WEBVTT

00:00:00.090 --> 00:00:01.680
The following
content is provided

00:00:01.680 --> 00:00:03.820
under a Creative
Commons license.

00:00:03.820 --> 00:00:06.550
Your support will help MIT
OpenCourseWare continue

00:00:06.550 --> 00:00:10.160
to offer high quality
educational resources for free.

00:00:10.160 --> 00:00:12.700
To make a donation or to
view additional materials

00:00:12.700 --> 00:00:16.620
from hundreds of MIT courses,
visit MIT OpenCourseWare

00:00:16.620 --> 00:00:17.275
at ocw.mit.edu.

00:00:26.024 --> 00:00:28.440
PROFESSOR: Last time we were
talking about the hair cells.

00:00:28.440 --> 00:00:30.710
And there's a picture
of a hair cell here.

00:00:32.340 --> 00:00:35.710
And what did we do
about the hair cells?

00:00:35.710 --> 00:00:39.980
We talked about the two types,
inner hair cells-- plain old,

00:00:39.980 --> 00:00:42.630
ordinary receptor cells.

00:00:42.630 --> 00:00:44.310
And we talked about
outer hair cells,

00:00:44.310 --> 00:00:46.100
which have this
wonderful characteristic

00:00:46.100 --> 00:00:47.360
of electromotility.

00:00:48.550 --> 00:00:52.320
When their hair bundle
is bent back and forth,

00:00:52.320 --> 00:00:54.165
their internal
potential changes.

00:00:55.200 --> 00:00:57.845
When they're depolarized,
the cell shortens.

00:00:59.330 --> 00:01:02.280
And somehow this
mechanical shortening

00:01:02.280 --> 00:01:06.455
adds to the vibrations
of the organ of Corti

00:01:06.455 --> 00:01:08.860
and basilar membrane that
were set up by sound.

00:01:09.950 --> 00:01:12.930
And it amplifies
those vibrations

00:01:12.930 --> 00:01:14.790
so that the inner
hair cells then

00:01:14.790 --> 00:01:16.625
are responding to an
amplified vibration.

00:01:18.450 --> 00:01:20.100
And the outer hair
cells are then

00:01:20.100 --> 00:01:23.680
dubbed by the name of
the cochlear amplifier.

00:01:23.680 --> 00:01:29.290
Without the amplification,
you lose 40 to 60 dB

00:01:29.290 --> 00:01:31.920
of hearing, which
is a big amount.

00:01:31.920 --> 00:01:33.790
A large hearing
loss would result

00:01:33.790 --> 00:01:36.050
if you didn't have
outer hair cells

00:01:36.050 --> 00:01:38.480
or without their
electromotility, which

00:01:38.480 --> 00:01:42.850
were shown by deleting
the gene for Preston

00:01:42.850 --> 00:01:47.730
and testing the knockout
mouse, which had a 40 to 60

00:01:47.730 --> 00:01:48.725
dB hearing loss.

00:01:51.690 --> 00:01:54.120
So questions about that?

00:01:54.120 --> 00:01:56.620
Before we talked about
the two hair cells,

00:01:56.620 --> 00:02:00.130
we talked about vibration in the
cochlear, what of tuning curve

00:02:00.130 --> 00:02:01.330
is.

00:02:01.330 --> 00:02:06.520
In that case for tuning
of the basilar membrane,

00:02:06.520 --> 00:02:09.100
a particular point
along the cochlear--

00:02:09.100 --> 00:02:11.460
so you could bring your
measurement device to one

00:02:11.460 --> 00:02:15.890
particular place and measure
its tuning in response to sounds

00:02:15.890 --> 00:02:18.520
of different
frequencies, and show

00:02:18.520 --> 00:02:21.660
that a single
place vibrates very

00:02:21.660 --> 00:02:24.720
nicely to a
particular frequency.

00:02:24.720 --> 00:02:27.190
That is you don't have
to send in very much

00:02:27.190 --> 00:02:29.240
sound into the ear.

00:02:29.240 --> 00:02:31.920
But if you go off that
particular frequency,

00:02:31.920 --> 00:02:33.700
you have to boost
the sound a lot

00:02:33.700 --> 00:02:36.015
to get that one
place to vibrate.

00:02:39.090 --> 00:02:42.170
And we had the example
of the vibration patterns

00:02:42.170 --> 00:02:48.500
that low frequencies stimulated
the cochlear apex the most, way

00:02:48.500 --> 00:02:51.020
up near the top of
the snail shell.

00:02:51.020 --> 00:02:53.800
And middle frequencies
stimulated the middle.

00:02:53.800 --> 00:02:57.480
And high frequencies
stimulated the basal part.

00:02:59.040 --> 00:03:01.480
We'll be talking a lot
more about that frequency

00:03:01.480 --> 00:03:04.640
organization along
the cochlear today,

00:03:04.640 --> 00:03:07.860
when we talk about
auditory nerve fibers.

00:03:09.340 --> 00:03:11.005
So here's a roadmap for today.

00:03:12.060 --> 00:03:14.320
We're going to concentrate
on the auditory nerve.

00:03:15.960 --> 00:03:17.936
And I just put down
some numbers so I

00:03:17.936 --> 00:03:20.060
wouldn't forget to tell
you how many auditory nerve

00:03:20.060 --> 00:03:21.304
fibers there are.

00:03:21.304 --> 00:03:25.690
There are approximately
30,000 auditory nerve fibers

00:03:25.690 --> 00:03:26.750
in humans.

00:03:26.750 --> 00:03:29.270
So that means in your
left ear you have 30,000.

00:03:29.270 --> 00:03:32.880
And in you're right ear you
have 30,000 sending messages

00:03:32.880 --> 00:03:34.950
from the ear to the brain.

00:03:34.950 --> 00:03:36.640
So that's a pretty
hefty number, right?

00:03:37.710 --> 00:03:41.110
How many optic
nerve fibers do you

00:03:41.110 --> 00:03:43.390
have or does a primate have?

00:03:43.390 --> 00:03:46.970
I'm sure Dr. Schiller
went over that number.

00:03:46.970 --> 00:03:49.880
We're pretty visual on animals.

00:03:51.470 --> 00:03:54.470
So our sense of vision
is well developed.

00:03:54.470 --> 00:03:58.310
So how many nerve fibers
go from the retina

00:03:58.310 --> 00:04:00.960
into the brain compared
to this number?

00:04:02.892 --> 00:04:03.600
Anybody remember?

00:04:05.100 --> 00:04:06.750
Well, that's a good
number to remember.

00:04:06.750 --> 00:04:11.610
It turns out there about 1
million optic nerve fibers

00:04:11.610 --> 00:04:13.200
from the retina into the brain.

00:04:13.200 --> 00:04:15.510
And here we have 30,000.

00:04:15.510 --> 00:04:19.470
So which is the most important
sense, vision or audition?

00:04:19.470 --> 00:04:23.390
Or which sense conveys
messages more efficiently,

00:04:23.390 --> 00:04:24.150
should we say?

00:04:26.310 --> 00:04:29.920
Well, obviously, primates
are very visual animals.

00:04:29.920 --> 00:04:33.831
So we have a lot more nerve
fibers sending messages

00:04:33.831 --> 00:04:35.830
into the brain about
vision than we do audition.

00:04:37.860 --> 00:04:41.440
So I may not have given you
numbers for the hair cells.

00:04:41.440 --> 00:04:46.250
In humans we have about
3,500 inner hair cells

00:04:46.250 --> 00:04:51.122
and about 12,000 outer
hair cells per cochlea.

00:04:51.122 --> 00:04:52.330
OK, so those are the numbers.

00:04:54.840 --> 00:04:59.110
So today, we'll talk about
the two types of nerve fibers.

00:04:59.110 --> 00:05:00.800
As we have two
types of hair cells,

00:05:00.800 --> 00:05:03.100
we have two types
of nerve fibers.

00:05:03.100 --> 00:05:05.250
We'll talk about
tuning curves now

00:05:05.250 --> 00:05:09.310
for the responses of
auditory nerve fibers.

00:05:09.310 --> 00:05:13.320
And we'll talk about
tonotopic organization.

00:05:13.320 --> 00:05:15.930
That is organization
of frequency

00:05:15.930 --> 00:05:18.960
to place within
the cochlea, which

00:05:18.960 --> 00:05:21.790
is one of the codes
for sound frequency.

00:05:21.790 --> 00:05:25.600
How do we know we're listening
to 1,000 Hertz and not 2,000

00:05:25.600 --> 00:05:28.660
Hertz by which place
along the cochlea

00:05:28.660 --> 00:05:31.380
and which group of auditory
nerve fibers is responding.

00:05:33.190 --> 00:05:35.190
Then we'll get away
from auditory nerve

00:05:35.190 --> 00:05:37.490
and have some listening
demonstrations.

00:05:37.490 --> 00:05:40.970
We'll see how good we
are at discriminating

00:05:40.970 --> 00:05:42.875
two frequencies that
are very close together.

00:05:44.170 --> 00:05:46.590
And we'll talk about
some tuning curves

00:05:46.590 --> 00:05:49.095
that are based on
psychophysical measure.

00:05:50.120 --> 00:05:50.970
That is listening.

00:05:52.190 --> 00:05:54.630
You can take some tuning curves
by just a human listener.

00:05:55.920 --> 00:05:57.970
Then we'll get back
to auditory nerve

00:05:57.970 --> 00:06:01.530
and talk about a different
code for sound frequency.

00:06:01.530 --> 00:06:05.780
That is the temporal code
for sound frequency, which

00:06:05.780 --> 00:06:08.580
involves a phenomenon
called phase

00:06:08.580 --> 00:06:10.340
locking of the auditory nerve.

00:06:11.430 --> 00:06:13.010
Then we'll talk
about how that's very

00:06:13.010 --> 00:06:17.070
important in your listening
to musical intervals.

00:06:17.070 --> 00:06:19.370
And the most important
musical interval

00:06:19.370 --> 00:06:22.380
is the octave, so we'll have
a demonstration of an octave.

00:06:25.280 --> 00:06:31.460
OK, so one of my problems as I
pass by the MIT Coop on the way

00:06:31.460 --> 00:06:34.360
to class, and I
always buy something.

00:06:34.360 --> 00:06:37.060
So I did a reading last week.

00:06:37.060 --> 00:06:39.800
And so we'll have a little
reading from this book,

00:06:39.800 --> 00:06:40.720
I Am Malala.

00:06:41.940 --> 00:06:46.460
She was the girl who was
shot right and recovered

00:06:46.460 --> 00:06:50.220
and was a candidate for
the Nobel Peace Prize.

00:06:50.220 --> 00:06:52.700
Maybe next year she'll
get the Peace Prize.

00:06:54.077 --> 00:06:54.910
I haven't read this.

00:06:54.910 --> 00:06:56.770
I just picked it up
a few minutes ago.

00:06:56.770 --> 00:06:59.186
But I went straight to the
section about her surgery.

00:07:00.290 --> 00:07:03.210
So she was shot in
the head on one side.

00:07:03.210 --> 00:07:05.390
And she said, "While
I was in surgery--"

00:07:05.390 --> 00:07:07.670
this is after recovery.

00:07:07.670 --> 00:07:11.140
This is a further
surgery she underwent.

00:07:11.140 --> 00:07:14.050
"While I was in surgery,
Mr. Irving, the surgeon who

00:07:14.050 --> 00:07:18.330
had repaired my nerve--"
that's her facial nerve--

00:07:18.330 --> 00:07:22.120
"also had a solution for
my damaged left ear drum.

00:07:22.120 --> 00:07:26.200
He put a small electronic
device called a cochlear implant

00:07:26.200 --> 00:07:31.800
inside my head near the ear,
and told me that in a month

00:07:31.800 --> 00:07:34.610
they would fit the
external part on my head.

00:07:34.610 --> 00:07:37.570
And then I should be able
to hear from that ear."

00:07:37.570 --> 00:07:41.990
OK, so the cochlear
implant is a device

00:07:41.990 --> 00:07:45.000
that stimulates the
auditory nerve fibers.

00:07:45.000 --> 00:07:49.480
And in a person who's
had a gunshot wound--

00:07:49.480 --> 00:07:53.570
either because of the loud
sound or the mechanical trauma

00:07:53.570 --> 00:07:56.450
to the ear or temporal
bone-- possibly

00:07:56.450 --> 00:08:00.090
the hair cells are damaged
or are completely missing.

00:08:00.090 --> 00:08:02.020
And the auditory
nerve fibers remain.

00:08:03.450 --> 00:08:05.870
The person is deaf
without the hair cells.

00:08:05.870 --> 00:08:09.190
But the device called
the cochlear implant

00:08:09.190 --> 00:08:12.830
can be inserted inside
this person's cochlear

00:08:12.830 --> 00:08:14.675
to stimulate the auditory nerve.

00:08:15.990 --> 00:08:19.760
And we'll have a discussion of
the cochlear implant next week

00:08:19.760 --> 00:08:23.980
when we have a demonstrator
come to class who's deaf.

00:08:23.980 --> 00:08:25.960
And she'll show you
about her implant.

00:08:25.960 --> 00:08:29.810
But we do need to know a
lot about the auditory nerve

00:08:29.810 --> 00:08:33.230
response before we can
really think about what

00:08:33.230 --> 00:08:36.500
is the good coding strategy
for cochlear implant.

00:08:36.500 --> 00:08:40.280
That is how do we take
the sound information

00:08:40.280 --> 00:08:43.039
and translate it
into the shocks that

00:08:43.039 --> 00:08:45.250
are provided by the cochlear
implant electrodes that

00:08:45.250 --> 00:08:47.300
stimulate the nerve fibers.

00:08:47.300 --> 00:08:50.630
Because little electric
currents in the cochlear implant

00:08:50.630 --> 00:08:53.610
are made to stimulate
the auditory nerve fibers

00:08:53.610 --> 00:08:55.745
that can then send
messages to the brain.

00:08:56.820 --> 00:08:59.130
So it's just a little
motivator for what's

00:08:59.130 --> 00:09:01.230
important about
auditory nerve code.

00:09:03.980 --> 00:09:08.780
So we'll start out today
with the hair cells.

00:09:08.780 --> 00:09:10.940
And these are the
auditory nerves here.

00:09:12.850 --> 00:09:16.150
One thing that's interesting
about vision and audition

00:09:16.150 --> 00:09:19.500
is the look of the
synapse between the hair

00:09:19.500 --> 00:09:24.820
cell and the nerve fiber, and
between the photoreceptor--

00:09:24.820 --> 00:09:27.520
you have rods and
cones in the retina.

00:09:27.520 --> 00:09:30.620
And they have associated
nerve terminals here.

00:09:30.620 --> 00:09:33.020
And these are
electron micrographs,

00:09:33.020 --> 00:09:36.630
taken with a very high powered
electron microscope that

00:09:36.630 --> 00:09:41.840
looks at the synapse between
the photoreceptor up here

00:09:41.840 --> 00:09:46.350
or the hair cell up here and the
associated nerve terminal down

00:09:46.350 --> 00:09:50.970
here, or the associated
either horizontal cell

00:09:50.970 --> 00:09:52.725
or bipolar cell down here.

00:09:53.990 --> 00:09:56.460
So in each case,
you have obviously

00:09:56.460 --> 00:09:58.290
the synapse here
is a little gap.

00:09:59.880 --> 00:10:03.200
And you have synaptic
vesicles that

00:10:03.200 --> 00:10:04.985
contain the neurotransmitter.

00:10:04.985 --> 00:10:07.500
And they're indicated
here in the photoreceptor.

00:10:07.500 --> 00:10:08.880
SV is the vesicle.

00:10:08.880 --> 00:10:11.075
And inside that vesicle
is the neurotransmitter.

00:10:12.330 --> 00:10:16.730
When the receptor
cell depolarizes,

00:10:16.730 --> 00:10:20.120
these synaptic vesicles
fuse and release

00:10:20.120 --> 00:10:24.920
their neurotransmitter
into the cleft and fire

00:10:24.920 --> 00:10:28.390
or activate their
post-synaptic element.

00:10:28.390 --> 00:10:32.060
In the case of the hair cell,
the auditory nerve fiber.

00:10:32.060 --> 00:10:37.270
This structure here is
called the synaptic ribbon.

00:10:37.270 --> 00:10:41.440
And it's supposed to coordinate
the release of the vesicles.

00:10:41.440 --> 00:10:44.380
And they call it a
ribbon in the hair cell

00:10:44.380 --> 00:10:47.460
here, even though it looks
like a big which ball.

00:10:47.460 --> 00:10:49.730
It doesn't look like
a ribbon at all.

00:10:49.730 --> 00:10:51.680
But it's called a
ribbon, because it

00:10:51.680 --> 00:10:53.230
has the same molecular basis.

00:10:53.230 --> 00:10:59.380
It has a lot of interesting
proteins and mechanisms

00:10:59.380 --> 00:11:02.390
to coordinate the release
of these neurotransmitter

00:11:02.390 --> 00:11:04.790
vesicles, which presumably
are synthesize up here

00:11:04.790 --> 00:11:08.450
in the cytoplasm and are
brought down to the ribbon

00:11:08.450 --> 00:11:13.290
and coordinated and released
at the hair cell to nerve fiber

00:11:13.290 --> 00:11:13.790
synapse.

00:11:13.790 --> 00:11:16.580
So I just wanted to
show you the look

00:11:16.580 --> 00:11:19.950
of the synapse in the
electron microscope.

00:11:19.950 --> 00:11:21.230
So that's what it looks like.

00:11:23.730 --> 00:11:29.440
And the next slide
here is this schematic

00:11:29.440 --> 00:11:33.360
of the two types of hair cells,
inner hair cells and the three

00:11:33.360 --> 00:11:36.185
rows of outer hair cells, and
their associated nerve fibers.

00:11:37.650 --> 00:11:41.310
And I think I mentioned
last time that almost all

00:11:41.310 --> 00:11:44.290
of the nerve fibers, the ones
that are sending messages

00:11:44.290 --> 00:11:47.350
to the brain at least, are
associated with the inner hair

00:11:47.350 --> 00:11:47.850
cells.

00:11:47.850 --> 00:11:50.430
So you can see how many
individual terminals there

00:11:50.430 --> 00:11:56.900
are-- as many as 20 on a
single inner hair cell.

00:11:56.900 --> 00:12:00.300
By contrast, the outer
hair cells-- you can see,

00:12:00.300 --> 00:12:02.370
well, this one
has three of them.

00:12:02.370 --> 00:12:04.520
But they're all coming
from the same fiber, which

00:12:04.520 --> 00:12:08.190
also innervates the
neighboring hair cells.

00:12:08.190 --> 00:12:11.030
So there are very few
of these so-called type

00:12:11.030 --> 00:12:13.686
two auditory nerve fibers.

00:12:13.686 --> 00:12:14.560
Here are the numbers.

00:12:14.560 --> 00:12:16.780
So this total is in cats.

00:12:16.780 --> 00:12:19.140
Cats have more nerve
fibers than humans,

00:12:19.140 --> 00:12:22.250
a total of maybe 50,000.

00:12:22.250 --> 00:12:28.110
About 45,000 of them are
the type ones, associated

00:12:28.110 --> 00:12:30.970
with inner hair
cells, and only 5,000

00:12:30.970 --> 00:12:33.645
or the type twos associated
with outer hair cells.

00:12:37.530 --> 00:12:39.870
So you can see by
this ratio then

00:12:39.870 --> 00:12:43.720
that most of the information
is being sent into the brain

00:12:43.720 --> 00:12:47.670
by the type one fibers, sending
messages from the inner hair

00:12:47.670 --> 00:12:48.170
cells.

00:12:49.740 --> 00:12:52.960
Those axons of the type
one fibers are thick.

00:12:54.410 --> 00:12:59.250
They have a myelin
covering, compared

00:12:59.250 --> 00:13:01.505
to the type two fibers,
which are very thin

00:13:01.505 --> 00:13:02.546
and they're unmyelinated.

00:13:04.370 --> 00:13:08.450
And actually, one of the very
interesting unknown facts

00:13:08.450 --> 00:13:13.860
about the auditory system
is that as far as we know,

00:13:13.860 --> 00:13:18.180
no recordings have ever been
made to sample the type two

00:13:18.180 --> 00:13:19.340
responses to sound.

00:13:20.501 --> 00:13:24.060
Do they respond to
different frequencies?

00:13:24.060 --> 00:13:25.147
Are they widely tuned?

00:13:25.147 --> 00:13:25.730
Narrowly tune?

00:13:25.730 --> 00:13:27.700
We don't know that at all.

00:13:27.700 --> 00:13:31.350
And it turns out that it's
just very difficult to sample

00:13:31.350 --> 00:13:36.745
from such thin axons as you
find in the type two fibers.

00:13:39.550 --> 00:13:43.610
So I actually have a grant
submitted to the National

00:13:43.610 --> 00:13:46.710
Institute of Health to use
a special type of electrodes

00:13:46.710 --> 00:13:49.050
to record from the type twos.

00:13:49.050 --> 00:13:51.550
I think it's being
reviewed next week.

00:13:51.550 --> 00:13:54.260
And I hope it gets funded
because then maybe I'll

00:13:54.260 --> 00:13:55.990
figure out this mystery.

00:13:57.540 --> 00:13:59.960
But it will be challenging
not only because they're thin,

00:13:59.960 --> 00:14:02.250
but because there
are fewer of them.

00:14:05.260 --> 00:14:09.160
So when I talk about auditory
nerve fiber recordings

00:14:09.160 --> 00:14:12.950
for this class, I'm going to
be talking about the type ones.

00:14:12.950 --> 00:14:14.700
That's the only kind we know of.

00:14:17.370 --> 00:14:22.220
And here is an example tuning
curve or receptive field

00:14:22.220 --> 00:14:24.330
for a type one
auditory nerve fiber.

00:14:26.180 --> 00:14:28.650
Now, I think Peter
Schiller probably

00:14:28.650 --> 00:14:32.960
talked about single
unit recordings

00:14:32.960 --> 00:14:36.275
with micro electrodes.

00:14:43.470 --> 00:14:44.600
So you have your nerve.

00:14:51.160 --> 00:14:52.700
It could be the optic nerve.

00:14:52.700 --> 00:14:56.270
It could be the
auditory nerve, which

00:14:56.270 --> 00:14:57.530
is what we're talking about.

00:14:59.482 --> 00:15:17.040
You have a microelectrode,
which is put into the nerve.

00:15:17.040 --> 00:15:18.820
And the tip of
the microelectrode

00:15:18.820 --> 00:15:20.230
is very, very tiny.

00:15:20.230 --> 00:15:23.440
It could be less than 1
micrometer in diameter.

00:15:25.790 --> 00:15:29.920
And usually the electrode
is filled with a conducting

00:15:29.920 --> 00:15:32.920
solution like
potassium chloride.

00:15:34.270 --> 00:15:37.830
And the pipette that's
filled with a KCL

00:15:37.830 --> 00:15:39.260
comes out to a big open end.

00:15:39.260 --> 00:15:45.450
And you can stick a wire in here
and run it to your amplifier

00:15:45.450 --> 00:15:48.520
and record the so-called spikes.

00:15:48.520 --> 00:15:51.410
You guys talked
about spikes, right?

00:15:51.410 --> 00:15:59.810
So you're recording the
spikes, AKA action potentials,

00:15:59.810 --> 00:16:01.186
AKA impulses.

00:16:03.330 --> 00:16:07.560
And if you want to do
this in a dramatic way,

00:16:07.560 --> 00:16:10.540
you send this signal
also to a loud speaker

00:16:10.540 --> 00:16:11.750
and you listen to them.

00:16:14.800 --> 00:16:16.560
And maybe we'll
have a demonstration

00:16:16.560 --> 00:16:18.610
at the end of the year on these.

00:16:18.610 --> 00:16:20.860
It's pretty nice
to listen to that.

00:16:20.860 --> 00:16:22.850
So you put your
electrode in there.

00:16:22.850 --> 00:16:26.230
And you move it
around until you have

00:16:26.230 --> 00:16:28.075
what's called a single unit.

00:16:34.320 --> 00:16:36.190
And why is it called
a single unit?

00:16:36.190 --> 00:16:38.490
Well, in the old
days, people didn't

00:16:38.490 --> 00:16:40.090
know what was being recorded.

00:16:40.090 --> 00:16:41.340
Is it a cell body?

00:16:43.440 --> 00:16:44.912
Is it a nerve axon?

00:16:44.912 --> 00:16:45.703
Is it the dendrite?

00:16:48.190 --> 00:16:50.200
What is it ?

00:16:50.200 --> 00:16:52.900
All they knew is that
coming out of the amplifier,

00:16:52.900 --> 00:16:54.735
they saw this spike.

00:17:00.430 --> 00:17:01.790
And that's what's plotted here.

00:17:01.790 --> 00:17:02.956
These are a bunch of spikes.

00:17:03.940 --> 00:17:06.890
And it's called a single
unit, because most

00:17:06.890 --> 00:17:09.260
of the time when you get
one of these recordings,

00:17:09.260 --> 00:17:11.109
the spikes look all the same.

00:17:11.109 --> 00:17:13.775
But every now and then you get a
recording that looks like this.

00:17:21.660 --> 00:17:26.340
And this is interpreted as
being fiber or axon number one.

00:17:27.369 --> 00:17:29.400
Here's another number one.

00:17:29.400 --> 00:17:32.510
And this is a second
fiber that's nearby.

00:17:32.510 --> 00:17:34.230
But it's a different one.

00:17:34.230 --> 00:17:37.429
Maybe there were
actually two fibers right

00:17:37.429 --> 00:17:38.220
next to each other.

00:17:38.220 --> 00:17:39.660
And you could
record both of them.

00:17:39.660 --> 00:17:40.695
That's very unusual.

00:17:41.730 --> 00:17:47.160
More commonly, you just have a
recording from one single unit.

00:17:47.160 --> 00:17:49.190
And the interpretation
is you are sampling

00:17:49.190 --> 00:17:56.840
from just one auditory nerve
fiber out of a total of 40,000.

00:17:56.840 --> 00:17:57.750
Is that clear?

00:18:00.580 --> 00:18:03.160
So such experiments are
done in the auditory nerve.

00:18:03.160 --> 00:18:08.410
In this case, I think
the experimental animal

00:18:08.410 --> 00:18:09.440
was a Guinea pig.

00:18:11.100 --> 00:18:13.750
And in this case,
it's recordings

00:18:13.750 --> 00:18:16.090
from a chinchilla
auditory nerve.

00:18:19.690 --> 00:18:21.370
So what's the stimulus?

00:18:21.370 --> 00:18:24.150
Well, this is a plot
of sound frequency,

00:18:24.150 --> 00:18:25.525
sound frequency in kilohertz.

00:18:28.470 --> 00:18:33.120
And this axis, on the y-axis,
is sound pressure level.

00:18:33.120 --> 00:18:35.470
So this is how loud
it is, if you will.

00:18:37.610 --> 00:18:41.030
And at a very low
or soft tone level,

00:18:41.030 --> 00:18:45.080
if this frequency is swept
from low to high frequencies,

00:18:45.080 --> 00:18:49.130
there was hardly any spikes
coming from that single unit.

00:18:50.150 --> 00:18:52.580
But if you boosted the
level up a little bit.

00:18:53.720 --> 00:18:58.210
And you came to a frequency
that it was about 10 kilohertz,

00:18:58.210 --> 00:19:00.840
there were a bunch of spikes
produced by that single unit.

00:19:03.030 --> 00:19:06.430
Then if you boosted the level
up so it was a moderate level,

00:19:06.430 --> 00:19:11.780
there were spikes anywhere from
8 kilohertz up to 11 kilohertz.

00:19:11.780 --> 00:19:15.260
All that band of frequencies
caused a response.

00:19:16.780 --> 00:19:21.180
Then at the highest level,
everything caused a response,

00:19:21.180 --> 00:19:25.440
from the lowest frequencies
up to about 12 kilohertz,

00:19:25.440 --> 00:19:26.450
and nothing above.

00:19:29.740 --> 00:19:32.150
What's this activity out here?

00:19:32.150 --> 00:19:34.490
I said nothing above
and nothing over here.

00:19:34.490 --> 00:19:36.850
Well, there's some
spontaneous firing.

00:19:43.970 --> 00:19:47.880
So even if you turn the
sound completely off,

00:19:47.880 --> 00:19:50.530
these nerve fibers have
a little bit of activity.

00:19:50.530 --> 00:19:51.765
They fire some impulses.

00:19:53.040 --> 00:19:54.260
There's an ongoing thing.

00:19:59.300 --> 00:20:02.960
If you outlined this
response area with a line--

00:20:02.960 --> 00:20:06.960
that line is the border, say,
between spontaneous firing

00:20:06.960 --> 00:20:11.400
or no firing and a response.

00:20:11.400 --> 00:20:14.654
So inside of the receptive
area there's a response.

00:20:14.654 --> 00:20:15.820
And outside there's nothing.

00:20:16.970 --> 00:20:19.605
Those lines are
called tuning curves.

00:20:20.690 --> 00:20:24.280
And here are a bunch of tuning
curves from a chinchilla.

00:20:24.280 --> 00:20:28.710
And there are one,
two, three, four, five,

00:20:28.710 --> 00:20:31.280
six different tuning curves.

00:20:31.280 --> 00:20:35.340
So what the experiment did was
they moved the electrode in

00:20:35.340 --> 00:20:36.930
and got one single unit.

00:20:39.290 --> 00:20:42.700
And then they moved
the electrode,

00:20:42.700 --> 00:20:44.350
let's say deeper into the nerve.

00:20:48.890 --> 00:20:51.150
And now they sampled
a different neuron,

00:20:51.150 --> 00:20:52.315
a different single unit.

00:20:53.960 --> 00:20:56.030
OK, maybe got this tuning curve.

00:20:56.030 --> 00:20:57.980
Then they went
deeper and sampled

00:20:57.980 --> 00:21:00.920
from this one and this one
and this one and this one.

00:21:00.920 --> 00:21:03.150
And the idea that it's
a different one-- well,

00:21:03.150 --> 00:21:04.940
the response is different.

00:21:04.940 --> 00:21:07.060
But also, as you
move the electrode,

00:21:07.060 --> 00:21:10.460
you lost the single
unit, number one.

00:21:10.460 --> 00:21:13.900
And you've maybe put it
deeper, a millimeter or so.

00:21:13.900 --> 00:21:15.642
It's a huge distance.

00:21:15.642 --> 00:21:16.725
And you've got a new unit.

00:21:17.960 --> 00:21:20.580
The action potentials
probably look different.

00:21:20.580 --> 00:21:21.295
That's a second.

00:21:24.320 --> 00:21:27.250
OK, so these are tuning
curves there then

00:21:27.250 --> 00:21:29.990
from six different single units.

00:21:29.990 --> 00:21:33.520
And each of them comes
down to a pretty nice tip.

00:21:34.970 --> 00:21:39.990
And if you take that
tip and the very lowest

00:21:39.990 --> 00:21:42.710
sound level they're
caused a response

00:21:42.710 --> 00:21:46.340
and extrapolate that to the
x-axis, you get a frequency.

00:21:47.470 --> 00:21:57.491
And that frequency is called
the CF, or characteristic

00:21:57.491 --> 00:21:57.990
frequency.

00:22:02.100 --> 00:22:04.720
OK, so CF is a very
important term.

00:22:04.720 --> 00:22:08.500
You should know that the CF
is the very tip of the tuning

00:22:08.500 --> 00:22:09.000
curve.

00:22:10.710 --> 00:22:13.610
And the CF is different
from frequency.

00:22:13.610 --> 00:22:16.430
Frequency is whatever
you want to dial in

00:22:16.430 --> 00:22:18.715
with your sound oscillator.

00:22:19.760 --> 00:22:24.490
But CF is a particular
characteristic of a neuron,

00:22:24.490 --> 00:22:26.705
in this case an
auditory nerve fiber,

00:22:26.705 --> 00:22:29.890
that you're recording from.

00:22:29.890 --> 00:22:31.920
And it's a
characteristic that it

00:22:31.920 --> 00:22:36.815
has that you measured from it.

00:22:38.170 --> 00:22:41.320
Many of these tuning
curves, in addition

00:22:41.320 --> 00:22:46.270
to having a CF and a so-called
tip region, also have a tail.

00:22:47.390 --> 00:22:52.340
And in this very high CF
neuron, the tail goes like this.

00:22:52.340 --> 00:22:53.890
And then there's
actually, I think,

00:22:53.890 --> 00:22:57.200
something that I dashed in
here, a dashed line here.

00:22:57.200 --> 00:22:59.730
And the tail continues
way down here.

00:23:01.350 --> 00:23:04.130
These experiments didn't
want to boost the sound level

00:23:04.130 --> 00:23:08.600
to get all the tail above 80
dBs because of possible damage.

00:23:08.600 --> 00:23:11.260
If you crank up too much
sound-- just like you

00:23:11.260 --> 00:23:13.510
get a gunshot to the head
is a very loud sound--

00:23:13.510 --> 00:23:15.694
it can cause damage
to the hair cells.

00:23:15.694 --> 00:23:16.860
They didn't want to do that.

00:23:16.860 --> 00:23:20.120
But you could see the tail
of this response area.

00:23:20.120 --> 00:23:22.985
It's a nice tip and a nice tail.

00:23:28.310 --> 00:23:32.290
OK, now, right away we have
a beautiful potential code

00:23:32.290 --> 00:23:33.790
for sound frequency.

00:23:33.790 --> 00:23:36.510
How do I know I'm
listening to 8 kilohertz?

00:23:37.690 --> 00:23:41.380
Well, this nerve fiber
responds very nicely,

00:23:41.380 --> 00:23:42.705
lots of action potentials.

00:23:44.270 --> 00:23:47.420
How do I know I'm
listening to 1 kilohertz?

00:23:47.420 --> 00:23:49.590
Well, that same nerve
fiber might respond.

00:23:49.590 --> 00:23:52.720
But I have to get the sound
level to very loud level,

00:23:52.720 --> 00:23:55.070
like 80 dBs ATL.

00:23:55.070 --> 00:23:58.290
But these other guys over
here with CFs of 1 kilohertz

00:23:58.290 --> 00:24:00.200
would respond at
a very low sound.

00:24:02.850 --> 00:24:05.620
So then we have a
code of which fiber

00:24:05.620 --> 00:24:09.050
you're listening
to tells you which

00:24:09.050 --> 00:24:10.320
frequency you're listening to.

00:24:10.320 --> 00:24:12.030
It's very important.

00:24:12.030 --> 00:24:15.160
You judge an instrument,
like a violin,

00:24:15.160 --> 00:24:17.920
by its combination
of frequencies.

00:24:17.920 --> 00:24:20.200
A guitar has a different
recombination frequencies.

00:24:21.260 --> 00:24:23.510
Male speakers generally
have deeper voices

00:24:23.510 --> 00:24:27.310
than female speakers, deeper
meaning more low frequencies.

00:24:28.500 --> 00:24:32.640
Female and children's voices
are higher in frequency.

00:24:32.640 --> 00:24:34.420
So frequency is
essential for you

00:24:34.420 --> 00:24:38.690
to identify what sound
stimulus you are listening.

00:24:41.930 --> 00:24:45.460
Why do we call this a place
code for sound frequency?

00:24:45.460 --> 00:24:50.400
Well, as we talked about before,
different parts of the cochlea

00:24:50.400 --> 00:24:52.390
respond to different
frequencies.

00:24:52.390 --> 00:24:57.340
Here is a beautiful
example of the place

00:24:57.340 --> 00:24:59.210
map for auditory nerve fibers.

00:25:00.290 --> 00:25:05.090
And in this case,
microelectrode recordings

00:25:05.090 --> 00:25:08.280
are done as we described before.

00:25:12.240 --> 00:25:16.700
But instead of just a plain
old potassium chloride solution

00:25:16.700 --> 00:25:23.280
in the microelectrode,
it's filled

00:25:23.280 --> 00:25:26.053
with a substance
called a neural tracer.

00:25:30.670 --> 00:25:33.080
What are examples
of neural tracers?

00:25:33.080 --> 00:25:36.950
Has anybody played around
with neural tracers before?

00:25:36.950 --> 00:25:39.000
Give me some
examples of chemicals

00:25:39.000 --> 00:25:40.000
that are neural tracers.

00:25:42.853 --> 00:25:43.353
Anybody?

00:25:47.720 --> 00:25:58.750
This one is a funny name
horseradish peroxidose,

00:25:58.750 --> 00:26:00.060
abbreviated HRP.

00:26:03.080 --> 00:26:10.900
Another one is biocytin,
OK, biotinylated

00:26:10.900 --> 00:26:13.050
dextran amine, PDA.

00:26:13.050 --> 00:26:14.965
There's millions of
them, Lucifer yellow.

00:26:20.320 --> 00:26:22.970
You can tell that I'm
a tracer kind of guy.

00:26:22.970 --> 00:26:25.740
I use tracers all the
time in my experiments.

00:26:25.740 --> 00:26:28.180
So what you do with
these neural tracers,

00:26:28.180 --> 00:26:32.070
it's convenient if
they are charged.

00:26:32.070 --> 00:26:33.820
For example,
horseradish peroxidose,

00:26:33.820 --> 00:26:35.040
this is a positive charge.

00:26:38.830 --> 00:26:45.580
And you can apply positive
current to the pipette up here.

00:26:45.580 --> 00:26:48.990
That's going to tend to
force positive charge

00:26:48.990 --> 00:26:50.710
out the tip of the electrode.

00:26:50.710 --> 00:26:55.520
You can expel a
positive ion out the tip

00:26:55.520 --> 00:26:57.750
by this technique, which
is called iontophoresis.

00:27:05.040 --> 00:27:07.010
And if it happens
that your tip is

00:27:07.010 --> 00:27:16.040
close to or ideally inside
an axon, some of that HRP

00:27:16.040 --> 00:27:18.940
is going to come out
the tip of the electrode

00:27:18.940 --> 00:27:20.250
and go into the axon.

00:27:23.730 --> 00:27:25.180
And why did we pick HRP?

00:27:26.440 --> 00:27:30.750
Because it's picked up by
chemical transport systems

00:27:30.750 --> 00:27:33.085
that transport
things along axons.

00:27:34.059 --> 00:27:35.350
And there are several of these.

00:27:35.350 --> 00:27:38.100
There's fast axonal transport.

00:27:38.100 --> 00:27:38.720
There's slow.

00:27:38.720 --> 00:27:39.575
There's medium.

00:27:39.575 --> 00:27:40.950
There's a whole
bunch of systems,

00:27:40.950 --> 00:27:45.765
because this axon is
coming from a cell here.

00:27:47.536 --> 00:27:49.260
It's connected to the cell body.

00:27:49.260 --> 00:27:53.750
In the cell body you make
things like neurotransmitter,

00:27:53.750 --> 00:27:55.510
because that's where
you can make protein.

00:27:55.510 --> 00:27:58.840
And that neurotransmitter
has to get down

00:27:58.840 --> 00:28:02.410
to the tip of the axon, which in
the case of the auditory nerve

00:28:02.410 --> 00:28:04.460
is in the cochlear
nucleus of the brain.

00:28:05.660 --> 00:28:08.370
So there are all these transport
systems transporting things.

00:28:10.010 --> 00:28:12.394
And it just turns out
that some chemicals

00:28:12.394 --> 00:28:13.310
are picked up by them.

00:28:13.310 --> 00:28:15.190
HRP is one of them.

00:28:15.190 --> 00:28:22.070
When you iontophorese
HRP into a nerve fiber,

00:28:22.070 --> 00:28:25.000
it's transported to all
parts of the nerve fiber,

00:28:25.000 --> 00:28:27.990
including to the cell
and including out

00:28:27.990 --> 00:28:31.360
to the tip of the nerve
fiber on the hair cell.

00:28:32.760 --> 00:28:37.480
So here is an example
of iontophoretically

00:28:37.480 --> 00:28:39.090
labeled nerve fibers.

00:28:39.090 --> 00:28:42.810
And there's five or six of them.

00:28:42.810 --> 00:28:45.440
The recording site was
here in the auditory nerve.

00:28:46.730 --> 00:28:48.494
This is a diagram
of the cochlea.

00:28:50.500 --> 00:28:53.020
This is the so-called
Schwann glial border,

00:28:53.020 --> 00:28:57.740
which defines the
periphery and the brain.

00:28:57.740 --> 00:28:58.950
So this would be the brain.

00:28:58.950 --> 00:29:01.074
So these are the nerve
fibers going into the brain.

00:29:02.440 --> 00:29:04.160
They were recorded in
the auditory nerve.

00:29:04.160 --> 00:29:08.190
And you can trace them
out into the periphery.

00:29:08.190 --> 00:29:13.320
Right here is the cell body
of the auditory nerve fiber.

00:29:13.320 --> 00:29:15.190
Every neuron has a cell body.

00:29:16.260 --> 00:29:18.110
Most neurons have axons.

00:29:18.110 --> 00:29:19.855
The axon was what was recorded.

00:29:21.380 --> 00:29:24.340
And the auditory nerve
neuron has a cell body.

00:29:24.340 --> 00:29:26.410
And it also has
a peripheral axon

00:29:26.410 --> 00:29:30.240
that goes out to the periphery
and contacts an inner hair

00:29:30.240 --> 00:29:30.740
cell.

00:29:32.710 --> 00:29:35.170
As we saw before, these
are type one auditory nerve

00:29:35.170 --> 00:29:36.950
fibers going to
inner hair cells.

00:29:36.950 --> 00:29:39.510
And it contacts usually
one inner hair cell.

00:29:41.310 --> 00:29:45.040
Now, you can know exactly
where that auditory nerve

00:29:45.040 --> 00:29:50.220
fiber started out by
tracing it and by tracing

00:29:50.220 --> 00:29:53.160
the base of the cochlea
through the spiral

00:29:53.160 --> 00:29:54.540
and all the way up to the apex.

00:29:55.660 --> 00:30:00.750
So starting at the base to the
apex-- so that's 100% distance,

00:30:00.750 --> 00:30:01.460
let's say.

00:30:03.110 --> 00:30:06.080
And if this were halfway
between the base and the apex,

00:30:06.080 --> 00:30:08.915
that would be the
50% distance place.

00:30:11.060 --> 00:30:14.700
This guy ending up near the
apex might be 80% distance

00:30:14.700 --> 00:30:16.635
from the base to the apex.

00:30:16.635 --> 00:30:18.760
OK, does everybody see how
I can make that mapping?

00:30:20.520 --> 00:30:24.160
These sausages here
are the outlines

00:30:24.160 --> 00:30:27.090
of the ganglion, the spiral
ganglion, where the cell

00:30:27.090 --> 00:30:28.810
bodies are of the
auditory nerve fibers.

00:30:30.850 --> 00:30:32.120
So what good is that mapping?

00:30:32.120 --> 00:30:39.300
Well, before we put the HRP in,
we measured the tuning curve.

00:30:40.940 --> 00:30:45.700
And we got the CF
from the tuning curve.

00:30:45.700 --> 00:30:48.442
So we measured the CF.

00:30:48.442 --> 00:30:53.390
We injected the neurotransmitter
to label the auditory nerve

00:30:53.390 --> 00:30:54.570
fiber.

00:30:54.570 --> 00:30:58.010
And we reconstructed
where the labeled ending

00:30:58.010 --> 00:31:01.250
of the auditory nerve fiber
contacted its inner hair cell.

00:31:03.180 --> 00:31:04.940
Why did we do this
for five of these?

00:31:04.940 --> 00:31:06.670
Well, in the
ultimate experiment,

00:31:06.670 --> 00:31:07.800
you just do it for one.

00:31:09.370 --> 00:31:12.450
But if you're getting good at
reconstructing the mapping,

00:31:12.450 --> 00:31:16.110
you can tell it should
be about the 50% place.

00:31:16.110 --> 00:31:17.460
And you go and find it's 51%.

00:31:17.460 --> 00:31:20.870
You know that fiber was
different than the one

00:31:20.870 --> 00:31:21.370
up there.

00:31:24.340 --> 00:31:30.810
Then you make your mapping--
characteristic frequency

00:31:30.810 --> 00:31:34.210
to position of enervation
along the cochlea.

00:31:34.210 --> 00:31:35.370
And here is the mapping.

00:31:35.370 --> 00:31:36.120
These are the CFs.

00:31:38.780 --> 00:31:42.630
And this is the percent
distance along the cochlea

00:31:42.630 --> 00:31:43.895
from the base.

00:31:45.570 --> 00:31:50.230
So 0% distance from the base
would be the extreme base.

00:31:50.230 --> 00:31:52.980
100% distance would
be the extreme apex.

00:31:54.370 --> 00:31:57.230
And you can see this
beautiful mapping of CF

00:31:57.230 --> 00:32:01.290
to position, almost
a straight line,

00:32:01.290 --> 00:32:02.920
until you get to the lowest CFs.

00:32:04.840 --> 00:32:07.230
And this, as usual, in
the auditory system,

00:32:07.230 --> 00:32:10.950
this frequency axis,
this is the CF axis now.

00:32:10.950 --> 00:32:12.185
It's on a log scale.

00:32:14.790 --> 00:32:19.560
So log frequency maps to linear
distance along the cochlea.

00:32:22.010 --> 00:32:30.210
Now, if the brain hears that
the 50% distance auditory nerve

00:32:30.210 --> 00:32:33.050
fiber is responding and
no other auditory nerve

00:32:33.050 --> 00:32:36.410
fiber is responding,
it knows it's

00:32:36.410 --> 00:32:38.835
listening to a 3
kilohertz frequency.

00:32:41.220 --> 00:32:46.130
Place to frequency
mapping is tonotopic.

00:32:47.700 --> 00:32:49.720
I said that opposite.

00:32:49.720 --> 00:32:52.460
Frequency to place is tonotopic.

00:32:52.460 --> 00:33:14.040
So this is a tonotopic mapping--
frequency to place, tonotopic.

00:33:14.040 --> 00:33:15.300
And why is that important?

00:33:15.300 --> 00:33:16.940
Well, it happens in the cochlea.

00:33:18.810 --> 00:33:21.090
It happens in the
auditory nerve.

00:33:21.090 --> 00:33:24.140
It happens in the cochlear
nucleus of the brain.

00:33:24.140 --> 00:33:26.780
It happens in almost
all the auditory

00:33:26.780 --> 00:33:30.290
centers in the entire brain,
all the way up to the cortex.

00:33:31.660 --> 00:33:35.400
You have neurons or
fibers responding

00:33:35.400 --> 00:33:38.220
to low CFs over
here in the brain.

00:33:38.220 --> 00:33:40.420
And if you move your
electrode over here,

00:33:40.420 --> 00:33:43.270
you find they're responding
to mid frequencies.

00:33:43.270 --> 00:33:47.750
And if you move them over here,
they're responding to high CFs.

00:33:49.790 --> 00:33:52.180
So this organization
is fundamental.

00:33:52.180 --> 00:33:56.030
It starts at the receptor
level in the cochlea.

00:33:56.030 --> 00:33:59.510
It's conveyed by the nerve
into the cochlear nucleus.

00:33:59.510 --> 00:34:01.770
And you have these
beautiful frequency--

00:34:01.770 --> 00:34:04.740
they're actually CF
organizations in the brain.

00:34:06.620 --> 00:34:09.000
So the place code
for sound frequency

00:34:09.000 --> 00:34:10.960
presumes that each
frequency stimulates

00:34:10.960 --> 00:34:12.695
a certain place along cochlea.

00:34:18.699 --> 00:34:21.219
And I guess, if
you generalize this

00:34:21.219 --> 00:34:25.290
from the auditory system
to the visual system--

00:34:25.290 --> 00:34:27.719
if you have a
particular light source,

00:34:27.719 --> 00:34:31.429
like that light over there, and
my eyes are looking this way,

00:34:31.429 --> 00:34:34.810
that light is going to
stimulate a particular place

00:34:34.810 --> 00:34:37.530
in my left retina and
in my right retina.

00:34:37.530 --> 00:34:42.179
So you have a coding for
where that light is along

00:34:42.179 --> 00:34:43.949
the place in the retina.

00:34:43.949 --> 00:34:45.540
In the auditory
system, you don't

00:34:45.540 --> 00:34:46.969
have that kind of a place code.

00:34:46.969 --> 00:34:49.902
You have a place code
for sound frequency.

00:34:49.902 --> 00:34:50.735
It's very different.

00:34:52.370 --> 00:34:54.349
The cochlea maps frequency.

00:35:01.400 --> 00:35:03.060
How can we use this code?

00:35:04.450 --> 00:35:07.810
We're actually very good at
distinguishing closely spaced

00:35:07.810 --> 00:35:09.030
frequencies.

00:35:09.030 --> 00:35:11.900
And here is now some
psychophyscial data

00:35:11.900 --> 00:35:13.630
from human listeners.

00:35:13.630 --> 00:35:17.630
We're going to get away from
the auditory nerve for awhile

00:35:17.630 --> 00:35:19.135
and talk about
listening studies.

00:35:20.610 --> 00:35:22.366
Here is to graph of frequency.

00:35:24.700 --> 00:35:27.525
And on the y-axis is delta f.

00:35:29.290 --> 00:35:30.215
What's delta f?

00:35:35.180 --> 00:35:42.765
Delta f is the just noticeable
difference for frequency.

00:35:45.732 --> 00:35:47.565
Of course, we're talk
about sound frequency.

00:35:53.430 --> 00:35:54.930
And how is the
experiment conducted?

00:35:54.930 --> 00:35:56.138
Well, you have your listener.

00:35:56.960 --> 00:35:59.110
Your listener is
listening to sound.

00:35:59.110 --> 00:36:01.115
And you give them a
1 kilohertz sound.

00:36:02.460 --> 00:36:05.470
And then you give them
a 2 kilohertz sound.

00:36:05.470 --> 00:36:08.560
The experimenter says, does it
sound the same or different?

00:36:08.560 --> 00:36:10.150
Ah, completely different.

00:36:10.150 --> 00:36:14.820
OK, 1 kilohertz sound
and a 1,100 kilohertz

00:36:14.820 --> 00:36:16.080
sound, same or different?

00:36:16.080 --> 00:36:17.490
Ah completely different.

00:36:17.490 --> 00:36:21.570
OK, 1,000 hertz sound
and 1,010 hertz sound.

00:36:21.570 --> 00:36:22.910
Ah, it's different.

00:36:22.910 --> 00:36:29.590
1,000 hertz and a
1,002 hertz sound?

00:36:29.590 --> 00:36:30.480
I'm not so sure.

00:36:30.480 --> 00:36:31.480
Give it to me again.

00:36:31.480 --> 00:36:34.920
OK, 1,000 hertz
sound, 1,002 hertz?

00:36:34.920 --> 00:36:36.770
Eh, it's just a
little bit different.

00:36:36.770 --> 00:36:39.350
1,000 hertz sound and
a 1,001 hertz sound?

00:36:39.350 --> 00:36:39.850
Same.

00:36:41.040 --> 00:36:43.160
OK, so that's the experiment.

00:36:43.160 --> 00:36:49.840
So we have the graph here for
the just noticeable difference

00:36:49.840 --> 00:36:52.450
in frequency, as a
function of frequency.

00:36:52.450 --> 00:36:54.270
And at 1,000
hertz-- that's right

00:36:54.270 --> 00:36:58.760
in the middle of your hearing
range-- the delta f-- well,

00:36:58.760 --> 00:37:01.700
it's hard to read that
axis-- the delta f

00:37:01.700 --> 00:37:05.285
is about 1 or 2 hertz.

00:37:06.840 --> 00:37:12.260
So 1,000 vs 1,002
hertz is just barely

00:37:12.260 --> 00:37:15.900
distinguishable for
human listeners.

00:37:18.280 --> 00:37:21.010
You can do that experimental
a little bit differently.

00:37:21.010 --> 00:37:25.140
Instead of giving two tones,
you can give one tone and vary

00:37:25.140 --> 00:37:26.570
its frequency a little.

00:37:28.920 --> 00:37:32.030
And that's kind of
a pleasing sound.

00:37:32.030 --> 00:37:34.890
Does everybody
know what a vibrato

00:37:34.890 --> 00:37:37.075
is on a stringed instrument?

00:37:40.420 --> 00:37:42.920
That's a plain A.
But if you vibrate

00:37:42.920 --> 00:37:49.130
it a little-- that's
the frequencies

00:37:49.130 --> 00:37:51.140
going back and forth.

00:37:51.140 --> 00:37:53.760
Everybody could hear
that vibrato right?

00:37:53.760 --> 00:37:57.140
Even though I'm changing the
frequency just a tiny bit.

00:37:57.140 --> 00:38:00.230
You could do the experiment
by vibrating the frequency

00:38:00.230 --> 00:38:01.520
just one single frequency.

00:38:02.750 --> 00:38:03.850
Is it vibrating?

00:38:03.850 --> 00:38:04.955
Or is it not vibrating?

00:38:06.010 --> 00:38:08.330
And you get about
the same result.

00:38:08.330 --> 00:38:10.400
That's what the second graph is.

00:38:11.780 --> 00:38:17.040
People who are tone deaf,
not proficient music,

00:38:17.040 --> 00:38:20.590
don't have any hearing
problems are almost always

00:38:20.590 --> 00:38:23.860
able to distinguish frequencies
with a little bit of training.

00:38:23.860 --> 00:38:27.810
The training is now here's the
task, that type of training.

00:38:27.810 --> 00:38:30.720
OK, so I have a
demonstration here.

00:38:30.720 --> 00:38:34.500
And we can listen to this and
see how good you guys are--

00:38:34.500 --> 00:38:39.290
you know, naive,
untrained listeners--

00:38:39.290 --> 00:38:41.970
and see if we're good at
distinguishing frequency.

00:38:41.970 --> 00:38:46.200
So the demonstration is
a little bit complicated.

00:38:46.200 --> 00:38:47.940
So I'll go through it.

00:38:47.940 --> 00:38:51.520
It's going to give you
1,000 hertz, standard,

00:38:51.520 --> 00:38:53.825
middle of your range
hearing frequency.

00:38:55.680 --> 00:38:58.249
And it's going to give you
a bunch of different groups.

00:38:58.249 --> 00:38:59.790
I'm going to go
through these slowly.

00:39:01.060 --> 00:39:05.340
And in each group, we
have 1,000 one hertz,

00:39:05.340 --> 00:39:08.670
and 1,000 hertz plus delta f.

00:39:08.670 --> 00:39:13.440
OK, delta f for group one is
10 hertz, big frequency space.

00:39:14.680 --> 00:39:17.630
And what you're
going to listen to

00:39:17.630 --> 00:39:27.120
is A, B, A, A, where is f--
1,000 hertz-- and f plus delta

00:39:27.120 --> 00:39:31.010
f-- 1,010 hertz.

00:39:31.010 --> 00:39:36.130
And B will be first, 1,010
hertz and then 1,000 hertz.

00:39:36.130 --> 00:39:40.190
Then A, 1,000
hertz, 1,010 hertz;

00:39:40.190 --> 00:39:43.690
and another A, 1,000
hertz, 1,010 hertz,

00:39:43.690 --> 00:39:45.970
just to give you a bunch
of different examples.

00:39:47.340 --> 00:39:53.370
Then group two, delta f will be
a little bit harder, 9 hertz,

00:39:53.370 --> 00:39:57.350
OK, so on and so forth,
down to group 10, which

00:39:57.350 --> 00:40:00.160
will be delta f of 1 hertz.

00:40:00.160 --> 00:40:08.490
Or seeing if we can distinguish
1,000 and 1,001 hertz.

00:40:08.490 --> 00:40:10.370
OK, so let's listen to this.

00:40:14.300 --> 00:40:18.360
MAN IN AUDIO: Frequency
difference file for J and D.

00:40:18.360 --> 00:40:21.276
You will hear 10 groups
of four tone pairs.

00:40:21.276 --> 00:40:24.358
In each group, there is a
small frequency difference

00:40:24.358 --> 00:40:26.850
between the tones
of the pairs, which

00:40:26.850 --> 00:40:29.190
decreases in each
successive group.

00:40:31.532 --> 00:40:35.004
[BEEPING OF TONE PAIRS]

00:40:39.834 --> 00:40:41.000
PROFESSOR: That's group one.

00:40:42.170 --> 00:40:43.620
Here's group two.

00:40:43.620 --> 00:40:47.120
[BEEPING OF TONE PAIRS]

00:42:35.000 --> 00:42:39.260
PROFESSOR: OK, could
everybody do the big interval,

00:42:39.260 --> 00:42:40.940
delta f equals 10?

00:42:40.940 --> 00:42:43.350
Raise your hand if
you could do that.

00:42:43.350 --> 00:42:47.770
Most-- some people can.

00:42:47.770 --> 00:42:49.570
I-- it's not problem.

00:42:49.570 --> 00:42:52.857
OK, how about your limits
for people who could do it.

00:42:52.857 --> 00:42:53.440
When did you--

00:42:53.440 --> 00:42:54.660
AUDIENCE: I heard eight.

00:42:54.660 --> 00:42:56.030
PROFESSOR: About eight, OK.

00:42:58.420 --> 00:43:00.693
And what was your
limit going down?

00:43:00.693 --> 00:43:01.318
AUDIENCE: Nine.

00:43:02.990 --> 00:43:04.920
PROFESSOR: Group
nine or delta f?

00:43:04.920 --> 00:43:06.515
OK, so delta f of 2.

00:43:08.070 --> 00:43:10.490
I cut out about
between two and three.

00:43:12.890 --> 00:43:16.110
Well, for those of us who could
do it, without any training

00:43:16.110 --> 00:43:22.350
at all, you get to what the
best results are-- people

00:43:22.350 --> 00:43:25.095
who have done this for
days and days and practice.

00:43:26.190 --> 00:43:28.340
And this is not an
ideal listening room.

00:43:28.340 --> 00:43:30.090
There's a lot of fan noise.

00:43:30.090 --> 00:43:33.310
There's some distractions too.

00:43:33.310 --> 00:43:36.470
Ideally, you'd be in a
completely quiet environment,

00:43:36.470 --> 00:43:37.640
perhaps wearing headphones.

00:43:39.180 --> 00:43:40.660
But it works pretty well.

00:43:40.660 --> 00:43:43.450
I'm not sure what it says
about people can't do it.

00:43:44.560 --> 00:43:48.074
And there certainly are.

00:43:48.074 --> 00:43:49.490
So I don't know
if you should have

00:43:49.490 --> 00:43:52.770
your hearing tested or whatever.

00:43:52.770 --> 00:43:55.220
But for those of
us who could do it,

00:43:55.220 --> 00:43:58.910
you get quickly to the
best possible results.

00:43:58.910 --> 00:44:02.770
So you could do a
calculation then on these.

00:44:02.770 --> 00:44:04.500
We know what delta f is.

00:44:04.500 --> 00:44:08.190
Let's say it's 2
hertz at 1,000 hertz.

00:44:08.190 --> 00:44:10.300
And let's go back to
our mapping experiment.

00:44:12.400 --> 00:44:18.600
So here's 1,000 hertz CF.

00:44:18.600 --> 00:44:20.740
Let's say we're
listening at the CF.

00:44:23.140 --> 00:44:28.640
And we're moving from 1,000 to
1,002 hertz, the best possible

00:44:28.640 --> 00:44:30.770
psychophysical performance.

00:44:30.770 --> 00:44:34.490
We can go up along this
cochlear frequency map

00:44:34.490 --> 00:44:37.340
and say, well, what
percent distance did

00:44:37.340 --> 00:44:44.250
we move from the 1,000 hertz
point to the 1,002 hertz point?

00:44:44.250 --> 00:44:46.120
And I don't know where it is.

00:44:46.120 --> 00:44:51.550
Well, it's about the 70%
distance place in this animal.

00:44:51.550 --> 00:44:53.130
This is a cat, of course.

00:44:53.130 --> 00:44:55.150
You can't do these kinds
of studies in humans.

00:44:55.150 --> 00:44:56.950
You could map it out in human.

00:44:58.300 --> 00:45:02.135
It turns out that if you know
how many inner hair cells there

00:45:02.135 --> 00:45:07.670
are-- we had that number before
along the base to apex spiral--

00:45:07.670 --> 00:45:09.760
and you know the
distance you're moving,

00:45:09.760 --> 00:45:12.920
it turns out you can make
the calculation, the best

00:45:12.920 --> 00:45:15.090
possible performance.

00:45:15.090 --> 00:45:18.990
You're moving from one inner
hair cell to its neighbor.

00:45:18.990 --> 00:45:22.450
So it's a very,
very small increment

00:45:22.450 --> 00:45:24.625
along the cochlear
spiral you're moving.

00:45:26.070 --> 00:45:29.110
That increment is
associated with the best

00:45:29.110 --> 00:45:32.570
possible psychophysical
performance

00:45:32.570 --> 00:45:35.180
in terms of frequency
distinction.

00:45:35.180 --> 00:45:38.380
OK, that's the
cochlear frequency map.

00:45:44.030 --> 00:45:45.778
OK, any questions about that?

00:45:48.050 --> 00:45:50.200
Now let's go back to
the auditory nerve

00:45:50.200 --> 00:45:54.430
and talk more about coding
for sound frequency.

00:45:56.670 --> 00:46:03.300
So far, we've just been
exploring single tone response

00:46:03.300 --> 00:46:03.800
areas.

00:46:03.800 --> 00:46:07.800
So now let's make the stimulus
a little bit more advanced

00:46:07.800 --> 00:46:10.325
and talk about
coding for two tones.

00:46:12.290 --> 00:46:14.615
What happens when
you have two tones?

00:46:14.615 --> 00:46:18.820
Well, here is a tuning curve,
plotted with open symbols

00:46:18.820 --> 00:46:22.270
here, for the kind
of tuning curve

00:46:22.270 --> 00:46:23.765
with one tone we had before.

00:46:25.830 --> 00:46:29.640
So everything within this
white area is excitatory.

00:46:29.640 --> 00:46:34.891
You put a frequency of 7
kilohertz in at 40 dB SPL.

00:46:34.891 --> 00:46:37.515
And the neuron is going to fire
all kinds of action potentials.

00:46:39.590 --> 00:46:43.390
Now, let's put in a tone
right at this triangle,

00:46:43.390 --> 00:46:45.160
called the probe tone.

00:46:45.160 --> 00:46:47.052
It's usually right at the CF.

00:46:47.052 --> 00:46:49.256
And it's above the threshold.

00:46:51.100 --> 00:46:53.770
In this case, it looks
like it's about 25 dB.

00:46:53.770 --> 00:46:55.855
And it gets the
neuron responding.

00:46:56.880 --> 00:47:01.260
You put that probe
tone in the neuron

00:47:01.260 --> 00:47:03.850
is going to fire some
action potentials.

00:47:03.850 --> 00:47:05.940
And keep that probe
tone in so the neuron

00:47:05.940 --> 00:47:07.570
is firing action potentials.

00:47:07.570 --> 00:47:09.400
And put a second tone in.

00:47:10.570 --> 00:47:14.645
And the second tone is often
outside the response areas.

00:47:16.680 --> 00:47:20.560
And it turns out that
anywhere in this shaded area

00:47:20.560 --> 00:47:25.540
above the CF or below
the CF, a second tone,

00:47:25.540 --> 00:47:28.990
as is illustrated here,
will decrease the response

00:47:28.990 --> 00:47:31.150
to the probe tone in
a dramatic fashion.

00:47:32.560 --> 00:47:35.480
Then, when you turn off
this second tone-- sometimes

00:47:35.480 --> 00:47:40.395
called a suppressing tone-- the
original activity comes back.

00:47:42.350 --> 00:47:46.245
And this phenomenon is
called two-tone suppression.

00:47:47.750 --> 00:47:50.170
At first, it was called
two-tone inhibition.

00:47:50.170 --> 00:47:55.330
People thought, oh, OK, there's
another nearby neighbor nerve

00:47:55.330 --> 00:47:57.440
fiber that's inhibiting
this first one.

00:47:57.440 --> 00:47:58.600
And they looked in
the cochlea and there

00:47:58.600 --> 00:47:59.933
weren't any inhibitory synapses.

00:48:01.880 --> 00:48:03.990
OK, so that was a problem.

00:48:03.990 --> 00:48:05.560
They started calling
it suppression.

00:48:05.560 --> 00:48:07.690
And they actually
ended up finding it

00:48:07.690 --> 00:48:10.730
in the movement of
the basilar membrane.

00:48:10.730 --> 00:48:13.210
So it's just something
about the vibration pattern

00:48:13.210 --> 00:48:17.770
of the cochlea that causes
the movement of the membranes

00:48:17.770 --> 00:48:20.200
to be diminished
by a second tone

00:48:20.200 --> 00:48:21.450
on either side of the first.

00:48:24.930 --> 00:48:26.810
Now, why do I bring this up?

00:48:26.810 --> 00:48:29.465
Well, it's kind of interesting
in a number of contexts.

00:48:30.520 --> 00:48:33.365
Two-tone suppression might
be a form of gain control.

00:48:34.660 --> 00:48:37.440
If you just had the
excitatory tuning curve,

00:48:37.440 --> 00:48:40.760
and you started listening in a
restaurant where everybody was

00:48:40.760 --> 00:48:43.450
talking and there
was a lot of sound,

00:48:43.450 --> 00:48:46.040
all your auditory
nerve fibers might

00:48:46.040 --> 00:48:49.020
be discharging at
their maximal rates.

00:48:49.020 --> 00:48:52.619
And you wouldn't be able to tell
the interesting conversation

00:48:52.619 --> 00:48:53.910
your two neighbors were having.

00:48:53.910 --> 00:48:55.330
You wouldn't be
able to eavesdrop.

00:48:55.330 --> 00:48:57.595
You wouldn't be having
a conversation yourself.

00:48:58.850 --> 00:49:01.770
So two-tone suppression
is a form of gain control,

00:49:01.770 --> 00:49:07.990
where the side bands reduce
the response to the main band,

00:49:07.990 --> 00:49:10.405
so that not everything's
being driven into saturation.

00:49:11.970 --> 00:49:12.850
That's one reason.

00:49:14.060 --> 00:49:16.270
And a second reason
is you can actually

00:49:16.270 --> 00:49:21.490
use this in a sort of a
tricky psychophysical paradigm

00:49:21.490 --> 00:49:26.060
to measure the tuning
of human listeners.

00:49:26.060 --> 00:49:29.950
We obviously can't go into
a human's auditory nerve

00:49:29.950 --> 00:49:31.910
with a microelectrode,
although it's

00:49:31.910 --> 00:49:33.580
been done a couple
times in surgery.

00:49:33.580 --> 00:49:34.260
But it's rare.

00:49:36.610 --> 00:49:41.680
It's easy to do a so-called
two-tone suppression paradigm,

00:49:41.680 --> 00:49:44.520
where you have the person
listen to the probe tone.

00:49:44.520 --> 00:49:46.490
You say to the
listener, here's a tone.

00:49:46.490 --> 00:49:48.650
I want you to listen to that.

00:49:48.650 --> 00:49:50.340
I'm going to put
in a second tone.

00:49:50.340 --> 00:49:51.266
I ignore that.

00:49:51.266 --> 00:49:52.140
Don't worry about it.

00:49:52.140 --> 00:49:54.170
Just listen to that
first probe tone.

00:49:56.650 --> 00:49:59.060
Tell me if you can hear
that original probe tone.

00:49:59.060 --> 00:50:00.220
Ah, yeah, sure I can hear.

00:50:00.220 --> 00:50:01.511
I'm going to put a second tone.

00:50:01.511 --> 00:50:03.395
Oh, I can't hear the
probe tone anymore.

00:50:04.720 --> 00:50:08.760
OK the second or side tone
has suppressed the response

00:50:08.760 --> 00:50:10.070
to the probe tone.

00:50:10.070 --> 00:50:12.200
And you can use that
as a measure of tuning,

00:50:12.200 --> 00:50:16.390
because these suppression areas
flank the excitatory area.

00:50:16.390 --> 00:50:20.140
And so here are some
results from humans

00:50:20.140 --> 00:50:24.480
in a so-called psychophysical
tuning curve paradigm.

00:50:25.660 --> 00:50:29.540
And these are a half a
dozen or so tuning curves.

00:50:29.540 --> 00:50:35.175
Each one has associated with
it a probe tone or a test tone.

00:50:36.360 --> 00:50:39.910
The task is listen
to that test tone

00:50:39.910 --> 00:50:43.965
and tell me if you still
hear it or if it's gone away.

00:50:45.070 --> 00:50:47.960
The experimenter
introduce a second tone,

00:50:47.960 --> 00:50:51.965
a so-called masker, at those
frequencies and levels.

00:50:53.490 --> 00:50:57.700
And at where the line
is drawn, the person

00:50:57.700 --> 00:50:59.850
who's listening
to the probe tone

00:50:59.850 --> 00:51:02.800
says, I can't hear that
probe tone anymore.

00:51:02.800 --> 00:51:04.620
Something happened to it.

00:51:04.620 --> 00:51:06.990
Well, two-tone suppression
happened to it.

00:51:06.990 --> 00:51:11.530
The masker masked the response
to the probe tone or test tone.

00:51:13.554 --> 00:51:15.470
And look at the shapes
of those tuning curves.

00:51:15.470 --> 00:51:18.050
They look like good old
auditory nerve fibers.

00:51:18.050 --> 00:51:19.350
They have a CF.

00:51:19.350 --> 00:51:21.940
The CF is right at the probe.

00:51:21.940 --> 00:51:23.125
They have a tip region.

00:51:24.240 --> 00:51:25.690
They have a tail region.

00:51:25.690 --> 00:51:30.120
If you measure the
sharpness, how wide they are,

00:51:30.120 --> 00:51:31.960
they're really
sharp at high CFs.

00:51:31.960 --> 00:51:35.445
And they get a little bit
broader as a CFs goes down.

00:51:36.830 --> 00:51:39.230
At high CFs they have
a tip and a tail.

00:51:39.230 --> 00:51:41.530
At low CFs they look
more like v-shape.

00:51:44.510 --> 00:51:48.430
We can go back to the
auditory nerve tuning curve

00:51:48.430 --> 00:51:49.530
with those in mind.

00:51:54.580 --> 00:51:55.970
And look how similar they are.

00:51:55.970 --> 00:52:00.690
Here's high CF,
tip and the tail.

00:52:00.690 --> 00:52:03.620
Low CF, just sort of a
plain v ad they're wider.

00:52:05.680 --> 00:52:08.560
Human psychophysical
tuning curves

00:52:08.560 --> 00:52:10.753
have that same general look.

00:52:14.750 --> 00:52:17.400
Now, remember, this is a
very different paradigm.

00:52:17.400 --> 00:52:19.770
Here there are two tones.

00:52:19.770 --> 00:52:21.670
The probe tone is one of them.

00:52:21.670 --> 00:52:24.870
And the masker or the
second suppressor tone

00:52:24.870 --> 00:52:26.200
is the second one.

00:52:26.200 --> 00:52:29.110
Whereas in good old fashioned
auditory nerve fiber

00:52:29.110 --> 00:52:34.160
tuning curve there was just
one, the excitatory tone.

00:52:36.550 --> 00:52:38.205
OK, so psychophysical
tuning curves

00:52:38.205 --> 00:52:40.980
are obtained from humans
in the following paradigm.

00:52:40.980 --> 00:52:42.860
We went over that.

00:52:42.860 --> 00:52:44.860
These tuning curves and
the neural tuning curves

00:52:44.860 --> 00:52:46.330
from animals are
roughly similar.

00:52:49.950 --> 00:52:54.240
Now, what would you expect
to happen to these tuning

00:52:54.240 --> 00:52:59.370
curves and the
neural tuning curves

00:52:59.370 --> 00:53:01.175
if you had an outer
hair cell problem?

00:53:15.630 --> 00:53:17.430
And this is kind of
the classic-- oh, yeah,

00:53:17.430 --> 00:53:20.490
you can sort of pass that
around-- a classic exam

00:53:20.490 --> 00:53:21.430
question.

00:53:21.430 --> 00:53:23.040
Draw a tuning curve.

00:53:23.040 --> 00:53:24.840
So you label this
with frequency.

00:53:27.560 --> 00:53:35.290
This is the sound pressure level
for a response-- I don't know.

00:53:35.290 --> 00:53:40.806
We can say whatever response you
want to-- 10 spikes per second.

00:53:45.490 --> 00:53:46.950
Label the CF.

00:53:46.950 --> 00:53:47.510
Here it is.

00:53:49.300 --> 00:53:50.010
This is a normal.

00:53:53.310 --> 00:53:54.390
What's the axis here?

00:53:54.390 --> 00:53:58.150
Well, the CF might be-- the
threshold might be at 0 dB.

00:53:59.400 --> 00:54:02.290
The tail comes in-- let's go
to our animal tuning curve

00:54:02.290 --> 00:54:03.420
just so we get this right.

00:54:05.344 --> 00:54:07.270
Oops, pressed the wrong button.

00:54:08.990 --> 00:54:13.630
OK, so the tip on this
one it's about 20.

00:54:13.630 --> 00:54:16.340
The tail is coming in about 60.

00:54:16.340 --> 00:54:19.330
So we are starting down--
well, let's say it's 20.

00:54:19.330 --> 00:54:25.055
This is going to be
60 dBs SPL-- normal.

00:54:26.760 --> 00:54:32.210
Draw the tuning curve in an
animal where the outer hair

00:54:32.210 --> 00:54:33.075
cells are damaged.

00:54:35.807 --> 00:54:37.515
Well, you could say,
there's no response.

00:54:37.515 --> 00:54:39.060
That wouldn't be quite right.

00:54:42.480 --> 00:54:51.300
OK, remember we're-- this is
the nerve fiber we're recording

00:54:51.300 --> 00:54:53.280
from, a type one.

00:54:53.280 --> 00:54:55.920
This is the inner hair cell.

00:54:55.920 --> 00:54:57.320
These are the outer hair cells.

00:55:05.450 --> 00:55:08.230
And we're saying damage
them, lesion them.

00:55:09.370 --> 00:55:11.810
You could have it
in a knockout animal

00:55:11.810 --> 00:55:13.745
where they had
lost their Preston.

00:55:18.690 --> 00:55:21.395
OK, so the cochlear
amplifier is lost.

00:55:23.070 --> 00:55:24.549
What sort of a
hearing loss do have

00:55:24.549 --> 00:55:26.090
when you lose the
cochlear amplifier?

00:55:29.670 --> 00:55:31.340
40 to 60 dB, right?

00:55:32.670 --> 00:55:34.510
Well, what's this interval?

00:55:34.510 --> 00:55:35.750
40 dB, right?

00:55:36.980 --> 00:55:40.180
And it turns out,
when you record

00:55:40.180 --> 00:55:47.760
from a preparation in which the
outer hair cells are lesioned,

00:55:47.760 --> 00:55:55.090
this is the kind of tuning curve
you find when the outer hair

00:55:55.090 --> 00:55:59.720
cells are killed or lesioned--
a tip-less tuning curve.

00:55:59.720 --> 00:56:03.842
At least from these high
frequencies that have a tip.

00:56:03.842 --> 00:56:07.260
And the lows they
look more bowl shaped.

00:56:07.260 --> 00:56:09.495
But there's a 40 to
60 dB hearing loss.

00:56:09.495 --> 00:56:13.180
You're not deaf, but you have
a greatly altered function.

00:56:15.610 --> 00:56:20.400
How good would this
function be for telling

00:56:20.400 --> 00:56:28.230
the difference between
1,000 hertz and 1,002 hertz?

00:56:28.230 --> 00:56:30.040
Not so good, right?

00:56:30.040 --> 00:56:34.020
You need a very
sharply tuned function

00:56:34.020 --> 00:56:39.350
to tell or discriminate between
two closely spaced frequencies.

00:56:39.350 --> 00:56:41.680
If you have an outer
hair cell problem,

00:56:41.680 --> 00:56:45.090
not only are your going
to be much less sensitive,

00:56:45.090 --> 00:56:47.612
but you're not going to be
so good at distinguishing

00:56:47.612 --> 00:56:48.445
between frequencies.

00:56:50.730 --> 00:56:53.100
Another way to think
about it is that if there

00:56:53.100 --> 00:56:59.190
were a whole bunch of
frequencies down here

00:56:59.190 --> 00:57:01.680
and your hearing
aid boosted them,

00:57:01.680 --> 00:57:04.290
you wouldn't be able to
listen to your characteristic

00:57:04.290 --> 00:57:08.250
frequency anymore, because
these side frequencies were

00:57:08.250 --> 00:57:10.570
getting into your response area.

00:57:10.570 --> 00:57:19.020
So these are non-selective
response areas,

00:57:19.020 --> 00:57:22.800
where the normal or sharply
tuned are very selective.

00:57:24.917 --> 00:57:26.250
And what are they selective for?

00:57:26.250 --> 00:57:28.795
For sound frequency.

00:57:35.520 --> 00:57:37.980
OK, so the outer
hair cells give you

00:57:37.980 --> 00:57:42.490
this big boost in sensitivity
and sharp tuning of the tip.

00:57:43.530 --> 00:57:46.175
That's the cochlear amplifier
part of the function.

00:57:48.540 --> 00:57:50.900
OK, now, how could we do this?

00:57:50.900 --> 00:57:52.624
Well, recently, within
the last 10 years,

00:57:52.624 --> 00:57:54.290
you can have a
[? knocked ?] out animal.

00:57:55.540 --> 00:57:58.180
But in the old days, you
could lesion outer hair cells

00:57:58.180 --> 00:57:59.840
by many means.

00:57:59.840 --> 00:58:02.950
You could lesion
them by loud sounds.

00:58:02.950 --> 00:58:04.580
Well, loud sounds
actually end up

00:58:04.580 --> 00:58:07.780
affecting inner hair cells
a little bit as well.

00:58:07.780 --> 00:58:10.800
So the preferred method of
lesioning outer hair cells

00:58:10.800 --> 00:58:13.070
was with drugs.

00:58:13.070 --> 00:58:18.646
For example, kanamycin is
a very good antibiotic.

00:58:20.410 --> 00:58:21.650
It kills bacteria.

00:58:21.650 --> 00:58:22.900
Unfortunately, it's audatoxic.

00:58:22.900 --> 00:58:25.440
It kills hair cells.

00:58:25.440 --> 00:58:28.380
And if you give it to animals
in just the right dose,

00:58:28.380 --> 00:58:31.620
you can kill the outer hair
cells, which for some reason--

00:58:31.620 --> 00:58:34.710
it's not known-- are
more sensitive to them.

00:58:34.710 --> 00:58:36.140
If you give them
a higher dose, it

00:58:36.140 --> 00:58:38.320
will also kill the
inner hair cell.

00:58:38.320 --> 00:58:40.680
But you can create
animal preparations

00:58:40.680 --> 00:58:43.490
in which the outer
hair cells are gone

00:58:43.490 --> 00:58:45.350
and the inner hassles
are remaining,

00:58:45.350 --> 00:58:48.560
at least over a particular
part of the cochlea.

00:58:48.560 --> 00:58:51.181
And from that part, you can
record these tip-less tuning

00:58:51.181 --> 00:58:51.680
curves.

00:58:58.730 --> 00:59:02.770
OK, so that is
mostly what I want

00:59:02.770 --> 00:59:05.680
to say about place coding
for sound frequency.

00:59:07.320 --> 00:59:11.230
And now, I want to get into the
second code for sound frequency

00:59:11.230 --> 00:59:15.190
that we have, which is
a temporal code that's

00:59:15.190 --> 00:59:20.320
based on the finding
of temporal synchrony

00:59:20.320 --> 00:59:21.540
in the auditory nerve.

00:59:23.110 --> 00:59:24.890
This is the so-called
phase-locking.

00:59:26.600 --> 00:59:30.270
Again, we're doing the same kind
of experimental preparation.

00:59:30.270 --> 00:59:33.630
We stick are recording
electrode in the auditory nerve.

00:59:33.630 --> 00:59:37.650
And we record from one
single auditory nerve fiber.

00:59:37.650 --> 00:59:39.670
And we measure it's spikes.

00:59:39.670 --> 00:59:41.490
Each one of these
little blips is a spike.

00:59:44.350 --> 00:59:48.120
The very top trace is
the sound wave form.

00:59:48.120 --> 00:59:51.210
The next trace is the response
of the auditory nerve fiber.

00:59:51.210 --> 00:59:55.070
And these are supra-imposed
multiple traces.

00:59:55.070 --> 00:59:57.780
And that trace is
with no stimulus.

00:59:57.780 --> 01:00:02.330
So this auditory nerve fiber
is obviously very happy firing

01:00:02.330 --> 01:00:03.850
long, spontaneous activity.

01:00:05.940 --> 01:00:07.630
Then let's turn the sound on.

01:00:07.630 --> 01:00:09.990
The top trace is on now.

01:00:09.990 --> 01:00:11.540
This is with the stimulus.

01:00:13.480 --> 01:00:18.440
And look at how these auditory
nerve fiber impulses tend

01:00:18.440 --> 01:00:23.540
to line up at a particular
phase of the sound stimulus.

01:00:25.170 --> 01:00:26.190
What's phase?

01:00:26.190 --> 01:00:39.230
Well, it's just the
degrees, the sine wave,

01:00:39.230 --> 01:00:44.860
as a function of time,
the sound pressure--

01:00:44.860 --> 01:00:51.590
this is sound
pressure-- and it's

01:00:51.590 --> 01:01:00.220
going through 360 degrees of
phase here-- 180 degrees here.

01:01:00.220 --> 01:01:03.490
And it looks like
many of the spikes

01:01:03.490 --> 01:01:10.580
are lining up around
80 degree point.

01:01:10.580 --> 01:01:13.280
So a lot of the
firing is right here.

01:01:13.280 --> 01:01:15.290
Not so much firing here.

01:01:15.290 --> 01:01:16.660
Not so much firing here.

01:01:17.670 --> 01:01:19.275
And then another
waveform comes along

01:01:19.275 --> 01:01:22.290
and you get some more
firing about the same time.

01:01:22.290 --> 01:01:25.740
Now, one very
common misconception

01:01:25.740 --> 01:01:29.930
about phase-locking is that
every time the sound wave form

01:01:29.930 --> 01:01:33.680
goes through-- in
this case 80 degrees--

01:01:33.680 --> 01:01:35.230
the fiber fires an impulse.

01:01:35.230 --> 01:01:36.830
That's not true at all.

01:01:36.830 --> 01:01:40.040
Here is a single trace, showing
excellent phase-locking.

01:01:41.920 --> 01:01:45.780
And there's a response
to the first wave form.

01:01:45.780 --> 01:01:48.650
But then the fiber takes
a break and doesn't

01:01:48.650 --> 01:01:49.830
respond during the second.

01:01:51.420 --> 01:01:54.680
And it looks like it responds
on the third and the fourth.

01:01:55.720 --> 01:01:57.750
But then it takes a
longer break and doesn't

01:01:57.750 --> 01:02:02.640
respond at the fifth or sixth,
but it responds at the seventh,

01:02:02.640 --> 01:02:06.190
and not at the eighth or ninth,
then on the 10th and 11th.

01:02:06.190 --> 01:02:07.450
So it doesn't matter.

01:02:07.450 --> 01:02:10.710
You don't have to respond
in every single waveform.

01:02:10.710 --> 01:02:13.570
You can respond in
one wave form and take

01:02:13.570 --> 01:02:19.080
a break for 100 waveforms,
as long as when you respond,

01:02:19.080 --> 01:02:22.100
the next time it's
on the same point

01:02:22.100 --> 01:02:25.285
or in the same phase
in the sound wave.

01:02:27.820 --> 01:02:31.070
So typically, to
get these data, you

01:02:31.070 --> 01:02:33.260
average over many
hundreds or even

01:02:33.260 --> 01:02:37.660
thousands of stimulus cycles,
where one complete cycle

01:02:37.660 --> 01:02:39.865
is 0 to 360 degrees.

01:02:41.720 --> 01:02:47.570
These are plots of
auditory nerve firing.

01:02:47.570 --> 01:02:51.590
So this is a firing rate access
percent of total impulses.

01:02:53.120 --> 01:02:55.030
This is now a time axis.

01:02:57.060 --> 01:02:59.810
So we're just saying when
does it fire along the time.

01:03:01.200 --> 01:03:03.290
And the stimulus,
I believe, here

01:03:03.290 --> 01:03:07.370
is 1,000 hertz, so it's the
middle of the hearing range.

01:03:07.370 --> 01:03:08.870
And this is excellent
phase-locking.

01:03:12.770 --> 01:03:15.110
If you were to
quantify this-- there

01:03:15.110 --> 01:03:16.860
are many ways to
quantify this-- but you

01:03:16.860 --> 01:03:22.140
could fit, for example, a
Fourier series, to that.

01:03:22.140 --> 01:03:24.530
And you could plot
just the fundamental

01:03:24.530 --> 01:03:26.340
of the Fourier series.

01:03:26.340 --> 01:03:29.030
And that's what's known as the
synchronization coefficient.

01:03:31.530 --> 01:03:33.480
And plot it as a
function of frequency.

01:03:34.520 --> 01:03:36.800
You could make your
measurements at 1,000 hertz,

01:03:36.800 --> 01:03:38.175
which is this
point on the graph.

01:03:39.280 --> 01:03:41.260
You could make them
at 5,000 hertz.

01:03:41.260 --> 01:03:43.025
You could make
them at 500 hertz.

01:03:45.320 --> 01:03:47.630
This synchronization
coefficient ends up

01:03:47.630 --> 01:03:50.740
being between 0.8 and
0.9 for low frequencies.

01:03:50.740 --> 01:03:53.370
And then it rolls
off essentially

01:03:53.370 --> 01:03:57.560
to be random firing at
around 3,000 or 4,000,

01:03:57.560 --> 01:04:00.100
certainly by 5,000 hertz.

01:04:00.100 --> 01:04:03.970
So this behavior,
this phase-locking

01:04:03.970 --> 01:04:07.205
goes away toward the high
end of our hearing range.

01:04:08.910 --> 01:04:14.360
It just means that the auditory
nerve can no longer synchronize

01:04:14.360 --> 01:04:15.910
at very high frequency.

01:04:15.910 --> 01:04:17.300
So what's going on here?

01:04:24.570 --> 01:04:28.400
The auditory nerve fiber
is getting its messages

01:04:28.400 --> 01:04:29.760
from the hair cell, right?

01:04:31.890 --> 01:04:34.160
Here's the auditory nerve fiber.

01:04:34.160 --> 01:04:36.172
And it's hooked up to
an inner hair cell.

01:04:41.260 --> 01:04:42.380
And it's sending messages.

01:04:43.700 --> 01:04:44.628
What are the messages?

01:04:44.628 --> 01:04:45.336
Neurotransmitter.

01:04:48.750 --> 01:04:51.600
When the wave form
goes like this,

01:04:51.600 --> 01:04:53.640
the auditory nerve
fiber is responding.

01:04:53.640 --> 01:04:55.480
Ah, it's getting lots
of neurotransmitter.

01:04:57.950 --> 01:05:00.870
Well, that was when
the stereocilia

01:05:00.870 --> 01:05:02.180
were bent one direction.

01:05:03.910 --> 01:05:06.040
Ions flowed in.

01:05:07.400 --> 01:05:09.069
The inner hair cell
was depolarized.

01:05:09.069 --> 01:05:10.610
It released lots of
neurotransmitter.

01:05:12.930 --> 01:05:16.550
Let's go a little
bit longer in time

01:05:16.550 --> 01:05:18.580
to this bottom part
of the phase curve.

01:05:19.770 --> 01:05:23.850
The stereocilia were bent
the opposite direction.

01:05:23.850 --> 01:05:25.645
The ion channels closed off.

01:05:26.760 --> 01:05:30.340
The inner hair cell went back to
its rest-- minus 80 millivolts,

01:05:30.340 --> 01:05:30.870
let's say.

01:05:31.960 --> 01:05:34.090
And it said, I'm
not excited anymore.

01:05:34.090 --> 01:05:38.440
I'm going to shut off the
flow of neurotransmitter.

01:05:38.440 --> 01:05:42.739
The auditory nerve fiber
goes, oh, we're quiet.

01:05:42.739 --> 01:05:43.780
We don't need to respond.

01:05:46.150 --> 01:05:50.680
Go back to the other direction,
then the stereocilia back

01:05:50.680 --> 01:05:51.850
the other way.

01:05:51.850 --> 01:05:53.340
Ah, I'm depolarized.

01:05:53.340 --> 01:05:55.492
I'm going to go to
minus 30 millivolts.

01:05:55.492 --> 01:05:57.200
Ah, well, let's release
neurotransmitter.

01:05:59.150 --> 01:06:01.100
Oh, wow, there's
something going on.

01:06:01.100 --> 01:06:02.600
I'm going to fire.

01:06:02.600 --> 01:06:05.060
I'm going to fire all
these action potentials.

01:06:05.060 --> 01:06:06.870
It's going back and
forth, back and forth.

01:06:08.420 --> 01:06:12.120
At some point though, this
is going back and forth

01:06:12.120 --> 01:06:16.710
so fast that this just
gets to be a blur.

01:06:18.330 --> 01:06:19.700
There is a sound there.

01:06:19.700 --> 01:06:22.040
It's depolarizing the hair cell.

01:06:22.040 --> 01:06:25.250
But it can't do this
push pull kind of thing.

01:06:25.250 --> 01:06:27.140
It's not fast enough.

01:06:27.140 --> 01:06:29.050
Even though there's a
nice synaptic ribbon

01:06:29.050 --> 01:06:32.070
there to coordinate the
release of the vesicles,

01:06:32.070 --> 01:06:33.790
it gets overwhelmed.

01:06:33.790 --> 01:06:38.730
Remember, at 1,000 hertz,
this is going back and forth

01:06:38.730 --> 01:06:39.480
in 1 millisecond.

01:06:41.050 --> 01:06:43.670
And 5,000 hertz, it's
going back and forth

01:06:43.670 --> 01:06:46.050
five times in 1 millisecond.

01:06:46.050 --> 01:06:47.381
That's pretty fast.

01:06:47.381 --> 01:06:48.380
And it gets overwhelmed.

01:06:49.900 --> 01:06:50.960
There's a response.

01:06:50.960 --> 01:06:54.428
There's more action potentials
with the stimulus than without.

01:06:54.428 --> 01:06:55.886
But they're no
longer synchronized.

01:06:57.130 --> 01:06:57.970
It gets overwhelmed.

01:06:57.970 --> 01:06:59.615
And phase-lacking goes away.

01:07:01.330 --> 01:07:05.960
We can distinguish
5,000 from 6,000 hertz

01:07:05.960 --> 01:07:07.465
very nicely when we listen.

01:07:08.690 --> 01:07:11.260
We're not using this
code, because there's

01:07:11.260 --> 01:07:15.440
no temporal synchrony in
the auditory nerve at very

01:07:15.440 --> 01:07:16.950
high frequencies.

01:07:16.950 --> 01:07:22.560
This is a kind of an interesting
code for sound frequency,

01:07:22.560 --> 01:07:25.060
because the timing
is going to be

01:07:25.060 --> 01:07:26.710
different for
different frequencies.

01:07:28.190 --> 01:07:38.220
Imagine at low frequencies--
and imagine just

01:07:38.220 --> 01:07:41.130
for the sake of argument--
that the auditory nerve

01:07:41.130 --> 01:07:46.280
fiber is going to respond on
every single stimulus peak.

01:07:46.280 --> 01:07:52.700
Let's say this is 1,000 hertz.

01:07:52.700 --> 01:08:00.040
And now let's say we dial
in 2,000 hertz, which

01:08:00.040 --> 01:08:04.830
is going to end up
going twice as fast.

01:08:04.830 --> 01:08:07.020
I'm not a very
good artists here.

01:08:07.020 --> 01:08:09.900
But you can imagine
that the firing

01:08:09.900 --> 01:08:15.020
is going to be twice as often,
if for the sake of argument

01:08:15.020 --> 01:08:20.410
we're firing in every stimulus
frequency, which may not

01:08:20.410 --> 01:08:21.340
happen.

01:08:21.340 --> 01:08:23.200
But this is kind of
an interesting code.

01:08:23.200 --> 01:08:25.370
Because if you're
sitting in the brain

01:08:25.370 --> 01:08:28.600
and you're getting
firing very far apart,

01:08:28.600 --> 01:08:31.094
you're going to say, OK,
that's a low frequency.

01:08:32.600 --> 01:08:34.699
But if you're getting
firing very close together,

01:08:34.699 --> 01:08:38.490
you're going to say, oh,
that's a higher frequency.

01:08:38.490 --> 01:08:42.210
So is there some little
detector in the brain

01:08:42.210 --> 01:08:44.370
that's detecting
these intervals?

01:08:44.370 --> 01:08:46.470
How fast the firing?

01:08:46.470 --> 01:08:48.180
Well, we don't know that.

01:08:48.180 --> 01:08:49.920
But we certainly
know that a code

01:08:49.920 --> 01:08:56.240
is available in the auditory
nerve at low frequencies,

01:08:56.240 --> 01:08:59.319
but not at high frequencies
like 5 kilohertz.

01:08:59.319 --> 01:09:02.020
So what's the
evidence that we're

01:09:02.020 --> 01:09:04.010
using one code or the other?

01:09:04.010 --> 01:09:08.660
Clearly, the place
code has to provide us

01:09:08.660 --> 01:09:10.899
with frequency information
at a higher frequency.

01:09:10.899 --> 01:09:14.890
There is no temporal code
at those high frequencies.

01:09:14.890 --> 01:09:17.130
Down low, which code do we use?

01:09:17.130 --> 01:09:18.610
Well, we probably use both.

01:09:20.790 --> 01:09:22.919
That's another way of
saying, I'm not really sure.

01:09:25.890 --> 01:09:30.810
But let me give you some data
from musical intervals that

01:09:30.810 --> 01:09:33.616
might suggest that
this time code is used.

01:09:35.000 --> 01:09:37.290
What are the data?

01:09:37.290 --> 01:09:40.380
We have to talk a little
bit about perception

01:09:40.380 --> 01:09:41.290
of musical intervals.

01:09:44.960 --> 01:09:47.800
And we might as well start
out with the most important

01:09:47.800 --> 01:09:50.983
musical interval,
which is the octave.

01:09:52.210 --> 01:09:55.200
Does everybody know
what an octave is?

01:09:55.200 --> 01:09:57.704
Yeah, what it is an octave?

01:09:59.950 --> 01:10:01.830
I can't explain
it, but I know it.

01:10:04.000 --> 01:10:04.960
What about on a piano?

01:10:04.960 --> 01:10:09.045
You go down and hit middle
C, where is the octave?

01:10:09.045 --> 01:10:10.382
AUDIENCE: The next C.

01:10:10.382 --> 01:10:11.590
PROFESSOR: The next C, right.

01:10:11.590 --> 01:10:13.920
You've even called
it the same letter,

01:10:13.920 --> 01:10:15.675
because it sounds so similar.

01:10:16.740 --> 01:10:20.200
But in precise physical
terms, an octave

01:10:20.200 --> 01:10:21.435
is a doubling of frequency.

01:10:23.330 --> 01:10:27.720
Whatever frequency middle C was,
if you double that frequency,

01:10:27.720 --> 01:10:31.310
you get an octave
above middle C.

01:10:31.310 --> 01:10:36.170
So we have some data
here for two intervals,

01:10:36.170 --> 01:10:40.310
one 440 hertz-- two
frequencies, one 440 hertz

01:10:40.310 --> 01:10:42.330
and another an
octave above, 880.

01:10:42.330 --> 01:10:43.760
So double it.

01:10:43.760 --> 01:10:48.750
And why did we pick 440 hertz?

01:10:48.750 --> 01:10:54.539
So that corresponds to a
note-- just-- yeah, right,

01:10:54.539 --> 01:10:56.080
it corresponds to
A. And I was trying

01:10:56.080 --> 01:10:59.460
to think if the A is
below or above middle C. I

01:10:59.460 --> 01:11:00.670
think it's above.

01:11:00.670 --> 01:11:03.600
So what's important-- you
guys knew that right away.

01:11:03.600 --> 01:11:04.770
What's important about that?

01:11:04.770 --> 01:11:06.805
AUDIENCE: Orchestras tune to it.

01:11:06.805 --> 01:11:08.180
PROFESSOR: Orchestras
tune to it.

01:11:08.180 --> 01:11:10.540
So can you give it to me?

01:11:11.880 --> 01:11:17.900
OK, I'll give it to you.
[WHISTLES] Sorry, that's A 440.

01:11:17.900 --> 01:11:20.240
And so here's A
440 on the violin.

01:11:20.240 --> 01:11:24.150
[PLUCKS A NOTE] OK,
now, how do I know that?

01:11:25.831 --> 01:11:27.080
Because orchestras tune to it.

01:11:28.710 --> 01:11:32.340
So for about 20 years,
I sat in an orchestra.

01:11:32.340 --> 01:11:36.800
And the first thing you did--
[LAUGHS] OK, tune, you guys.

01:11:36.800 --> 01:11:38.790
And what instrument
gives the tuning note?

01:11:38.790 --> 01:11:41.310
If you're in junior high, it's
this little electronic thing.

01:11:41.310 --> 01:11:44.760
But if you're in the BSO, what
instrument gives the tuning

01:11:44.760 --> 01:11:45.468
note?

01:11:45.468 --> 01:11:46.472
AUDIENCE: The violin.

01:11:46.472 --> 01:11:48.513
PROFESSOR: No, violins go
out of tune like crazy.

01:11:48.513 --> 01:11:49.400
AUDIENCE: Oboe.

01:11:49.400 --> 01:11:51.150
PROFESSOR: Oboe, right,
because the oboe's

01:11:51.150 --> 01:11:52.990
a very stable instrument.

01:11:52.990 --> 01:11:55.030
And if the barometric
pressure goes up

01:11:55.030 --> 01:11:56.870
and the humidity
goes down, the oboe's

01:11:56.870 --> 01:11:59.550
still going to give you A 440.

01:11:59.550 --> 01:12:04.840
So the A 440 is a very
important musical note.

01:12:04.840 --> 01:12:07.040
And all these
instruments, of course,

01:12:07.040 --> 01:12:08.519
have a whole bunch of harmonics.

01:12:08.519 --> 01:12:11.060
This string is vibrating in a
whole bunch of different modes.

01:12:12.380 --> 01:12:17.140
But the fundamental, the length
where the whole string vibrates

01:12:17.140 --> 01:12:18.490
is A 440.

01:12:18.490 --> 01:12:22.020
OK, so here's A 440
or approximately.

01:12:26.300 --> 01:12:29.350
Now, an octave above that
is a very nice sounds.

01:12:29.350 --> 01:12:36.500
It's another A. That's
A 880, the fundamental.

01:12:36.500 --> 01:12:42.235
And if I sound them together,
they sound very beautiful.

01:12:43.720 --> 01:12:46.570
And in any musical
culture, an octave

01:12:46.570 --> 01:12:49.240
is a very predominant
interval, because it

01:12:49.240 --> 01:12:51.620
sounds so wonderful to your ear.

01:12:53.030 --> 01:12:56.520
And violinists, I can tell you
from experience, practice a lot

01:12:56.520 --> 01:13:00.140
of time trying to tune
their octaves perfectly.

01:13:00.140 --> 01:13:02.710
And if you've ever listened to
a professionals go like this,

01:13:02.710 --> 01:13:04.510
and every time they
go up and down,

01:13:04.510 --> 01:13:06.690
the octave is just beautiful.

01:13:06.690 --> 01:13:10.395
But if you've been to middle
school or elementary school,

01:13:10.395 --> 01:13:11.395
it's a little different.

01:13:13.010 --> 01:13:16.710
Because sometimes when those
students play an octave,

01:13:16.710 --> 01:13:20.090
it doesn't really hit
to be exactly an octave.

01:13:20.090 --> 01:13:22.850
And now I'm going to give
you a demonstration that's

01:13:22.850 --> 01:13:26.320
440 and not quite 880.

01:13:26.320 --> 01:13:29.200
OK And it's not going to
sound exactly the same.

01:13:29.200 --> 01:13:30.162
So here it is.

01:13:34.900 --> 01:13:38.190
And that's an interval I've
listened to many times.

01:13:39.240 --> 01:13:40.910
But it's not a desired interval.

01:13:40.910 --> 01:13:42.530
It's a very dissonant interval.

01:13:43.790 --> 01:13:49.070
And what is terribly
displeasing about something

01:13:49.070 --> 01:13:52.760
that's not quite an
octave versus an octave.

01:13:53.850 --> 01:13:56.830
That is a question
that the place code

01:13:56.830 --> 01:13:59.350
has a lot of problems with.

01:13:59.350 --> 01:14:05.520
Because, for example, along
the cochlea there is a place--

01:14:05.520 --> 01:14:10.220
it's quite near the
apex-- for the 440.

01:14:10.220 --> 01:14:12.680
And then if you go
more basally, there's

01:14:12.680 --> 01:14:15.170
another place for the 880.

01:14:15.170 --> 01:14:18.684
And there's a place
for the 879 and 878.

01:14:18.684 --> 01:14:20.100
And those would
be very dissonant.

01:14:21.190 --> 01:14:26.980
But there's no reason that
those two things have any links

01:14:26.980 --> 01:14:29.240
to one another in
the place code.

01:14:29.240 --> 01:14:32.420
There's a place for 1,000
and a place for 2,000.

01:14:32.420 --> 01:14:34.440
Why do they sound so
wonderful together?

01:14:35.740 --> 01:14:38.940
The timing code though
has an answer for that.

01:14:38.940 --> 01:14:43.410
And here is some
data to show you

01:14:43.410 --> 01:14:46.060
why those two intervals
[? meld ?] very good together.

01:14:48.050 --> 01:14:50.270
If you look at
the spike pattern,

01:14:50.270 --> 01:14:54.560
in response to either
one of these frequencies,

01:14:54.560 --> 01:14:58.010
and compute what are
called the intervals

01:14:58.010 --> 01:15:02.840
between the spike-- so-called
interspike intervals--

01:15:02.840 --> 01:15:05.715
every time you get a spike,
you start your clock ticking.

01:15:07.320 --> 01:15:11.210
And that interval is timed
until the next spike fires.

01:15:11.210 --> 01:15:13.160
That's an interspike interval.

01:15:14.570 --> 01:15:17.530
And obviously, if
this is phase-locked,

01:15:17.530 --> 01:15:20.740
these intervals are going
to have a close relationship

01:15:20.740 --> 01:15:23.760
to the stimulus period.

01:15:23.760 --> 01:15:28.150
So here's a spike and here's
an approximately two-cycle

01:15:28.150 --> 01:15:29.710
interspike interval.

01:15:29.710 --> 01:15:34.470
Here's a short interval,
but it's one complete phase.

01:15:34.470 --> 01:15:37.660
Here's a long interval, but
it's now three complete phases.

01:15:38.670 --> 01:15:40.610
Here's another
three-phase interval.

01:15:40.610 --> 01:15:42.705
Here's a one-phase interval.

01:15:43.720 --> 01:15:48.450
You could make a very nice
plot of the interspike interval

01:15:48.450 --> 01:15:50.900
in milliseconds, the
time between the spikes.

01:15:52.430 --> 01:15:55.430
And these are the number of
occurrences on the y-axis.

01:15:55.430 --> 01:15:59.060
So for 440 hertz it's
the dashed curve here.

01:15:59.060 --> 01:16:03.180
And you get a big peak
here that's a multiple

01:16:03.180 --> 01:16:05.040
of the period.

01:16:05.040 --> 01:16:08.170
So at 440 hertz,
the sound wave form

01:16:08.170 --> 01:16:11.690
is taking about 2
and 1/2 milliseconds

01:16:11.690 --> 01:16:14.610
to go through one
complete cycle.

01:16:14.610 --> 01:16:18.830
And these intervals would be
firing on successive periods,

01:16:18.830 --> 01:16:21.050
which, obviously, the
nerve fiber can do.

01:16:22.170 --> 01:16:24.500
But sometimes they
take a break, and they

01:16:24.500 --> 01:16:27.290
fire only every other period.

01:16:27.290 --> 01:16:30.060
And that's a double
of this interval.

01:16:30.060 --> 01:16:33.120
And so you have a lot
of firing at about 4

01:16:33.120 --> 01:16:37.450
and 1/2 milliseconds, a lot of
firing at about 7 milliseconds,

01:16:37.450 --> 01:16:39.290
and so on and so forth.

01:16:39.290 --> 01:16:40.880
So this is an
interspike interval

01:16:40.880 --> 01:16:45.317
from auditory nerve firing in
response to this low frequency.

01:16:45.317 --> 01:16:46.650
Now, let's double the frequency.

01:16:48.830 --> 01:16:51.520
Now, the sound wave form
is going back and forth

01:16:51.520 --> 01:16:52.340
twice as fast.

01:16:54.180 --> 01:16:57.920
And you have-- no surprise--
firing, in some cases,

01:16:57.920 --> 01:17:00.500
twice as short intervals.

01:17:00.500 --> 01:17:03.930
So here's an interval
for the 880 hertz

01:17:03.930 --> 01:17:06.420
that's about 1 and
1/2 milliseconds.

01:17:08.520 --> 01:17:10.860
But here we have
a firing pattern

01:17:10.860 --> 01:17:15.290
that's exactly-- within the
limits of experimental error--

01:17:15.290 --> 01:17:18.320
exactly the same as
for the 440 hertz.

01:17:19.660 --> 01:17:22.520
Here we have an interval
that's representative

01:17:22.520 --> 01:17:26.870
of skipping a stimulus waveform
or two stimulus waveforms

01:17:26.870 --> 01:17:27.460
for 880.

01:17:27.460 --> 01:17:32.840
But here we have a peek at
exactly the same as 440 hertz,

01:17:32.840 --> 01:17:35.240
because these
intervals are lining up

01:17:35.240 --> 01:17:39.685
every other one for the
presentation of the octave.

01:17:40.760 --> 01:17:44.030
When you put those
two sounds together,

01:17:44.030 --> 01:17:46.670
you're going to get the
combination pattern.

01:17:46.670 --> 01:17:48.130
And many of the
intervals are going

01:17:48.130 --> 01:17:51.260
to be precisely on one another.

01:17:51.260 --> 01:17:55.290
And that is a very pleasing
sensation for you to listen to.

01:17:56.570 --> 01:17:59.880
If you look at other very
common musical intervals,

01:17:59.880 --> 01:18:03.890
like the fifth or
the fourth, you

01:18:03.890 --> 01:18:07.390
will have many
overlapping periods

01:18:07.390 --> 01:18:10.370
of interspike intervals
in auditory nerve firing.

01:18:10.370 --> 01:18:13.650
And those are very
common musical intervals.

01:18:13.650 --> 01:18:19.830
If you look at dissonant
interval, like 440 and 870,

01:18:19.830 --> 01:18:25.050
there will be no overlap
amongst those two frequencies

01:18:25.050 --> 01:18:26.722
in the auditory nerve firing.

01:18:30.020 --> 01:18:33.250
Now, let's go back
to psychophysics

01:18:33.250 --> 01:18:37.270
and give you one more
interesting piece of the puzzle

01:18:37.270 --> 01:18:42.120
here for why temporal codes
might be important is active

01:18:42.120 --> 01:18:45.740
matches become more difficult--
and actually impossible--

01:18:45.740 --> 01:18:47.015
above 5 kilohertz.

01:18:48.840 --> 01:18:50.230
OK, well, how does that fit in?

01:18:50.230 --> 01:18:54.370
Well, we just said
that phase-locking--

01:18:54.370 --> 01:18:57.110
because the hair cell and
auditory nerve can't keep up

01:18:57.110 --> 01:18:59.260
with one another,
the phase-locking

01:18:59.260 --> 01:19:01.060
diminishes for these
high frequencies.

01:19:01.060 --> 01:19:04.030
And there it becomes
impossible to match

01:19:04.030 --> 01:19:05.660
because you don't
have this timing

01:19:05.660 --> 01:19:07.166
code in the auditory nerve.

01:19:08.760 --> 01:19:13.310
So most of musical sounds
are confined to the spectrum

01:19:13.310 --> 01:19:14.635
below 3 kilohertz.

01:19:16.000 --> 01:19:19.690
If you look even at the upper
limit of the piano keyboard,

01:19:19.690 --> 01:19:21.710
you're sort of right
at 3 kilohertz or so.

01:19:23.380 --> 01:19:24.990
And that's probably the reason.

01:19:24.990 --> 01:19:26.120
It's a very likely reason.

01:19:28.390 --> 01:19:31.510
Now, we have a research
paper for today.

01:19:31.510 --> 01:19:33.850
And I'll give you the
bottom line or the take home

01:19:33.850 --> 01:19:36.930
message for that.

01:19:36.930 --> 01:19:40.120
And this is an interesting
neuralphysiological study

01:19:40.120 --> 01:19:45.260
based on a psychophysical
phenomenon called

01:19:45.260 --> 01:19:48.080
the octave enlargement effect.

01:19:49.750 --> 01:19:53.860
So I wasn't quite
truthful when I told you

01:19:53.860 --> 01:19:58.110
that octaves are the
most perfect interval

01:19:58.110 --> 01:20:02.130
to listen to, because it turns
out if you have people-- if you

01:20:02.130 --> 01:20:05.830
give people a low tone and
give them an oscillator.

01:20:05.830 --> 01:20:08.540
And say, dial in an
octave above that,

01:20:08.540 --> 01:20:11.390
they'll dial it in and say,
ah, that sounds so great.

01:20:11.390 --> 01:20:13.300
But if you look
really carefully,

01:20:13.300 --> 01:20:15.960
it's not exactly an octave.

01:20:15.960 --> 01:20:18.220
It's a small deviation.

01:20:18.220 --> 01:20:22.510
What they dial in-- especially
at high frequencies-- remember,

01:20:22.510 --> 01:20:24.590
you can't do this at
really high frequencies,

01:20:24.590 --> 01:20:28.450
but at 2,500 or 1,500
hertz, toward the high end

01:20:28.450 --> 01:20:32.060
of where you can do it-- they
dial in actually a little bit

01:20:32.060 --> 01:20:33.145
more than an octave.

01:20:33.145 --> 01:20:35.480
And they say, ah,
that sounds great.

01:20:35.480 --> 01:20:37.510
But it's not exactly an octave.

01:20:37.510 --> 01:20:42.000
So this paper looked and said,
what about auditory nerve fiber

01:20:42.000 --> 01:20:46.440
firing can explain this
octave enlargement effect?

01:20:46.440 --> 01:20:49.580
The fact that people
dial in a little bit more

01:20:49.580 --> 01:20:52.720
than an octave for
the upper tone.

01:20:52.720 --> 01:20:54.595
So these are the
psychophysical measurements.

01:20:55.600 --> 01:20:57.540
And they just give
you previous studies.

01:20:57.540 --> 01:21:00.305
What they did was they recorded
from the auditory nerve.

01:21:01.390 --> 01:21:03.920
And they looked at interspike
interval histograms,

01:21:03.920 --> 01:21:05.660
like we've just
been talking about.

01:21:06.970 --> 01:21:10.250
And they saw something
really interesting

01:21:10.250 --> 01:21:11.765
at the very high frequencies.

01:21:12.830 --> 01:21:15.380
So they're going to especially
concentrate in here.

01:21:17.400 --> 01:21:20.970
They found that the
very first interval

01:21:20.970 --> 01:21:24.710
didn't match exactly
what was predicted.

01:21:24.710 --> 01:21:27.730
So this is a stimulus
of 1,750 hertz,

01:21:27.730 --> 01:21:30.400
so toward the right
end of the graph here.

01:21:30.400 --> 01:21:34.570
And where you'd predict
the intervals to happen

01:21:34.570 --> 01:21:38.360
is shown by the
vertical dashed lines.

01:21:38.360 --> 01:21:41.130
And they didn't fire right
on those predictions,

01:21:41.130 --> 01:21:44.055
except for the very
shortest intervals.

01:21:45.530 --> 01:21:47.220
And they said,
what's going on here?

01:21:47.220 --> 01:21:52.160
Well, when you get to
very high frequencies,

01:21:52.160 --> 01:21:54.710
what are we talking about
for these intervals?

01:21:54.710 --> 01:21:57.880
Even at 1,000 hertz,
what's the time scale here?

01:22:01.390 --> 01:22:02.490
This is 1 millisecond.

01:22:05.430 --> 01:22:10.510
What problems do you get when
you ask a nerve fiber to fire

01:22:10.510 --> 01:22:13.780
and then you ask it to fire
again a millisecond later?

01:22:13.780 --> 01:22:15.030
That's a very brief interval.

01:22:15.030 --> 01:22:15.765
Anybody know?

01:22:21.820 --> 01:22:24.010
What is the limit of firing?

01:22:24.010 --> 01:22:28.514
Can nerve fibers fire closely
spaced action potentials,

01:22:28.514 --> 01:22:29.930
you know, less
than a millisecond?

01:22:29.930 --> 01:22:31.346
What's the problem
that they have?

01:22:31.346 --> 01:22:33.110
AUDIENCE: Is it that
they are polarized?

01:22:33.110 --> 01:22:35.740
PROFESSOR: They're
hyperpolarized, right.

01:22:35.740 --> 01:22:36.355
What else?

01:22:36.355 --> 01:22:38.770
AUDIENCE: Because of
the refractory period.

01:22:38.770 --> 01:22:39.770
PROFESSOR: That's right.

01:22:39.770 --> 01:22:47.580
There's something called
the refractory period, which

01:22:47.580 --> 01:22:49.230
can cause them to hyperpolarize.

01:22:49.230 --> 01:22:54.320
So what's happening-- if this
were a nerve cell membrane,

01:22:54.320 --> 01:22:58.770
is what happens is sodium
channels can open up

01:22:58.770 --> 01:23:01.420
to allow sodium to come in
and depolarize the neuron

01:23:01.420 --> 01:23:02.765
and fire an action potential.

01:23:04.550 --> 01:23:06.490
And then those channels
are turned off.

01:23:06.490 --> 01:23:09.600
And potassium channels
open up and allow potassium

01:23:09.600 --> 01:23:11.233
to go back and
even hyperpolarize.

01:23:12.370 --> 01:23:16.640
But these channels take a
little bit of time to recover.

01:23:16.640 --> 01:23:19.180
It takes a little bit
of time for the sodium

01:23:19.180 --> 01:23:23.250
to turn off and get ready to
fire another action potential.

01:23:23.250 --> 01:23:26.020
It takes a lot longer time
for the potassium channel

01:23:26.020 --> 01:23:29.650
to close and get ready to
fire another action potential.

01:23:29.650 --> 01:23:32.580
And the refractory
period is the time

01:23:32.580 --> 01:23:35.370
it takes for everything
to recover fully

01:23:35.370 --> 01:23:36.845
so we're ready to fire again.

01:23:38.570 --> 01:23:41.351
And in the limit, that's
supposed to be about 1

01:23:41.351 --> 01:23:41.850
millisecond.

01:23:44.700 --> 01:23:48.320
That's the absolute
real refractory period.

01:23:48.320 --> 01:23:51.040
So when I was drawing here
things a millisecond and less,

01:23:51.040 --> 01:23:53.280
I wasn't really being truthful.

01:23:53.280 --> 01:23:56.180
There's something called the
relative refractory period,

01:23:56.180 --> 01:23:57.737
which is a couple
of milliseconds.

01:24:01.640 --> 01:24:04.880
And the nerve fiber can
respond, but it's not

01:24:04.880 --> 01:24:07.955
going to respond quite
as quickly as before.

01:24:08.970 --> 01:24:11.510
All these channels
aren't completely reset.

01:24:11.510 --> 01:24:16.520
It's going to take a little
bit longer time to respond.

01:24:16.520 --> 01:24:20.100
That's what's going on
in this very first peak.

01:24:20.100 --> 01:24:22.990
Remember, this first
peak indicates firing

01:24:22.990 --> 01:24:30.130
at successive cycles of the
sound waveform at 1,750 hertz.

01:24:30.130 --> 01:24:32.780
It's a very brief interval.

01:24:32.780 --> 01:24:34.890
And what happened is you fired.

01:24:34.890 --> 01:24:36.470
And then the next
action potential

01:24:36.470 --> 01:24:39.120
you fired, but you
delayed a little bit.

01:24:39.120 --> 01:24:41.140
So the interval is
a little bit longer.

01:24:43.210 --> 01:24:46.270
That pushed this one
up to the next peak,

01:24:46.270 --> 01:24:50.030
so that this interval
was actually too short.

01:24:50.030 --> 01:24:52.380
What the brain is
getting is an interval

01:24:52.380 --> 01:24:54.560
that's a little bit too short.

01:24:54.560 --> 01:24:57.110
When it hears that,
it says, I want

01:24:57.110 --> 01:24:59.900
to recreate the higher frequency
to be an interval that's

01:24:59.900 --> 01:25:02.920
a little too short, I'm going
to dial in a little bit too high

01:25:02.920 --> 01:25:03.490
in frequency.

01:25:05.310 --> 01:25:09.190
OK, because that
frequency sounds better

01:25:09.190 --> 01:25:10.890
with this too short
of an interval.

01:25:12.590 --> 01:25:14.090
So go ahead and read that paper.

01:25:14.090 --> 01:25:19.170
It's a very interesting study
of how neuronal firing can

01:25:19.170 --> 01:25:21.390
give you a
psychophysical phenomena.

01:25:22.610 --> 01:25:23.790
That's quite interesting.

01:25:23.790 --> 01:25:27.230
And it happens especially
at the high frequencies.

01:25:27.230 --> 01:25:28.970
Octave matches the
low frequencies

01:25:28.970 --> 01:25:32.746
are sort of as you would predict
as is auditory nerve firing.

01:25:37.610 --> 01:25:39.310
OK, any questions?

01:25:39.310 --> 01:25:41.430
If not, we'll meet
back on Wednesday.

01:25:41.430 --> 01:25:43.560
And we'll talk about the
cochlear nucleus, which

01:25:43.560 --> 01:25:47.310
is the beginning of the
central auditory pathway.