This computer is learning to read your mind | DIY Neuroscience, a TED series

126,401 views ・ 2018-09-15

TED


Please double-click on the English subtitles below to play the video.

00:00
Translator: Joseph Geni Reviewer: Krystian Aparta
0
0
7000
00:12
Greg Gage: Mind-reading. You've seen this in sci-fi movies:
1
12203
2859
00:15
machines that can read our thoughts.
2
15086
1857
00:16
However, there are devices today
3
16967
1798
00:18
that can read the electrical activity from our brains.
4
18789
2524
00:21
We call this the EEG.
5
21337
1272
00:23
Is there information contained in these brainwaves?
6
23695
2829
00:26
And if so, could we train a computer to read our thoughts?
7
26548
2813
00:29
My buddy Nathan has been working to hack the EEG
8
29385
2904
00:32
to build a mind-reading machine.
9
32313
1676
00:34
[DIY Neuroscience]
10
34013
2457
00:36
So this is how the EEG works.
11
36939
1561
00:38
Inside your head is a brain,
12
38524
1844
00:40
and that brain is made out of billions of neurons.
13
40392
2558
00:42
Each of those neurons sends an electrical message to each other.
14
42974
3007
00:46
These small messages can combine to make an electrical wave
15
46005
2814
00:48
that we can detect on a monitor.
16
48843
1570
00:50
Now traditionally, the EEG can tell us large-scale things,
17
50437
2724
00:53
for example if you're asleep or if you're alert.
18
53185
2350
00:55
But can it tell us anything else?
19
55559
1594
00:57
Can it actually read our thoughts?
20
57177
1702
00:58
We're going to test this,
21
58903
1219
01:00
and we're not going to start with some complex thoughts.
22
60146
2623
01:02
We're going to do something very simple.
23
62793
1977
01:04
Can we interpret what someone is seeing using only their brainwaves?
24
64794
3253
01:08
Nathan's going to begin by placing electrodes on Christy's head.
25
68071
3000
01:11
Nathan: My life is tangled.
26
71095
1523
01:12
(Laughter)
27
72642
1150
01:14
GG: And then he's going to show her a bunch of pictures
28
74152
2584
01:16
from four different categories.
29
76760
1521
01:18
Nathan: Face, house, scenery and weird pictures.
30
78305
2654
01:20
GG: As we show Christy hundreds of these images,
31
80983
2498
01:23
we are also capturing the electrical waves onto Nathan's computer.
32
83505
3543
01:27
We want to see if we can detect any visual information about the photos
33
87072
3386
01:30
contained in the brainwaves,
34
90482
1352
01:31
so when we're done, we're going to see if the EEG
35
91858
2331
01:34
can tell us what kind of picture Christy is looking at,
36
94213
2598
01:36
and if it does, each category should trigger a different brain signal.
37
96835
3584
01:40
OK, so we collected all the raw EEG data,
38
100443
2628
01:43
and this is what we got.
39
103095
1150
01:45
It all looks pretty messy, so let's arrange them by picture.
40
105389
2938
01:48
Now, still a bit too noisy to see any differences,
41
108826
2656
01:51
but if we average the EEG across all image types
42
111506
3040
01:54
by aligning them to when the image first appeared,
43
114570
2436
01:57
we can remove this noise,
44
117030
1617
01:58
and pretty soon, we can see some dominant patterns
45
118671
2334
02:01
emerge for each category.
46
121029
1564
02:02
Now the signals all still look pretty similar.
47
122617
2156
02:04
Let's take a closer look.
48
124797
1215
02:06
About a hundred milliseconds after the image comes on,
49
126036
2525
02:08
we see a positive bump in all four cases,
50
128585
2628
02:11
and we call this the P100, and what we think that is
51
131237
2789
02:14
is what happens in your brain when you recognize an object.
52
134050
3075
02:17
But damn, look at that signal for the face.
53
137149
2086
02:19
It looks different than the others.
54
139259
1711
02:20
There's a negative dip about 170 milliseconds
55
140994
2890
02:23
after the image comes on.
56
143908
1540
02:25
What could be going on here?
57
145472
1750
02:27
Research shows that our brain has a lot of neurons that are dedicated
58
147246
3240
02:30
to recognizing human faces,
59
150510
1459
02:31
so this N170 spike could be all those neurons
60
151993
2844
02:34
firing at once in the same location,
61
154861
1985
02:36
and we can detect that in the EEG.
62
156870
1634
02:39
So there are two takeaways here.
63
159083
1820
02:40
One, our eyes can't really detect the differences in patterns
64
160927
3085
02:44
without averaging out the noise,
65
164036
1571
02:45
and two, even after removing the noise,
66
165631
2237
02:47
our eyes can only pick up the signals associated with faces.
67
167892
3001
02:50
So this is where we turn to machine learning.
68
170917
2268
02:53
Now, our eyes are not very good at picking up patterns in noisy data,
69
173209
3976
02:57
but machine learning algorithms are designed to do just that,
70
177209
2946
03:00
so could we take a lot of pictures and a lot of data
71
180179
3201
03:03
and feed it in and train a computer
72
183404
1790
03:05
to be able to interpret what Christy is looking at in real time?
73
185218
3381
03:09
We're trying to code the information that's coming out of her EEG
74
189088
4117
03:13
in real time
75
193229
1175
03:14
and predict what it is that her eyes are looking at.
76
194428
2461
03:16
And if it works, what we should see
77
196913
1727
03:18
is every time that she gets a picture of scenery,
78
198664
2381
03:21
it should say scenery, scenery, scenery, scenery.
79
201069
2286
03:23
A face -- face, face, face, face,
80
203379
1957
03:25
but it's not quite working that way, is what we're discovering.
81
205360
3531
03:33
(Laughter)
82
213385
3548
03:36
OK.
83
216957
1151
03:38
Director: So what's going on here? GG: We need a new career, I think.
84
218132
3382
03:41
(Laughter)
85
221538
1070
03:42
OK, so that was a massive failure.
86
222632
2444
03:45
But we're still curious: How far could we push this technology?
87
225100
3212
03:48
And we looked back at what we did.
88
228336
1640
03:50
We noticed that the data was coming into our computer very quickly,
89
230000
3143
03:53
without any timing of when the images came on,
90
233167
2241
03:55
and that's the equivalent of reading a very long sentence
91
235432
2876
03:58
without spaces between the words.
92
238332
1605
03:59
It would be hard to read,
93
239961
1438
04:01
but once we add the spaces, individual words appear
94
241423
3713
04:05
and it becomes a lot more understandable.
95
245160
2044
04:07
But what if we cheat a little bit?
96
247228
1847
04:09
By using a sensor, we can tell the computer when the image first appears.
97
249099
3537
04:12
That way, the brainwave stops being a continuous stream of information,
98
252660
3602
04:16
and instead becomes individual packets of meaning.
99
256286
2711
04:19
Also, we're going to cheat a little bit more,
100
259021
2368
04:21
by limiting the categories to two.
101
261413
1812
04:23
Let's see if we can do some real-time mind-reading.
102
263249
2383
04:25
In this new experiment,
103
265656
1235
04:26
we're going to constrict it a little bit more
104
266915
2097
04:29
so that we know the onset of the image
105
269036
2252
04:31
and we're going to limit the categories to "face" or "scenery."
106
271312
3382
04:35
Nathan: Face. Correct.
107
275097
1511
04:37
Scenery. Correct.
108
277780
1351
04:40
GG: So right now, every time the image comes on,
109
280251
2373
04:42
we're taking a picture of the onset of the image
110
282648
2266
04:44
and decoding the EEG.
111
284938
1695
04:46
It's getting correct.
112
286657
1256
04:47
Nathan: Yes. Face. Correct.
113
287937
1579
04:49
GG: So there is information in the EEG signal, which is cool.
114
289540
2859
04:52
We just had to align it to the onset of the image.
115
292423
2537
04:55
Nathan: Scenery. Correct.
116
295307
1311
04:59
Face. Yeah.
117
299344
1150
05:00
GG: This means there is some information there,
118
300518
2288
05:02
so if we know at what time the picture came on,
119
302830
2913
05:05
we can tell what type of picture it was,
120
305767
1999
05:07
possibly, at least on average, by looking at these evoked potentials.
121
307790
5096
05:12
Nathan: Exactly.
122
312910
1325
05:14
GG: If you had told me at the beginning of this project this was possible,
123
314259
3521
05:17
I would have said no way.
124
317804
1251
05:19
I literally did not think we could do this.
125
319079
2000
05:21
Did our mind-reading experiment really work?
126
321103
2066
05:23
Yes, but we had to do a lot of cheating.
127
323193
1975
05:25
It turns out you can find some interesting things in the EEG,
128
325192
2905
05:28
for example if you're looking at someone's face,
129
328121
2290
05:30
but it does have a lot of limitations.
130
330435
2157
05:32
Perhaps advances in machine learning will make huge strides,
131
332616
2946
05:35
and one day we will be able to decode what's going on in our thoughts.
132
335586
3390
05:39
But for now, the next time a company says that they can harness your brainwaves
133
339000
4077
05:43
to be able to control devices,
134
343101
1750
05:44
it is your right, it is your duty to be skeptical.
135
344875
3310
About this website

This site will introduce you to YouTube videos that are useful for learning English. You will see English lessons taught by top-notch teachers from around the world. Double-click on the English subtitles displayed on each video page to play the video from there. The subtitles scroll in sync with the video playback. If you have any comments or requests, please contact us using this contact form.

https://forms.gle/WvT1wiN1qDtmnspy7