TEDxCaltech
Michael Dickinson: How a fly flies
Filmed:
Readability: 4.2
1,787,704 views
An insect's ability to fly is one of the greatest feats of evolution. Michael Dickinson looks at how a fruit fly takes flight with such delicate wings, thanks to a clever flapping motion and flight muscles that are both powerful and nimble. But the secret ingredient: the incredible fly brain.
Michael Dickinson - Biologist
Most people are irritated by the buzzing of a fly's wings. But biologist Michael Dickinson views the sound with a deep sense of wonder. Full bio
Most people are irritated by the buzzing of a fly's wings. But biologist Michael Dickinson views the sound with a deep sense of wonder. Full bio
Double-click the English transcript below to play the video.
00:16
I grew up watching Star Trek. I love Star Trek.
0
545
3532
00:19
Star Trek made me want to see alien creatures,
1
4077
4462
00:24
creatures from a far-distant world.
2
8539
2303
00:26
But basically, I figured out that I could find
3
10842
2787
00:29
those alien creatures right on Earth.
4
13629
2977
00:32
And what I do is I study insects.
5
16606
2653
00:35
I'm obsessed with insects, particularly insect flight.
6
19259
3256
00:38
I think the evolution of insect flight is perhaps
7
22515
3141
00:41
one of the most important events in the history of life.
8
25656
2742
00:44
Without insects, there'd be no flowering plants.
9
28398
2237
00:46
Without flowering plants, there would be no
10
30635
1916
00:48
clever, fruit-eating primates giving TED Talks.
11
32551
3137
00:51
(Laughter)
12
35688
2300
00:53
Now,
13
37988
1987
00:55
David and Hidehiko and Ketaki
14
39975
3039
00:58
gave a very compelling story about
15
43014
3445
01:02
the similarities between fruit flies and humans,
16
46459
2805
01:05
and there are many similarities,
17
49264
1489
01:06
and so you might think that if humans are similar to fruit flies,
18
50753
3002
01:09
the favorite behavior of a fruit fly might be this, for example --
19
53755
3797
01:13
(Laughter)
20
57552
2282
01:15
but in my talk, I don't want to emphasize on the similarities
21
59834
3191
01:18
between humans and fruit flies, but rather the differences,
22
63025
3067
01:21
and focus on the behaviors that I think fruit flies excel at doing.
23
66092
5287
01:27
And so I want to show you a high-speed video sequence
24
71379
2856
01:30
of a fly shot at 7,000 frames per second in infrared lighting,
25
74235
3935
01:34
and to the right, off-screen, is an electronic looming predator
26
78170
4210
01:38
that is going to go at the fly.
27
82380
1435
01:39
The fly is going to sense this predator.
28
83815
1838
01:41
It is going to extend its legs out.
29
85653
2455
01:44
It's going to sashay away
30
88108
1613
01:45
to live to fly another day.
31
89721
2565
01:48
Now I have carefully cropped this sequence
32
92286
2362
01:50
to be exactly the duration of a human eye blink,
33
94648
3160
01:53
so in the time that it would take you to blink your eye,
34
97808
2834
01:56
the fly has seen this looming predator,
35
100642
3265
01:59
estimated its position, initiated a motor pattern to fly it away,
36
103907
6168
02:05
beating its wings at 220 times a second as it does so.
37
110075
4464
02:10
I think this is a fascinating behavior
38
114539
1973
02:12
that shows how fast the fly's brain can process information.
39
116512
3921
02:16
Now, flight -- what does it take to fly?
40
120433
2842
02:19
Well, in order to fly, just as in a human aircraft,
41
123275
2864
02:22
you need wings that can generate sufficient aerodynamic forces,
42
126139
2735
02:24
you need an engine sufficient to generate the power required for flight,
43
128874
3546
02:28
and you need a controller,
44
132420
1709
02:30
and in the first human aircraft, the controller was basically
45
134129
2626
02:32
the brain of Orville and Wilbur sitting in the cockpit.
46
136755
4312
02:36
Now, how does this compare to a fly?
47
141067
2753
02:39
Well, I spent a lot of my early career trying to figure out
48
143820
3251
02:42
how insect wings generate enough force to keep the flies in the air.
49
147071
4336
02:47
And you might have heard how engineers proved
50
151407
1610
02:48
that bumblebees couldn't fly.
51
153017
2634
02:51
Well, the problem was in thinking that the insect wings
52
155651
2620
02:54
function in the way that aircraft wings work. But they don't.
53
158271
3119
02:57
And we tackle this problem by building giant,
54
161390
2854
03:00
dynamically scaled model robot insects
55
164244
3432
03:03
that would flap in giant pools of mineral oil
56
167676
3336
03:06
where we could study the aerodynamic forces.
57
171012
2274
03:09
And it turns out that the insects flap their wings
58
173286
2158
03:11
in a very clever way, at a very high angle of attack
59
175444
2592
03:13
that creates a structure at the leading edge of the wing,
60
178036
3121
03:17
a little tornado-like structure called a leading edge vortex,
61
181157
3199
03:20
and it's that vortex that actually enables the wings
62
184356
2954
03:23
to make enough force for the animal to stay in the air.
63
187310
3359
03:26
But the thing that's actually most -- so, what's fascinating
64
190669
2428
03:28
is not so much that the wing has some interesting morphology.
65
193097
2975
03:31
What's clever is the way the fly flaps it,
66
196072
3645
03:35
which of course ultimately is controlled by the nervous system,
67
199717
3136
03:38
and this is what enables flies to perform
68
202853
2647
03:41
these remarkable aerial maneuvers.
69
205500
2807
03:44
Now, what about the engine?
70
208307
2097
03:46
The engine of the fly is absolutely fascinating.
71
210404
2492
03:48
They have two types of flight muscle:
72
212896
1898
03:50
so-called power muscle, which is stretch-activated,
73
214794
2985
03:53
which means that it activates itself and does not need to be controlled
74
217779
3726
03:57
on a contraction-by-contraction basis by the nervous system.
75
221505
3339
04:00
It's specialized to generate the enormous power required for flight,
76
224844
4609
04:05
and it fills the middle portion of the fly,
77
229453
2079
04:07
so when a fly hits your windshield,
78
231532
1547
04:08
it's basically the power muscle that you're looking at.
79
233079
2406
04:11
But attached to the base of the wing
80
235485
2146
04:13
is a set of little, tiny control muscles
81
237631
2638
04:16
that are not very powerful at all, but they're very fast,
82
240269
3301
04:19
and they're able to reconfigure the hinge of the wing
83
243570
3206
04:22
on a stroke-by-stroke basis,
84
246776
1762
04:24
and this is what enables the fly to change its wing
85
248538
3142
04:27
and generate the changes in aerodynamic forces
86
251680
2971
04:30
which change its flight trajectory.
87
254651
2573
04:33
And of course, the role of the nervous system is to control all this.
88
257224
3563
04:36
So let's look at the controller.
89
260787
1512
04:38
Now flies excel in the sorts of sensors
90
262299
2647
04:40
that they carry to this problem.
91
264946
2284
04:43
They have antennae that sense odors and detect wind detection.
92
267230
4127
04:47
They have a sophisticated eye which is
93
271357
1675
04:48
the fastest visual system on the planet.
94
273032
2456
04:51
They have another set of eyes on the top of their head.
95
275488
2036
04:53
We have no idea what they do.
96
277524
2052
04:55
They have sensors on their wing.
97
279576
2954
04:58
Their wing is covered with sensors, including sensors
98
282530
3760
05:02
that sense deformation of the wing.
99
286290
2046
05:04
They can even taste with their wings.
100
288336
2109
05:06
One of the most sophisticated sensors a fly has
101
290445
2555
05:08
is a structure called the halteres.
102
293000
1807
05:10
The halteres are actually gyroscopes.
103
294807
1879
05:12
These devices beat back and forth about 200 hertz during flight,
104
296686
4449
05:17
and the animal can use them to sense its body rotation
105
301135
2673
05:19
and initiate very, very fast corrective maneuvers.
106
303808
3968
05:23
But all of this sensory information has to be processed
107
307776
2329
05:26
by a brain, and yes, indeed, flies have a brain,
108
310105
3720
05:29
a brain of about 100,000 neurons.
109
313825
3159
05:32
Now several people at this conference
110
316984
2193
05:35
have already suggested that fruit flies could serve neuroscience
111
319177
4808
05:39
because they're a simple model of brain function.
112
323985
3247
05:43
And the basic punchline of my talk is,
113
327232
2077
05:45
I'd like to turn that over on its head.
114
329309
2658
05:47
I don't think they're a simple model of anything.
115
331967
2628
05:50
And I think that flies are a great model.
116
334595
2477
05:52
They're a great model for flies.
117
337072
2516
05:55
(Laughter)
118
339588
2481
05:57
And let's explore this notion of simplicity.
119
342069
3003
06:00
So I think, unfortunately, a lot of neuroscientists,
120
345072
2431
06:03
we're all somewhat narcissistic.
121
347503
1832
06:05
When we think of brain, we of course imagine our own brain.
122
349335
3433
06:08
But remember that this kind of brain,
123
352768
1960
06:10
which is much, much smaller
124
354728
1768
06:12
— instead of 100 billion neurons, it has 100,000 neurons —
125
356496
2678
06:15
but this is the most common form of brain on the planet
126
359174
2882
06:17
and has been for 400 million years.
127
362056
2904
06:20
And is it fair to say that it's simple?
128
364960
2288
06:23
Well, it's simple in the sense that it has fewer neurons,
129
367248
2095
06:25
but is that a fair metric?
130
369343
1754
06:26
And I would propose it's not a fair metric.
131
371097
2276
06:29
So let's sort of think about this. I think we have to compare --
132
373373
3100
06:32
(Laughter) —
133
376473
1559
06:33
we have to compare the size of the brain
134
378032
5121
06:39
with what the brain can do.
135
383153
2030
06:41
So I propose we have a Trump number,
136
385183
2881
06:43
and the Trump number is the ratio of this man's
137
388064
2865
06:46
behavioral repertoire to the number of neurons in his brain.
138
390929
3679
06:50
We'll calculate the Trump number for the fruit fly.
139
394608
2668
06:53
Now, how many people here think the Trump number
140
397276
2684
06:55
is higher for the fruit fly?
141
399960
2489
06:58
(Applause)
142
402449
2431
07:00
It's a very smart, smart audience.
143
404880
3428
07:04
Yes, the inequality goes in this direction, or I would posit it.
144
408308
3327
07:07
Now I realize that it is a little bit absurd
145
411635
2382
07:09
to compare the behavioral repertoire of a human to a fly.
146
414017
3558
07:13
But let's take another animal just as an example. Here's a mouse.
147
417575
4143
07:17
A mouse has about 1,000 times as many neurons as a fly.
148
421718
4305
07:21
I used to study mice. When I studied mice,
149
426023
2027
07:23
I used to talk really slowly.
150
428050
2837
07:26
And then something happened when I started to work on flies.
151
430887
2576
07:29
(Laughter)
152
433463
2412
07:31
And I think if you compare the natural history of flies and mice,
153
435875
3460
07:35
it's really comparable. They have to forage for food.
154
439335
3313
07:38
They have to engage in courtship.
155
442648
2447
07:40
They have sex. They hide from predators.
156
445095
3471
07:44
They do a lot of the similar things.
157
448566
1980
07:46
But I would argue that flies do more.
158
450546
1718
07:48
So for example, I'm going to show you a sequence,
159
452264
3378
07:51
and I have to say, some of my funding comes from the military,
160
455642
4205
07:55
so I'm showing this classified sequence
161
459847
2072
07:57
and you cannot discuss it outside of this room. Okay?
162
461919
4093
08:01
So I want you to look at the payload
163
466012
1908
08:03
at the tail of the fruit fly.
164
467920
3026
08:06
Watch it very closely,
165
470946
2101
08:08
and you'll see why my six-year-old son
166
473047
4297
08:13
now wants to be a neuroscientist.
167
477344
4729
08:17
Wait for it.
168
482073
1179
08:19
Pshhew.
169
483252
1569
08:20
So at least you'll admit that if fruit flies are not as clever as mice,
170
484821
3084
08:23
they're at least as clever as pigeons. (Laughter)
171
487905
4916
08:28
Now, I want to get across that it's not just a matter of numbers
172
492821
3967
08:32
but also the challenge for a fly to compute
173
496788
2598
08:35
everything its brain has to compute with such tiny neurons.
174
499386
2849
08:38
So this is a beautiful image of a visual interneuron from a mouse
175
502235
2988
08:41
that came from Jeff Lichtman's lab,
176
505223
2768
08:43
and you can see the wonderful images of brains
177
507991
3247
08:47
that he showed in his talk.
178
511238
3193
08:50
But up in the corner, in the right corner, you'll see,
179
514431
2368
08:52
at the same scale, a visual interneuron from a fly.
180
516799
4112
08:56
And I'll expand this up.
181
520911
1841
08:58
And it's a beautifully complex neuron.
182
522752
2170
09:00
It's just very, very tiny, and there's lots of biophysical challenges
183
524922
3485
09:04
with trying to compute information with tiny, tiny neurons.
184
528407
3623
09:07
How small can neurons get? Well, look at this interesting insect.
185
532030
3537
09:11
It looks sort of like a fly. It has wings, it has eyes,
186
535567
2212
09:13
it has antennae, its legs, complicated life history,
187
537779
2799
09:16
it's a parasite, it has to fly around and find caterpillars
188
540578
3096
09:19
to parasatize,
189
543674
1382
09:20
but not only is its brain the size of a salt grain,
190
545056
4115
09:25
which is comparable for a fruit fly,
191
549171
1969
09:27
it is the size of a salt grain.
192
551140
2926
09:29
So here's some other organisms at the similar scale.
193
554066
3635
09:33
This animal is the size of a paramecium and an amoeba,
194
557701
4130
09:37
and it has a brain of 7,000 neurons that's so small --
195
561831
3880
09:41
you know these things called cell bodies you've been hearing about,
196
565711
2456
09:44
where the nucleus of the neuron is?
197
568167
1651
09:45
This animal gets rid of them because they take up too much space.
198
569818
3460
09:49
So this is a session on frontiers in neuroscience.
199
573278
2473
09:51
I would posit that one frontier in neuroscience is to figure out how the brain of that thing works.
200
575751
5360
09:57
But let's think about this. How can you make a small number of neurons do a lot?
201
581111
5633
10:02
And I think, from an engineering perspective,
202
586744
2522
10:05
you think of multiplexing.
203
589266
1729
10:06
You can take a hardware and have that hardware
204
590995
2703
10:09
do different things at different times,
205
593698
1613
10:11
or have different parts of the hardware doing different things.
206
595311
2995
10:14
And these are the two concepts I'd like to explore.
207
598306
3271
10:17
And they're not concepts that I've come up with,
208
601577
1658
10:19
but concepts that have been proposed by others in the past.
209
603235
4545
10:23
And one idea comes from lessons from chewing crabs.
210
607780
3075
10:26
And I don't mean chewing the crabs.
211
610855
1867
10:28
I grew up in Baltimore, and I chew crabs very, very well.
212
612722
3599
10:32
But I'm talking about the crabs actually doing the chewing.
213
616321
2857
10:35
Crab chewing is actually really fascinating.
214
619178
2030
10:37
Crabs have this complicated structure under their carapace
215
621208
3259
10:40
called the gastric mill
216
624467
1310
10:41
that grinds their food in a variety of different ways.
217
625777
2430
10:44
And here's an endoscopic movie of this structure.
218
628207
5259
10:49
The amazing thing about this is that it's controlled
219
633466
2560
10:51
by a really tiny set of neurons, about two dozen neurons
220
636026
3432
10:55
that can produce a vast variety of different motor patterns,
221
639458
4963
11:00
and the reason it can do this is that this little tiny ganglion
222
644421
4347
11:04
in the crab is actually inundated by many, many neuromodulators.
223
648768
4184
11:08
You heard about neuromodulators earlier.
224
652952
2141
11:10
There are more neuromodulators
225
655093
2225
11:13
that alter, that innervate this structure than actually neurons in the structure,
226
657318
5485
11:18
and they're able to generate a complicated set of patterns.
227
662803
4242
11:22
And this is the work by Eve Marder and her many colleagues
228
667045
3441
11:26
who've been studying this fascinating system
229
670486
2295
11:28
that show how a smaller cluster of neurons
230
672781
2152
11:30
can do many, many, many things
231
674933
1825
11:32
because of neuromodulation that can take place on a moment-by-moment basis.
232
676758
4856
11:37
So this is basically multiplexing in time.
233
681614
2439
11:39
Imagine a network of neurons with one neuromodulator.
234
684053
2785
11:42
You select one set of cells to perform one sort of behavior,
235
686838
3478
11:46
another neuromodulator, another set of cells,
236
690316
2618
11:48
a different pattern, and you can imagine
237
692934
1713
11:50
you could extrapolate to a very, very complicated system.
238
694647
3878
11:54
Is there any evidence that flies do this?
239
698525
2094
11:56
Well, for many years in my laboratory and other laboratories around the world,
240
700619
3375
11:59
we've been studying fly behaviors in little flight simulators.
241
703994
2648
12:02
You can tether a fly to a little stick.
242
706642
1706
12:04
You can measure the aerodynamic forces it's creating.
243
708348
2501
12:06
You can let the fly play a little video game
244
710849
2546
12:09
by letting it fly around in a visual display.
245
713395
3878
12:13
So let me show you a little tiny sequence of this.
246
717273
2337
12:15
Here's a fly
247
719610
1227
12:16
and a large infrared view of the fly in the flight simulator,
248
720837
3437
12:20
and this is a game the flies love to play.
249
724274
1955
12:22
You allow them to steer towards the little stripe,
250
726229
2437
12:24
and they'll just steer towards that stripe forever.
251
728666
2825
12:27
It's part of their visual guidance system.
252
731491
3558
12:30
But very, very recently, it's been possible
253
735049
2345
12:33
to modify these sorts of behavioral arenas for physiologies.
254
737394
4940
12:38
So this is the preparation that one of my former post-docs,
255
742334
2488
12:40
Gaby Maimon, who's now at Rockefeller, developed,
256
744822
2443
12:43
and it's basically a flight simulator
257
747265
1686
12:44
but under conditions where you actually can stick an electrode
258
748951
3075
12:47
in the brain of the fly and record
259
752026
2264
12:50
from a genetically identified neuron in the fly's brain.
260
754290
3656
12:53
And this is what one of these experiments looks like.
261
757946
2298
12:56
It was a sequence taken from another post-doc in the lab,
262
760244
2971
12:59
Bettina Schnell.
263
763215
1199
13:00
The green trace at the bottom is the membrane potential
264
764414
3392
13:03
of a neuron in the fly's brain,
265
767806
2030
13:05
and you'll see the fly start to fly, and the fly is actually
266
769836
2942
13:08
controlling the rotation of that visual pattern itself
267
772778
3279
13:11
by its own wing motion,
268
776057
1479
13:13
and you can see this visual interneuron
269
777536
2110
13:15
respond to the pattern of wing motion as the fly flies.
270
779646
3908
13:19
So for the first time we've actually been able to record
271
783554
2376
13:21
from neurons in the fly's brain while the fly
272
785930
2908
13:24
is performing sophisticated behaviors such as flight.
273
788838
4468
13:29
And one of the lessons we've been learning
274
793306
1855
13:31
is that the physiology of cells that we've been studying
275
795161
2420
13:33
for many years in quiescent flies
276
797581
2421
13:35
is not the same as the physiology of those cells
277
800002
2648
13:38
when the flies actually engage in active behaviors
278
802650
2736
13:41
like flying and walking and so forth.
279
805386
2539
13:43
And why is the physiology different?
280
807925
2925
13:46
Well it turns out it's these neuromodulators,
281
810850
2057
13:48
just like the neuromodulators in that little tiny ganglion in the crabs.
282
812907
3951
13:52
So here's a picture of the octopamine system.
283
816858
2550
13:55
Octopamine is a neuromodulator
284
819408
1754
13:57
that seems to play an important role in flight and other behaviors.
285
821162
4336
14:01
But this is just one of many neuromodulators
286
825498
2472
14:03
that's in the fly's brain.
287
827970
1071
14:04
So I really think that, as we learn more,
288
829041
2666
14:07
it's going to turn out that the whole fly brain
289
831707
2527
14:10
is just like a large version of this stomatogastric ganglion,
290
834234
3089
14:13
and that's one of the reasons why it can do so much with so few neurons.
291
837323
4360
14:17
Now, another idea, another way of multiplexing
292
841683
2787
14:20
is multiplexing in space,
293
844470
1656
14:22
having different parts of a neuron
294
846126
1694
14:23
do different things at the same time.
295
847820
2122
14:25
So here's two sort of canonical neurons
296
849942
1833
14:27
from a vertebrate and an invertebrate,
297
851775
2285
14:29
a human pyramidal neuron from Ramon y Cajal,
298
854060
3250
14:33
and another cell to the right, a non-spiking interneuron,
299
857310
4003
14:37
and this is the work of Alan Watson and Malcolm Burrows many years ago,
300
861313
4147
14:41
and Malcolm Burrows came up with a pretty interesting idea
301
865460
3075
14:44
based on the fact that this neuron from a locust
302
868535
2882
14:47
does not fire action potentials.
303
871417
1959
14:49
It's a non-spiking cell.
304
873376
1748
14:51
So a typical cell, like the neurons in our brain,
305
875124
2780
14:53
has a region called the dendrites that receives input,
306
877904
2752
14:56
and that input sums together
307
880656
2589
14:59
and will produce action potentials
308
883245
2296
15:01
that run down the axon and then activate
309
885541
2331
15:03
all the output regions of the neuron.
310
887872
2296
15:06
But non-spiking neurons are actually quite complicated
311
890168
2876
15:08
because they can have input synapses and output synapses
312
893044
3112
15:12
all interdigitated, and there's no single action potential
313
896156
3663
15:15
that drives all the outputs at the same time.
314
899819
3126
15:18
So there's a possibility that you have computational compartments
315
902945
3907
15:22
that allow the different parts of the neuron
316
906852
3978
15:26
to do different things at the same time.
317
910830
2560
15:29
So these basic concepts of multitasking in time
318
913390
4671
15:33
and multitasking in space,
319
918061
2361
15:36
I think these are things that are true in our brains as well,
320
920422
2832
15:39
but I think the insects are the true masters of this.
321
923254
2577
15:41
So I hope you think of insects a little bit differently next time,
322
925831
3116
15:44
and as I say up here, please think before you swat.
323
928947
2935
15:47
(Applause)
324
931882
2953
ABOUT THE SPEAKER
Michael Dickinson - BiologistMost people are irritated by the buzzing of a fly's wings. But biologist Michael Dickinson views the sound with a deep sense of wonder.
Why you should listen
Some things are so commonplace that they barely register our attention. Michael Dickinson has dedicated much of his research to one such thing -- the flight of the fly. Dickinson aims to understand how a fly's nervous system allows it to accomplish such incredible aerodynamic feats. Affectionately dubbed the "Fly Guy" by The Scientist, Dickinson's research brings together zoology, neuroscience and fluid mechanics.
Dickinson was named a MacArthur Fellow in 2001. He is now a professor of biology at the University of Washington, where he heads The Dickinson Lab. The lab conducts research into insect flight control, animal brain recordings, animal/robot interactions and animal visual navigation and welcomes students with an interest in studying insect flight, behavior and evolution from an interdisciplinary approach perspective.
Michael Dickinson | Speaker | TED.com