:01:50,320 --> 00:01:53,740 But let's talk a little bit about critical periods. 38 00:01:53,740 --> 00:01:54,880 I think actually, you know, 39 00:01:54,880 --> 00:01:56,840 probably everybody on the planet who's read 40 00:01:56,840 --> 00:02:00,000 a basic science book knows about critical periods. 41 00:02:00,000 --> 00:02:02,720 And probably baked into their thinking is, you know, 42 00:02:02,720 --> 00:02:06,120 there's a very small window of time during neurodevelopment 43 00:02:06,120 --> 00:02:08,840 when these periods open up and they close. 44 00:02:08,840 --> 00:02:10,400 And if you haven't learned what you need to learn 45 00:02:10,400 --> 00:02:12,580 in those periods of time, the game is up. 46 00:02:13,500 --> 00:02:14,920 Disabuse us of that notion. 47 00:02:15,840 --> 00:02:17,840 Right, well, throughout human history, 48 00:02:17,840 --> 00:02:20,800 we've appreciated the importance of early childhood 49 00:02:20,800 --> 00:02:23,980 and infancy and shaping our identities. 50 00:02:23,980 --> 00:02:28,040 And this taps into some fundamental biology. 51 00:02:28,040 --> 00:02:31,120 From mouse to human, we see that brain functions 52 00:02:31,120 --> 00:02:35,280 are shaped early in life and that many of them 53 00:02:35,280 --> 00:02:37,680 are kind of locked in. 54 00:02:37,680 --> 00:02:41,280 The degree to which that's true in terms of reversibility 55 00:02:41,280 --> 00:02:44,840 is something that's being actively researched now. 56 00:02:44,840 --> 00:02:47,940 And it's the advent of modern neurobiological techniques 57 00:02:47,940 --> 00:02:50,840 that allow us to understand what opens these windows, 58 00:02:50,840 --> 00:02:52,160 determines their duration, 59 00:02:52,160 --> 00:02:54,840 and ultimately might close them or not. 60 00:02:54,840 --> 00:02:59,000 And by following those paths, we can try to understand 61 00:02:59,000 --> 00:03:01,720 whether they really are closed for good. 62 00:03:01,720 --> 00:03:06,080 Right, and so is the timing of these windows 63 00:03:06,080 --> 00:03:08,960 of opportunity to learn specific functions, 64 00:03:08,960 --> 00:03:11,480 it's stereotyped across individuals, 65 00:03:11,480 --> 00:03:13,200 is it driven by the environment? 66 00:03:13,200 --> 00:03:15,740 What's turning on and off these windows? 67 00:03:15,740 --> 00:03:19,480 Well, it's the classic gene environment interaction question. 68 00:03:19,480 --> 00:03:23,720 And in fact, we know that there are many cellular components 69 00:03:23,720 --> 00:03:26,600 that contribute to the timing mechanism now. 70 00:03:26,600 --> 00:03:29,960 And those are in fact sensitive to environment. 71 00:03:29,960 --> 00:03:32,840 So the answer is of course both are involved. 72 00:03:32,840 --> 00:03:36,920 And in certain extreme cases, mental illness, 73 00:03:36,920 --> 00:03:40,520 you might see this play out in a very dramatic way 74 00:03:40,520 --> 00:03:42,240 as a shift in timing. 75 00:03:42,240 --> 00:03:44,760 Right, and are these developmental windows, 76 00:03:44,760 --> 00:03:48,240 these critical periods, they're in utero as well as 77 00:03:48,240 --> 00:03:49,520 after the birth, is that? 78 00:03:49,520 --> 00:03:52,280 That's right, so some systems like the auditory system 79 00:03:52,280 --> 00:03:54,320 is coming online before birth. 80 00:03:54,320 --> 00:03:59,160 And there's a first important notion 81 00:03:59,160 --> 00:04:01,480 that critical periods are staggered 82 00:04:01,480 --> 00:04:05,080 and not happening synchronously across brain modalities. 83 00:04:05,080 --> 00:04:07,740 So different functions coming online over time, 84 00:04:07,740 --> 00:04:09,640 and maybe the sense is rolling out 85 00:04:09,640 --> 00:04:12,620 in a pseudo sequential form or? 86 00:04:12,620 --> 00:04:15,260 That's right, so there's a rough sense of hierarchy 87 00:04:15,260 --> 00:04:18,440 that primary sensory areas, the first filters 88 00:04:18,440 --> 00:04:21,120 to the outside world seem to be shaped earlier 89 00:04:21,120 --> 00:04:22,940 than higher cognitive functions. 90 00:04:22,940 --> 00:04:25,920 But this is a very rough approximation. 91 00:04:25,920 --> 00:04:28,860 The important notion is that there's not one critical period. 92 00:04:28,860 --> 00:04:30,500 There are multiple critical periods. 93 00:04:30,500 --> 00:04:34,920 Right, right, and then your research really looks at 94 00:04:34,920 --> 00:04:38,720 the molecular biology and the physiology, 95 00:04:38,720 --> 00:04:40,520 neurophysiology of that. 96 00:04:40,520 --> 00:04:43,760 And are there insights that, if you were to say, 97 00:04:43,760 --> 00:04:45,560 like what are my top three things that I know 98 00:04:45,560 --> 00:04:47,640 about this plasticity? 99 00:04:47,640 --> 00:04:49,920 What is plasticity and what are those things 100 00:04:49,920 --> 00:04:53,080 that really give rise to the ability to learn? 101 00:04:53,080 --> 00:04:56,260 Right, well plasticity of course is the ability 102 00:04:56,260 --> 00:05:00,120 to adapt to change, and our brain is a plastic machine. 103 00:05:00,120 --> 00:05:01,320 That's its job. 104 00:05:01,320 --> 00:05:05,460 But we know that this degree of plasticity 105 00:05:05,460 --> 00:05:08,200 changes dynamically across the lifespan. 106 00:05:08,200 --> 00:05:11,880 And so critical periods or sensitive periods arise 107 00:05:11,880 --> 00:05:15,540 because early life experiences are particularly potent 108 00:05:15,540 --> 00:05:17,140 in shaping the brain. 109 00:05:17,140 --> 00:05:21,620 And the experimental work is trying to understand 110 00:05:21,620 --> 00:05:23,220 why that is. 111 00:05:23,220 --> 00:05:25,320 You know, a reasonable question would be, 112 00:05:25,320 --> 00:05:28,060 if this plasticity is so great for learning, 113 00:05:28,060 --> 00:05:30,240 why on earth do we shut it down at all? 114 00:05:30,240 --> 00:05:32,780 Why wouldn't we stay that way throughout the duration 115 00:05:32,780 --> 00:05:33,620 of a lifetime? 116 00:05:33,620 --> 00:05:35,340 Right, that is a great question. 117 00:05:35,340 --> 00:05:40,340 And I think it's always tempting to think more is better. 118 00:05:40,340 --> 00:05:43,980 But as we learn how these windows come about, 119 00:05:43,980 --> 00:05:46,960 and there's some surprises that we've come across 120 00:05:46,960 --> 00:05:49,740 along the way, we understand better why it's important 121 00:05:49,740 --> 00:05:52,660 to dial down and stabilize circuitry. 122 00:05:52,660 --> 00:05:54,540 And in fact, we spend most of our life 123 00:05:54,540 --> 00:05:57,340 in this more stable processing mode. 124 00:05:57,340 --> 00:06:01,360 I guess there are two insights I could elaborate on. 125 00:06:01,360 --> 00:06:04,380 Just computationally speaking, it wouldn't make sense 126 00:06:04,380 --> 00:06:07,800 to rewire with every possible experience. 127 00:06:07,800 --> 00:06:10,540 And in fact, that might be a condition akin 128 00:06:10,540 --> 00:06:13,660 to certain mental illnesses where mechanisms 129 00:06:13,660 --> 00:06:16,900 of closing critical periods are not fully active. 130 00:06:18,140 --> 00:06:19,380 And then from- 131 00:06:19,380 --> 00:06:20,660 Could you give us, sorry to interrupt, 132 00:06:20,660 --> 00:06:21,780 but could you give us an example of that? 133 00:06:21,780 --> 00:06:23,540 I mean, that's a fascinating notion 134 00:06:23,540 --> 00:06:25,740 that it's plasticity run awry 135 00:06:25,740 --> 00:06:27,580 that might give rise to a mental illness. 136 00:06:27,580 --> 00:06:30,860 Right, so at the cellular level, 137 00:06:30,860 --> 00:06:34,100 plasticity is ultimately about rewiring connections. 138 00:06:34,100 --> 00:06:37,960 And one example of that is the pruning of dendritic spines 139 00:06:37,960 --> 00:06:40,260 on the predominant excitatory neurons 140 00:06:40,260 --> 00:06:42,020 in the cortex, for example. 141 00:06:42,020 --> 00:06:44,700 And in mental illnesses like schizophrenia, 142 00:06:44,700 --> 00:06:47,820 a hallmark signature is excessive pruning. 143 00:06:47,820 --> 00:06:50,260 And so this could be because critical periods 144 00:06:50,260 --> 00:06:54,460 have not fully closed and excessive remodeling 145 00:06:54,460 --> 00:06:56,340 has been ongoing. 146 00:06:56,340 --> 00:06:58,320 And in fact, that's how we got interested 147 00:06:58,320 --> 00:07:01,420 in the mental illness angle of critical periods. 148 00:07:01,420 --> 00:07:05,740 As we started to unearth the different mechanisms 149 00:07:05,740 --> 00:07:08,460 that are involved in closure, 150 00:07:08,460 --> 00:07:12,860 they were being linked separately in GWAS studies 151 00:07:12,860 --> 00:07:14,820 to schizophrenia, for example. 152 00:07:14,820 --> 00:07:17,500 So in the big genome-wide association studies, 153 00:07:17,500 --> 00:07:19,700 the genes responsible for this plasticity 154 00:07:19,700 --> 00:07:24,340 are popping up as candidates in this terrible disease. 155 00:07:24,340 --> 00:07:25,740 That's right. 156 00:07:25,740 --> 00:07:27,380 That's very interesting, right, 157 00:07:27,380 --> 00:07:30,020 which would also give some explanation, 158 00:07:30,020 --> 00:07:31,540 which I think fascinates people to, 159 00:07:31,540 --> 00:07:34,060 like, why does schizophrenia emerge 160 00:07:34,060 --> 00:07:37,140 in the late teens and the early 20s, 161 00:07:37,140 --> 00:07:39,980 rather than, you know, it's got that peculiar time course 162 00:07:39,980 --> 00:07:41,140 to it, and of course, this would provide 163 00:07:41,140 --> 00:07:42,300 an explanation for that. 164 00:07:42,300 --> 00:07:45,500 Right, as well as the fact that these windows happen 165 00:07:45,500 --> 00:07:47,700 at different times in different brain regions. 166 00:07:47,700 --> 00:07:50,100 And so the kind of executive functions 167 00:07:50,100 --> 00:07:52,140 that are compromised in schizophrenia 168 00:07:52,140 --> 00:07:54,380 related to prefrontal brain function 169 00:07:54,380 --> 00:07:58,140 are naturally where these windows are closing last. 170 00:07:58,140 --> 00:07:59,340 Right, right. 171 00:07:59,340 --> 00:08:00,900 It gives me two questions. 172 00:08:00,900 --> 00:08:04,500 So, you know, I think people, again, 173 00:08:04,500 --> 00:08:06,460 will be aware now that we now know 174 00:08:06,460 --> 00:08:09,940 that some of this development of brain architecture, 175 00:08:09,940 --> 00:08:11,660 particularly in the prefrontal lobes, 176 00:08:11,660 --> 00:08:13,340 continues right into the 20s. 177 00:08:13,340 --> 00:08:17,300 So some of these critical periods are really late in life, 178 00:08:17,300 --> 00:08:20,020 you know, not an infancy business. 179 00:08:20,020 --> 00:08:21,060 Is that the case? 180 00:08:21,060 --> 00:08:21,900 That's right. 181 00:08:21,900 --> 00:08:24,340 And in fact, in the human, in some brain, 182 00:08:24,340 --> 00:08:27,340 higher-order brain areas, they may never really close. 183 00:08:27,340 --> 00:08:29,700 And that's the second insight I was alluding to. 184 00:08:29,700 --> 00:08:33,860 At a cellular level, the genes that are related 185 00:08:33,860 --> 00:08:37,780 to critical period closure seem to be, surprisingly, 186 00:08:37,780 --> 00:08:42,060 break-like factors that inhibit physically 187 00:08:42,060 --> 00:08:44,740 or functionally the plastic process, 188 00:08:44,740 --> 00:08:48,340 which would mean that critical period closure 189 00:08:48,340 --> 00:08:51,380 is an active process, not the traditional thinking 190 00:08:51,380 --> 00:08:53,660 that plasticity fades away with age. 191 00:08:53,660 --> 00:08:55,620 That's the phenomenology. 192 00:08:55,620 --> 00:08:58,780 But it's actually not because of the loss of plasticity 193 00:08:58,780 --> 00:09:03,180 so much as the active prevention of plasticity. 194 00:09:03,180 --> 00:09:06,740 And it suggests that if you look at brain regions 195 00:09:06,740 --> 00:09:10,380 that have evolved in humans that are not present in mice, 196 00:09:10,380 --> 00:09:14,140 for example, and tend to stay plastic longer, 197 00:09:14,140 --> 00:09:17,900 sure enough, we find fewer of these break-like factors there, 198 00:09:17,900 --> 00:09:21,260 consistent with the idea that human intelligence 199 00:09:21,260 --> 00:09:23,580 has benefited from adding areas 200 00:09:23,580 --> 00:09:25,980 that don't close this plastic window. 201 00:09:25,980 --> 00:09:28,340 Good, that's absolutely fascinating. 202 00:09:28,340 --> 00:09:30,300 So now, of course, that brings us to, 203 00:09:30,300 --> 00:09:32,660 can you get in there and turn on and off these switches 204 00:09:32,660 --> 00:09:34,700 that be an obvious benefit? 205 00:09:34,700 --> 00:09:37,900 I suppose, if you think about things like stroke 206 00:09:37,900 --> 00:09:40,180 or that where people lose a function 207 00:09:40,180 --> 00:09:42,500 and they don't have the plasticity to remap, 208 00:09:42,500 --> 00:09:46,860 to bring brain circuits on to compensate, 209 00:09:46,860 --> 00:09:49,220 is there, there's opportunity there, right? 210 00:09:49,220 --> 00:09:50,660 And that's a big piece of what you do. 211 00:09:50,660 --> 00:09:54,900 Yes, so the kind of science fiction-like notion 212 00:09:54,900 --> 00:09:58,940 of reopening or rejuvenating plasticity in the adult brain 213 00:09:58,940 --> 00:10:02,660 has a very powerful therapeutic implication 214 00:10:02,660 --> 00:10:06,500 for recovery from brain injury, stroke, and adulthood. 215 00:10:06,500 --> 00:10:08,700 Of course, as you've mentioned, 216 00:10:08,700 --> 00:10:11,220 we have to do this in a very measured, careful way 217 00:10:11,220 --> 00:10:15,100 because evolution has turned on these breaks for a reason, 218 00:10:15,100 --> 00:10:16,980 and we can talk more about that. 219 00:10:16,980 --> 00:10:21,420 But the goal with our work at Children's Hospital 220 00:10:21,420 --> 00:10:26,300 and other clinically relevant venues is exactly this. 221 00:10:26,300 --> 00:10:28,300 Can we leverage critical period biology 222 00:10:28,300 --> 00:10:30,300 to recover brain function? 223 00:10:30,300 --> 00:10:32,260 And what, would an aspect of that be 224 00:10:32,260 --> 00:10:35,780 spatially specific targeting and turning on and off circuits? 225 00:10:35,780 --> 00:10:37,460 I mean, there's the obvious thing 226 00:10:37,460 --> 00:10:40,300 to do something systemic and you open up the whole brain 227 00:10:40,300 --> 00:10:43,180 and this may be not where you wanna go. 228 00:10:43,180 --> 00:10:47,180 Right, so now that we know that critical periods 229 00:10:47,180 --> 00:10:49,980 happen sequentially in a well-orchestrated manner 230 00:10:49,980 --> 00:10:51,380 thanks to triggers and breaks 231 00:10:51,380 --> 00:10:53,260 that open and close these windows, 232 00:10:54,180 --> 00:10:56,180 you could imagine how damaging it would be 233 00:10:56,180 --> 00:10:59,380 to reopen the whole brain at once. 234 00:10:59,380 --> 00:11:03,500 But there are probably some fail-safe mechanisms 235 00:11:03,500 --> 00:11:07,500 there as well, and so it's not that we are making 236 00:11:07,500 --> 00:11:10,300 the brain plastic so much as opening a gate 237 00:11:10,300 --> 00:11:13,340 that's permissive for training to change 238 00:11:13,340 --> 00:11:15,340 in a modality-specific way. 239 00:11:15,340 --> 00:11:17,380 Amazing, absolutely astounding. 240 00:11:17,380 --> 00:11:20,820 Takao, you were talking about this business of 241 00:11:20,820 --> 00:11:23,420 the closing of critical periods 242 00:11:23,420 --> 00:11:25,900 and at a relatively high level. 243 00:11:25,900 --> 00:11:27,980 Can we talk a little bit at a more granular level 244 00:11:27,980 --> 00:11:29,540 about some of the molecular biology 245 00:11:29,540 --> 00:11:31,420 and some insights that you might have there? 246 00:11:31,420 --> 00:11:32,540 Yeah, sure. 247 00:11:32,540 --> 00:11:37,540 I think there have been two very surprising discoveries 248 00:11:37,960 --> 00:11:40,180 in the study of critical periods. 249 00:11:40,180 --> 00:11:43,020 One is that they are very sensitive 250 00:11:43,020 --> 00:11:45,820 to the development of inhibitory neurons 251 00:11:45,820 --> 00:11:50,100 and that the timing of these windows can be movable 252 00:11:50,100 --> 00:11:51,820 depending on when these inhibitory, 253 00:11:51,820 --> 00:11:55,020 particular class of inhibitory cell matures. 254 00:11:55,020 --> 00:11:59,100 It's surprising because plasticity is often studied 255 00:11:59,100 --> 00:12:02,220 at excitatory connections onto excitatory neurons 256 00:12:02,220 --> 00:12:04,620 which are far more abundant, 257 00:12:04,620 --> 00:12:07,620 but the inhibitory cells seem to drive the bus 258 00:12:07,620 --> 00:12:09,860 and they are of a particular type. 259 00:12:09,860 --> 00:12:12,060 They are fast spiking cells. 260 00:12:12,060 --> 00:12:15,220 They use a lot of energy and so are vulnerable 261 00:12:15,220 --> 00:12:18,020 to oxidative stress which they generate. 262 00:12:18,020 --> 00:12:21,540 And in fact, some of the closure mechanisms 263 00:12:21,540 --> 00:12:25,100 are related to dampening this oxidative stress. 264 00:12:25,100 --> 00:12:27,940 And so closing the critical period 265 00:12:27,940 --> 00:12:30,340 might actually be neuroprotective 266 00:12:30,340 --> 00:12:34,900 and that if we remove the brakes on the plasticity, 267 00:12:34,900 --> 00:12:38,640 then yes, there is a moment for rewiring, 268 00:12:38,640 --> 00:12:40,780 but too much of a good thing could lead 269 00:12:40,780 --> 00:12:44,060 to disruption or degeneration. 270 00:12:44,060 --> 00:12:46,580 And a change in this excitatory inhibitory balance 271 00:12:46,580 --> 00:12:49,460 which is a big piece of our important theory 272 00:12:49,460 --> 00:12:50,660 in neurodegenerative disorder. 273 00:12:50,660 --> 00:12:53,020 And that's how we got into autism in fact. 274 00:12:53,020 --> 00:12:56,380 So discovering that inhibition was pivotal 275 00:12:56,380 --> 00:12:58,820 and that this balance is what we should be looking at, 276 00:12:58,820 --> 00:13:02,700 not just LTP of excitatory connections 277 00:13:02,700 --> 00:13:07,700 has really made headway into autism research. 278 00:13:07,700 --> 00:13:08,540 Yes. 279 00:13:08,540 --> 00:13:10,300 I first came to know about your work 280 00:13:10,300 --> 00:13:13,580 because you were working on mouse models of autism. 281 00:13:13,580 --> 00:13:16,980 I wanted to get a little bit controversial 282 00:13:16,980 --> 00:13:19,540 because I think there are a lot of folks out there 283 00:13:19,540 --> 00:13:23,540 say how on earth is it possible to model a disease 284 00:13:23,540 --> 00:13:25,500 as complex as autism? 285 00:13:25,500 --> 00:13:28,340 Where the symptoms we think about are social communications 286 00:13:28,340 --> 00:13:30,300 and repetitive behaviors and stuff, 287 00:13:30,300 --> 00:13:32,740 the sort of defining symptom clusters. 288 00:13:32,740 --> 00:13:35,140 How can you possibly be studying that in the mouse? 289 00:13:35,140 --> 00:13:38,180 Do you want to give us a little tutorial on that? 290 00:13:38,180 --> 00:13:39,740 Oh, certainly, yes. 291 00:13:39,740 --> 00:13:43,700 We would never be so bold as to claim a mouse has autism. 292 00:13:43,700 --> 00:13:47,440 But they are a very powerful living test tube 293 00:13:47,440 --> 00:13:49,580 of what happens to a complex circuit 294 00:13:49,580 --> 00:13:53,620 when a gene implicated in autism is disrupted. 295 00:13:53,620 --> 00:13:57,060 And so that's how we treat the model system. 296 00:13:57,060 --> 00:14:00,740 Said this with modern machine learning approaches 297 00:14:00,740 --> 00:14:03,840 and more sophisticated analysis of behavior 298 00:14:03,840 --> 00:14:06,420 and being able to think more like a mouse, 299 00:14:06,420 --> 00:14:10,100 we might be able to extract a complex phenotype 300 00:14:10,100 --> 00:14:11,180 even in these animals. 301 00:14:11,180 --> 00:14:14,600 So I think these two thought processes 302 00:14:14,600 --> 00:14:16,160 are running in parallel. 303 00:14:16,160 --> 00:14:19,440 But we've learned quite a bit about how local circuits 304 00:14:19,440 --> 00:14:23,020 and synapses develop, taking the approach 305 00:14:23,020 --> 00:14:27,420 that genes linked to human autism can be probed in mice. 306 00:14:27,420 --> 00:14:28,260 Fantastic. 307 00:14:28,260 --> 00:14:29,940 Okay, that's crystal clear. 308 00:14:29,940 --> 00:14:31,220 Very well put, I must say. 309 00:14:31,220 --> 00:14:34,800 If we can, can we dial the clock back? 310 00:14:34,800 --> 00:14:37,240 And we had a little chat before 311 00:14:37,240 --> 00:14:41,640 and I got to hear a little bit about your trajectory. 312 00:14:41,640 --> 00:14:43,560 You weren't born on these shores. 313 00:14:43,560 --> 00:14:45,080 Can we go back to the beginning, 314 00:14:45,080 --> 00:14:47,160 the genesis of Takawa Hench 315 00:14:47,160 --> 00:14:50,320 and what got you to where you're at at Harvard, 316 00:14:50,320 --> 00:14:51,920 one of the great institutions on the planet 317 00:14:51,920 --> 00:14:54,440 studying autism and mouse models? 318 00:14:55,600 --> 00:14:57,160 Where were you born? 319 00:14:57,160 --> 00:14:58,000 Sure. 320 00:14:58,000 --> 00:14:59,800 How did that impact your development? 321 00:14:59,800 --> 00:15:02,440 Right, well, as you can guess from my name, 322 00:15:02,440 --> 00:15:04,880 I'm half Japanese, half German, 323 00:15:04,880 --> 00:15:09,000 and my father met my mother in Tokyo. 324 00:15:09,000 --> 00:15:14,000 And he was an engineer, computer scientist back in the day. 325 00:15:14,360 --> 00:15:19,360 And he was sent to Japan by IBM 326 00:15:19,360 --> 00:15:23,200 to design the first Chinese character keyboards. 327 00:15:23,200 --> 00:15:25,240 And as you can imagine, 328 00:15:25,240 --> 00:15:27,560 there are thousands of Chinese characters 329 00:15:27,560 --> 00:15:31,880 to encode all of that in keyboards was quite a challenge. 330 00:15:31,880 --> 00:15:34,760 And he came up with a coding scheme, 331 00:15:34,760 --> 00:15:38,960 double-byte character set to encode the characters, 332 00:15:38,960 --> 00:15:41,640 even though he didn't speak a word of Japanese, 333 00:15:41,640 --> 00:15:44,520 taking a very German engineering approach. 334 00:15:44,520 --> 00:15:45,920 And that's still used today 335 00:15:45,920 --> 00:15:48,880 in encoding these characters on keyboards. 336 00:15:48,880 --> 00:15:49,720 Oh, extraordinary. 337 00:15:49,720 --> 00:15:53,640 And during that time, he met my mother. 338 00:15:53,640 --> 00:15:56,680 It was around the time of the first Tokyo Olympic Games. 339 00:15:56,680 --> 00:16:01,240 And I was raised in a multilingual environment. 340 00:16:01,240 --> 00:16:02,500 They took it upon themselves 341 00:16:02,500 --> 00:16:04,960 to speak only their native language with me. 342 00:16:04,960 --> 00:16:07,720 And then we moved to New York. 343 00:16:07,720 --> 00:16:09,320 My father was moved to IBM. 344 00:16:09,320 --> 00:16:10,160 Let me jump in. 345 00:16:10,160 --> 00:16:13,400 So your father was learning Japanese at this point? 346 00:16:13,400 --> 00:16:14,800 Yes, yes, of course. 347 00:16:14,800 --> 00:16:17,480 Given the job he was asked to do. 348 00:16:17,480 --> 00:16:20,660 And my mother was studying German actually, as it turned out. 349 00:16:20,660 --> 00:16:25,660 And so they happened to meet in that way. 350 00:16:25,660 --> 00:16:29,940 And then his job took him to New York 351 00:16:29,940 --> 00:16:33,220 and the whole family moved to the United States 352 00:16:33,220 --> 00:16:34,060 when I was three. 353 00:16:34,060 --> 00:16:35,780 So my- 354 00:16:35,780 --> 00:16:36,860 But at three years of age, 355 00:16:36,860 --> 00:16:39,360 your bilingual Japanese, 356 00:16:39,360 --> 00:16:43,260 or at least a proto bilingual Japanese German speaker, 357 00:16:43,260 --> 00:16:44,580 you haven't heard a word of English at this point. 358 00:16:44,580 --> 00:16:46,500 Not a word of English, that's right. 359 00:16:46,500 --> 00:16:49,580 And then English came in. 360 00:16:49,580 --> 00:16:52,420 But fortunately for me, everything was compartmentalized. 361 00:16:52,420 --> 00:16:56,420 So English was friends and outside the house. 362 00:16:56,420 --> 00:16:57,740 And I grew up in that way. 363 00:16:57,740 --> 00:17:00,520 I also attended a Japanese school in New York 364 00:17:03,020 --> 00:17:04,900 in parallel to the American school. 365 00:17:04,900 --> 00:17:08,500 And so I was able to keep the languages separate 366 00:17:08,500 --> 00:17:09,700 in that way. 367 00:17:09,700 --> 00:17:12,020 And that's what drew my interest to the brain. 368 00:17:12,020 --> 00:17:15,180 And that goes to extraordinary plasticity. 369 00:17:15,180 --> 00:17:17,540 I mean, I think this is one of the things, of course, 370 00:17:17,540 --> 00:17:20,340 about being in America and painting in broad strokes, 371 00:17:20,340 --> 00:17:23,460 but a great number of people in America grow up 372 00:17:23,460 --> 00:17:25,620 in a monolingual environment. 373 00:17:25,620 --> 00:17:28,980 And just the idea that you can pack three entire languages 374 00:17:28,980 --> 00:17:30,980 simultaneously into a child's brain. 375 00:17:30,980 --> 00:17:34,140 I mean, the plasticity must be extraordinary. 376 00:17:34,140 --> 00:17:37,380 Yes, it was surprising to me actually that, 377 00:17:37,380 --> 00:17:39,540 in school we learn a second language. 378 00:17:39,540 --> 00:17:40,780 And so I took French. 379 00:17:40,780 --> 00:17:43,900 So just for good measure, you decided to do number four. 380 00:17:43,900 --> 00:17:48,500 And the French class really opened my eyes 381 00:17:48,500 --> 00:17:51,140 that most other kids were not growing up 382 00:17:51,140 --> 00:17:53,260 in a trilingual environment. 383 00:17:53,260 --> 00:17:54,220 And- 384 00:17:54,220 --> 00:17:55,300 I'm really struggling with it. 385 00:17:55,300 --> 00:17:59,060 Yeah, learning French was somehow easier 386 00:17:59,060 --> 00:18:01,580 because I guess I was used to the idea 387 00:18:01,580 --> 00:18:04,780 of multiple representations for the same objects. 388 00:18:06,020 --> 00:18:09,240 And so that's when I started to develop this fascination 389 00:18:09,240 --> 00:18:13,660 with how early life experience can change brain function. 390 00:18:13,660 --> 00:18:14,500 Right, right, right. 391 00:18:14,500 --> 00:18:15,320 Amazing. 392 00:18:15,320 --> 00:18:16,780 Any other languages that we need to know about? 393 00:18:16,780 --> 00:18:18,340 Well, my wife is Italian. 394 00:18:18,340 --> 00:18:19,820 Goodness me. 395 00:18:19,820 --> 00:18:20,660 Working on that. 396 00:18:20,660 --> 00:18:21,500 You're working on that too. 397 00:18:21,500 --> 00:18:23,180 And finding it easy or? 398 00:18:23,180 --> 00:18:26,500 Yes, it's more confusing because of the French. 399 00:18:26,500 --> 00:18:27,340 Yeah. 400 00:18:27,340 --> 00:18:29,220 And so it's a latecomer. 401 00:18:29,220 --> 00:18:33,460 And of course, different words pop in at that point. 402 00:18:33,460 --> 00:18:34,380 Yeah, yeah, yeah. 403 00:18:34,380 --> 00:18:35,940 I'm actually in the process of trying to learn 404 00:18:35,940 --> 00:18:37,860 a little bit of Spanish. 405 00:18:37,860 --> 00:18:41,740 But growing up in Ireland where we have Irish and English, 406 00:18:41,740 --> 00:18:45,260 the fact that that facility is there as well, 407 00:18:45,260 --> 00:18:47,020 I think it makes it a little bit easier 408 00:18:47,020 --> 00:18:49,220 to catch on to a new language. 409 00:18:49,220 --> 00:18:50,300 Right. 410 00:18:50,300 --> 00:18:53,340 And so that is actually a form of, 411 00:18:53,340 --> 00:18:56,980 this business of being multilingual, polyglot, 412 00:18:56,980 --> 00:19:00,940 has really shaped your thinking in a lot of ways. 413 00:19:00,940 --> 00:19:01,860 Very much so. 414 00:19:01,860 --> 00:19:05,500 It shaped my journey into neuroscience 415 00:19:06,840 --> 00:19:11,340 and also how I traversed the trajectory 416 00:19:11,340 --> 00:19:13,220 where I sought out training. 417 00:19:13,220 --> 00:19:14,140 Right, right. 418 00:19:14,140 --> 00:19:14,980 Tell us about that. 419 00:19:14,980 --> 00:19:18,460 So you moved to Long Island, right? 420 00:19:18,460 --> 00:19:19,740 And then you moved to Westchester County, 421 00:19:19,740 --> 00:19:20,940 to Tarrytown, I believe. 422 00:19:20,940 --> 00:19:21,780 Yes. 423 00:19:21,780 --> 00:19:23,220 And that's where you did your schooling, 424 00:19:23,220 --> 00:19:24,620 high school and all that. 425 00:19:24,620 --> 00:19:26,100 And then college from there? 426 00:19:26,100 --> 00:19:30,300 Yes, I went to Harvard and I was very gung-ho 427 00:19:30,300 --> 00:19:33,220 on the first wave, I'm dating myself now, 428 00:19:33,220 --> 00:19:36,800 of AI really, where expert systems, 429 00:19:36,800 --> 00:19:38,660 as they were called in the day, 430 00:19:38,660 --> 00:19:40,180 were achieving great things. 431 00:19:40,180 --> 00:19:41,780 So not neuroscience in the beginning, 432 00:19:41,780 --> 00:19:43,620 you were thinking, was this because 433 00:19:43,620 --> 00:19:45,300 of your dad's engineering background? 434 00:19:45,300 --> 00:19:46,380 Probably, most likely. 435 00:19:46,380 --> 00:19:48,140 Yes, the summer before college, 436 00:19:48,140 --> 00:19:51,220 I worked at IBM Watson Laboratories 437 00:19:51,220 --> 00:19:52,540 in a natural language processing lab. 438 00:19:52,540 --> 00:19:53,380 Which is right there, right? 439 00:19:53,380 --> 00:19:54,220 That's right. 440 00:19:54,220 --> 00:19:55,060 In the northwest, right? 441 00:19:55,060 --> 00:19:57,900 Yeah, in Yorktown Heights, that's right. 442 00:19:57,900 --> 00:20:00,780 And so with that enthusiasm, 443 00:20:00,780 --> 00:20:03,280 I arrived at Harvard thinking computer science 444 00:20:03,280 --> 00:20:04,740 will solve everything. 445 00:20:04,740 --> 00:20:07,460 And then of course, realizing that we know 446 00:20:07,460 --> 00:20:09,120 precious little about the brain. 447 00:20:10,180 --> 00:20:13,020 At the time, it was shortly after Hubel and Wiesel 448 00:20:13,020 --> 00:20:15,820 had won the Nobel Prize for understanding 449 00:20:15,820 --> 00:20:18,580 the fundamental organization of the visual cortex 450 00:20:18,580 --> 00:20:20,320 and critical periods. 451 00:20:20,320 --> 00:20:25,320 And so I switched more into neurobiology at that point. 452 00:20:25,580 --> 00:20:26,580 As an undergraduate? 453 00:20:26,580 --> 00:20:27,420 As an undergraduate. 454 00:20:27,420 --> 00:20:28,260 I see, yeah. 455 00:20:28,260 --> 00:20:29,100 Yeah, and... 456 00:20:29,100 --> 00:20:32,700 And you were really in a hotbed of scientific discovery 457 00:20:32,700 --> 00:20:34,380 in the neurosciences at that point at Harvard. 458 00:20:34,380 --> 00:20:36,700 Yeah, so much going on. 459 00:20:36,700 --> 00:20:40,540 And I kind of worked my way up the noraxis. 460 00:20:40,540 --> 00:20:44,060 I did my undergraduate thesis with Alan Hobson, 461 00:20:44,060 --> 00:20:47,220 the late Alan Hobson in sleep research. 462 00:20:47,220 --> 00:20:52,220 And from there, took a fellowship to Masao Ito's lab 463 00:20:52,580 --> 00:20:53,940 at the University of Tokyo. 464 00:20:53,940 --> 00:20:57,220 He's of course the godfather of the cerebellum 465 00:20:57,220 --> 00:21:00,820 and plasticity in the cerebellar cortex. 466 00:21:00,820 --> 00:21:04,060 And then had a full bright year with Wolf Singer 467 00:21:04,060 --> 00:21:06,760 and the Max Planck in Frankfurt. 468 00:21:06,760 --> 00:21:10,260 And ended up with Michael Stryker at UCSF 469 00:21:10,260 --> 00:21:12,940 to really get to work on critical period mechanisms. 470 00:21:12,940 --> 00:21:15,580 Wow, you've really collected the institutions 471 00:21:15,580 --> 00:21:19,460 and some extraordinarily prominent mentors. 472 00:21:19,460 --> 00:21:20,700 I've been very fortunate, yes. 473 00:21:20,700 --> 00:21:21,540 Amazing, amazing. 474 00:21:21,540 --> 00:21:25,660 And going back to Tokyo, was that a specific choice 475 00:21:25,660 --> 00:21:29,340 based on being of Japanese origin and that? 476 00:21:29,340 --> 00:21:32,140 Oh yes, after college and having grown up in the States, 477 00:21:32,140 --> 00:21:34,220 I really wanted to follow my roots 478 00:21:34,220 --> 00:21:37,020 and spend some time in Japan and Germany. 479 00:21:37,020 --> 00:21:38,560 Outstanding, outstanding. 480 00:21:38,560 --> 00:21:40,620 We wouldn't be complete here without getting back 481 00:21:40,620 --> 00:21:42,660 to your original motivation, 482 00:21:42,660 --> 00:21:44,860 which was around computer science and AI. 483 00:21:44,860 --> 00:21:47,100 And this has come full circle now, right? 484 00:21:47,100 --> 00:21:48,860 With the interface, I mean, 485 00:21:48,860 --> 00:21:51,720 we're in the AI revolution at the moment. 486 00:21:51,720 --> 00:21:53,060 Tell us a little bit about your thoughts 487 00:21:53,060 --> 00:21:55,420 about the role of AI in the neurosciences 488 00:21:55,420 --> 00:21:57,780 as we speak, as we're sitting here. 489 00:21:57,780 --> 00:21:59,140 Yes, isn't this fascinating? 490 00:21:59,140 --> 00:22:02,380 So it really has come full circle, 491 00:22:02,380 --> 00:22:04,500 but now we have several decades 492 00:22:04,500 --> 00:22:08,140 of neurobiological understanding behind it as well. 493 00:22:08,140 --> 00:22:12,700 The current excitement about AI is loosely modeled 494 00:22:12,700 --> 00:22:16,080 on brain structure in terms of deep neural networks. 495 00:22:16,080 --> 00:22:18,900 And with the brute force power of computing 496 00:22:18,900 --> 00:22:21,420 that's available, amazing things are happening 497 00:22:21,420 --> 00:22:25,740 and everyone is familiar with GPTs and so forth. 498 00:22:25,740 --> 00:22:27,980 But I think the question still remains, 499 00:22:27,980 --> 00:22:30,860 why is it seemingly effortless for a child 500 00:22:30,860 --> 00:22:34,720 with very little exposure to learn how to speak one 501 00:22:34,720 --> 00:22:39,260 or multiple languages or learn a variety of skills 502 00:22:39,260 --> 00:22:44,260 in ways that are not yet possible with current AI? 503 00:22:44,340 --> 00:22:48,000 And so the motivation behind our work now 504 00:22:48,000 --> 00:22:51,260 is to understand principles of brain development, 505 00:22:51,260 --> 00:22:55,420 which might bring the AI a little closer 506 00:22:55,420 --> 00:22:58,560 to the way humans acquire their knowledge 507 00:22:58,560 --> 00:23:00,200 or intelligent behaviors. 508 00:23:01,420 --> 00:23:04,660 But at the same time, honoring the fact 509 00:23:04,660 --> 00:23:08,700 that the world has changed and that humans will need 510 00:23:08,700 --> 00:23:13,700 to coexist with the AI and rely on it in ways 511 00:23:14,020 --> 00:23:16,960 that were not possible even a year and a half ago. 512 00:23:16,960 --> 00:23:20,780 So the alignment problem of human 513 00:23:20,780 --> 00:23:23,580 and artificial intelligence and having AIs 514 00:23:23,580 --> 00:23:25,880 that understand the motivations of humans 515 00:23:25,880 --> 00:23:29,660 when they're interacting with them is extremely important. 516 00:23:29,660 --> 00:23:30,740 Right, right. 517 00:23:30,740 --> 00:23:34,700 And on that front, I mean, do you think that AI, 518 00:23:34,700 --> 00:23:39,240 you know, watching AIs develop human-like skills 519 00:23:39,240 --> 00:23:42,440 is a pathway to understanding disease, for example? 520 00:23:42,440 --> 00:23:43,780 Is that a? 521 00:23:43,780 --> 00:23:44,780 Yes, I think so. 522 00:23:44,780 --> 00:23:47,540 So from a conceptual point of view, 523 00:23:47,540 --> 00:23:51,100 intelligent behavior can exist in a vacuum or a void 524 00:23:51,100 --> 00:23:54,380 like AI might in the virtual space. 525 00:23:54,380 --> 00:23:57,940 But at the same time, it's a good reminder 526 00:23:57,940 --> 00:24:00,580 that the brain exists in a body 527 00:24:00,580 --> 00:24:05,220 and it's bringing our kind of reductionist approach back out 528 00:24:05,220 --> 00:24:09,200 to consider the brain as a homeostatic organ 529 00:24:09,200 --> 00:24:13,440 within an individual whose job is to move 530 00:24:13,440 --> 00:24:16,380 and adapt to environments and things that AI 531 00:24:16,380 --> 00:24:18,620 doesn't necessarily need to worry about. 532 00:24:18,620 --> 00:24:20,900 And so our brain has evolved to deal 533 00:24:20,900 --> 00:24:23,340 with a particular set of constraints, 534 00:24:23,340 --> 00:24:27,180 namely existing within a body, which AI doesn't. 535 00:24:27,180 --> 00:24:31,140 And this might be one major difference in the way 536 00:24:31,140 --> 00:24:35,660 these two forms of intelligence are being developed. 537 00:24:35,660 --> 00:24:36,260 Right. 538 00:24:36,260 --> 00:24:38,580 But we have the ability to embody an AI, right, 539 00:24:38,580 --> 00:24:41,260 in robotics and so on. 540 00:24:41,260 --> 00:24:42,340 That's right. 541 00:24:42,340 --> 00:24:45,980 So that's part of the work we do with some of our colleagues 542 00:24:45,980 --> 00:24:49,660 in Japan, developmental robotics, as it's called, 543 00:24:49,660 --> 00:24:53,660 to have actual physical robots that are then 544 00:24:53,660 --> 00:24:57,500 programmed with models of the developing brain 545 00:24:57,500 --> 00:25:00,540 to test our understanding of how brain development works, 546 00:25:00,540 --> 00:25:02,900 but also as tools for interacting 547 00:25:02,900 --> 00:25:06,620 with autistic children, for example, who often 548 00:25:06,620 --> 00:25:11,940 might prefer to interact with a robot rather than another human. 549 00:25:11,940 --> 00:25:12,700 Embodied human. 550 00:25:12,700 --> 00:25:13,540 Yeah. 551 00:25:13,540 --> 00:25:18,460 It's absolutely fascinating. 552 00:25:18,460 --> 00:25:22,140 When people worry about the sort of apocalyptic things, 553 00:25:22,140 --> 00:25:27,460 do you have any worries about that, AI jumping the shark? 554 00:25:27,460 --> 00:25:29,820 Of course, it is a concern. 555 00:25:29,820 --> 00:25:37,300 And ethics and the proper use of AI and how it's advanced 556 00:25:37,300 --> 00:25:41,100 needs to really catch up quickly now to make sure 557 00:25:41,100 --> 00:25:43,100 that guidelines are in place. 558 00:25:43,100 --> 00:25:44,020 That's for sure. 559 00:25:44,020 --> 00:25:45,820 So it's in our own control. 560 00:25:45,820 --> 00:25:46,500 Right. 561 00:25:46,500 --> 00:25:49,300 You've had an extraordinary career. 562 00:25:49,300 --> 00:25:50,620 I mean, you really, really have. 563 00:25:50,620 --> 00:25:55,460 And looking back on the things that happened 564 00:25:55,460 --> 00:25:59,060 and the people that you met, has it given you 565 00:25:59,060 --> 00:26:04,100 a sense for the path to success? 566 00:26:04,100 --> 00:26:06,460 And if you were to take that and turn it 567 00:26:06,460 --> 00:26:08,940 into advice for a youngster today, 568 00:26:08,940 --> 00:26:10,740 you have some pearls of wisdom. 569 00:26:10,740 --> 00:26:13,380 We always ask this question. 570 00:26:13,380 --> 00:26:13,860 Yes. 571 00:26:13,860 --> 00:26:17,140 Well, the world is changing dramatically. 572 00:26:17,140 --> 00:26:21,940 So I feel like I'm a bit dated already. 573 00:26:21,940 --> 00:26:26,540 Serendipity has been a big part of this. 574 00:26:26,540 --> 00:26:31,540 I started my lab out of graduate school from UCSF, 575 00:26:31,540 --> 00:26:33,300 went back to Japan. 576 00:26:33,300 --> 00:26:36,460 It was not something that I had imagined doing. 577 00:26:36,460 --> 00:26:41,540 But they were launching a new institute, the Riken Brain 578 00:26:41,540 --> 00:26:43,020 Science Institute. 579 00:26:43,020 --> 00:26:47,700 And the goal was to create something more Western, 580 00:26:47,700 --> 00:26:50,460 not the traditional hierarchical system 581 00:26:50,460 --> 00:26:52,260 of Japanese university. 582 00:26:52,260 --> 00:26:54,820 And I thought, well, who am I? 583 00:26:54,820 --> 00:26:56,820 I'm just starting out. 584 00:26:56,820 --> 00:27:00,300 And I had a lot of interesting ideas, I thought, 585 00:27:00,300 --> 00:27:05,220 for the research, but was just untested. 586 00:27:05,220 --> 00:27:08,780 But I felt that I could bring the kind 587 00:27:08,780 --> 00:27:13,300 of Western infrastructure or ecosystem to a new institution. 588 00:27:13,300 --> 00:27:17,340 And in that way, felt like I could contribute right away 589 00:27:17,340 --> 00:27:21,900 in addition to eventually science. 590 00:27:21,900 --> 00:27:25,580 That was a very, very lucky break. 591 00:27:25,580 --> 00:27:29,420 I was, of course, anxious about doing that, 592 00:27:29,420 --> 00:27:31,860 but excited at the same time. 593 00:27:31,860 --> 00:27:34,660 It was an institution that had no tenure. 594 00:27:34,660 --> 00:27:37,300 It was on a five-year review cycle. 595 00:27:37,300 --> 00:27:41,260 And the challenge was there, also moving 596 00:27:41,260 --> 00:27:44,620 far from the familiar. 597 00:27:44,620 --> 00:27:49,100 But chances come around very rarely. 598 00:27:49,100 --> 00:27:51,420 And so I think my advice would be, 599 00:27:51,420 --> 00:27:56,020 if a young person had an opportunity, 600 00:27:56,020 --> 00:27:58,460 they shouldn't be afraid. 601 00:27:58,460 --> 00:28:02,380 And they should seize the day. 602 00:28:02,380 --> 00:28:04,820 You know, you said something very interesting there 603 00:28:04,820 --> 00:28:07,380 about the business of bringing sort of a Western model, 604 00:28:07,380 --> 00:28:11,100 I assume, like an American model to science. 605 00:28:11,100 --> 00:28:12,820 Is there something about that that 606 00:28:12,820 --> 00:28:16,740 makes it intrinsically better at getting the work done, 607 00:28:16,740 --> 00:28:18,180 do you think? 608 00:28:18,180 --> 00:28:23,340 Better in the sense that young people are brimming with ideas. 609 00:28:23,340 --> 00:28:27,740 And in a hierarchical system like the traditional Japanese 610 00:28:27,740 --> 00:28:29,980 university system. 611 00:28:29,980 --> 00:28:30,940 And much of Europe. 612 00:28:30,940 --> 00:28:31,900 And much of Europe. 613 00:28:31,900 --> 00:28:34,860 Gaining the independence to follow your own ideas 614 00:28:34,860 --> 00:28:40,660 takes years of patience and moving up the hierarchy. 615 00:28:40,660 --> 00:28:44,420 The US system empowers assistant professors 616 00:28:44,420 --> 00:28:47,140 to be independent with all the risk that comes along 617 00:28:47,140 --> 00:28:48,260 with that, of course. 618 00:28:48,260 --> 00:28:51,860 But also the independence to follow your dream. 619 00:28:51,860 --> 00:28:54,140 So stripping away the politics a little bit. 620 00:28:54,140 --> 00:28:57,660 And maybe is there a more business model to it, 621 00:28:57,660 --> 00:28:58,140 do you think? 622 00:28:58,140 --> 00:29:00,620 Is that part of it? 623 00:29:00,620 --> 00:29:03,460 A business model to the American way 624 00:29:03,460 --> 00:29:07,100 is very successful for advancing science. 625 00:29:07,100 --> 00:29:10,220 The freshest, brightest minds are given the opportunity 626 00:29:10,220 --> 00:29:10,740 right away. 627 00:29:13,780 --> 00:29:16,860 What's more difficult is the long vision. 628 00:29:16,860 --> 00:29:19,580 So I've noticed in Europe and certainly in Japan, 629 00:29:19,580 --> 00:29:23,180 grants are designed around long term vision. 630 00:29:23,180 --> 00:29:26,700 And if you're studying something like Alzheimer's disease, 631 00:29:26,700 --> 00:29:29,940 which takes years to manifest, it's 632 00:29:29,940 --> 00:29:35,940 very hard to imagine doing a holistic study with short three 633 00:29:35,940 --> 00:29:37,700 five year grant cycles. 634 00:29:37,700 --> 00:29:38,220 Very good. 635 00:29:38,220 --> 00:29:39,940 Yeah, absolutely. 636 00:29:39,940 --> 00:29:43,660 So short, sharp kind of bolus of money for a youngster 637 00:29:43,660 --> 00:29:44,780 to get some ideas going. 638 00:29:44,780 --> 00:29:47,580 But the bigger the wicked problems 639 00:29:47,580 --> 00:29:51,140 are a little less tangible with that model. 640 00:29:51,140 --> 00:29:52,660 Well, that's really, really interesting. 641 00:29:52,660 --> 00:29:55,020 And thank you for those exceptional insights. 642 00:29:55,020 --> 00:29:57,300 Takau, it's an absolute pleasure to have you here. 643 00:29:57,300 --> 00:30:00,500 And thanks for joining us on Neuroscience Perspectives. 644 00:30:00,500 --> 00:30:01,700 Thank you for having me. 645 00:30:01,700 --> 00:30:02,700 Thank you. 646 00:30:02,700 --> 00:30:27,660 Thank you.