These are very unique times for brain research. The aperitif for the course will thus highlight the present “brain-excitements” worldwide. You will then become intimately acquainted with the operational principles of neuronal “life-ware” (synapses, neurons and the networks that they form) and consequently, on how neurons behave as computational microchips and how they plastically and constantly change - a process that underlies learning and memory. Recent heroic attempts to realistically simulate large cortical networks in the computer will be highlighted (e.g., “the Blue Brain Project”) and processes related to perception, cognition and emotions in the brain will be discussed. For dessert we will deliberate on the future of brain research, including the questions of “brain and art”, consciousness and free will. For more information see the course promo below and read “About the course.”
Early Processing of Sensory Information
Loading...
En provenance du cours de Hebrew University of Jerusalem
Synapses, Neurons and Brains
635 notes
À partir de la leçon
Perception, Action, Cognition and Emotions
During this module we will have a special lecture which will be given by Prof. Israel Nelken from the Hebrew University in Jerusalem, who will discuss "Perception, action, cognition and emotions". Until today we have focused mainly on the function of single cells and small networks. Module #8 dwells into higher level computations and especially "the story of sound". The auditory system converts sound into electrical signals. Hair cells, basilar membrane, the cochlea and others are all tools which the brain uses in order to translate the outer world into neuronal activity. The next step is perception – processing the sensory information into useful representation. One beautiful example is the case of using binaural cues for the localization of sound. Perception leads the organism to actions – the brain predicts the world and computes the action that will yield a maximal reward and avoid punishment. What happens in the case of surprises? Here, higher level computation is needed. We will end the lecture with emotions – What are they, how are they represented in the brain and how they affect our actions.
Rencontrer les enseignants
Idan SegevProfessor
Computational Neuroscience
After having talked about auditory transduction, the way that sounds become
electrical activity, I'm going to show you now, one case of an interesting
processing of sounds and specific of the electrical activity that results from
sounds. And specifically, I want to discuss the
issue of auditory localization. We can identify the direction of sound
source of, of a, of objects by the sound that they emit.
So we hear a sound, and we'll turn our heads towards the, towards the sound.
This is a very interesting use of sound, and the reason is that there is no
explicit representation of sound direction in the ear.
The ear has well, a sound. So identification of sound direction
requires comparing the sound to each of the two ears.
And therefore the auditory localization requires computation.
It requires information from the two ears to which the central nervous system and
get processed there. So how do we know which the direction of
sound, what are the physical queues that we can, we can use.
in order to identify the direction of a sound source.
these cues have to do with the sounds reaching the two ears and so they are
called Binaural cues, Binaural means things that happen in the two ears
separately. W are going to discuss only the
localization of pure sine waves, pure frequencies and when it feel sine waves
reaches the ears, it has two types of physical differwences.
It produces two types of physical differences in the two ears.
First of all it reaches first the near ear, and only then the far ear.
So there is a slight shift in time, a minute shift in time of arrival of sounds
to the two ears. And secondly the head acts roughly as a
shadow. So the sound is going to be louder in the
near ear and slightly softer in the other ear.
That's if I put microphones in the two ears and I present a sine wave, and I
record a sound in the two ears I will find two sign waves.
The sine wave in the near ear is going to be earlier and, earlier and louder.
We call the time differences between the two waves, the interaural time
difference, it can again, interaural between the two ears.
So interaural time difference or ITDs, and the level differences, are the
interaural level differences, ILDs. I'm going to discuss mostly the
calculation of the ITDs, of the interaural time differences.
This time different cells between the eh, peak of the sine waves reaching eh, each
of the ear are really small. We are talking, when we are talking about
eh, human heads, the distance between the two ears is something like 25
centimeters. And the time that takes the sounds to
move from one ear to the next to the other is less than one millisecond.
If we are talking about the ability to descriminate between the location of two
sound sources in front of you. We can identify differences on the order
of two degrees. And the, the time differences between a
sound that is straight, the, it's, a sound that is straight ahead relay to the
ear, the two ears at the same time. And, and the, a sound that is shifted by
a few degrees from the center Will reach the close yield few times of micro
seconds a, earlier than a a the other yield.
So we're capable of disseminating interaural time difference of few tens of
micro seconds that's hundreds of a millisecond.
The important property that allows the audotory system to know when these peaks
of the sound wave occur is called Phase Locking.
It turns out that when the auditory nerve fibers, that are sensitive to the element
range of frequencies, low frequencies, when the auditory nerve fibers fire, they
fire at a fixed relationship to the shape of the incoming sound.
So if I excite the, if, if I play a sine wave, to the ear, and I, eh, record the
spikes that occur during, the presentation of this sound, these spikes
are going to happen, in this case, mostly it's the peak of the sign waves.
So the time which spikes occur, mark the peaks, the pressure maximum in the
incoming sound. And this is true for both ears.
So now when the sound reaches the two ears at slightly different times.
These spikes are going to occur at slightly different times in one ear
relative to the other. This is a complex slide, I'm going to
explain to go over it very slowly and it describes the pathway.
That eh, eh, is eh, involved in the calculation of the inter aural time
differences. So first of all, we have the ears, right
ear, and left ear, and the sound reaching the two ears.
Eh, the ear canal, the tympanic membrane, and here's the cochlea.
These eh, large ci, circular structures are part off the vestibular system that's
involved in eh, eh, eh, balance. So the auditory nerve come out of the
cochlea, of the cochlea, goes into the, the central nervous system, and ends in a
structure called the cochlear nucleus. So we have the ear, the auditory nerve
and the cochlear nucleus on this one side.
And the ea, the ear, the cochlear nucleus from the other side.
The auditory nerve fiber terminate in the cochlear nucleus.
They make eh, very large synapses on the relevant neurons in the cochlear nucleus,
and these neurons now send the axons to a third structure that sits not far from
the cochlear nuclei called the medial superior olive.
I wanted to show you a little bit a rough view of the brain in order to for you to
have an idea of where these structures lie.
So this is a presentation of part of the brain called the brain stem.
The brain stem comes, comes out of the skull so from here down we have the
spinal chord, the spinal chord goes into the skull and becomes brain stem.
On the brain stem we have the two cochlear nuclei.
So the ear, one ear sits here the other ear sits here.
And, the cochlear, the auditory nerves come of the ear, crosses the bone and
goes into the cochlear nucleus. The right ear to this cochlear nucleus
and the left ear to that cochlear nucleus.
Slightly further forward in the brainstem, we have the nuclei that
together form the so called superior olivary complex, and the MSO, the medial
superior olive is part of the superior olivary complex.
The neurons of the MSO have dendrites pointing, to, to the two sides.
One side gets the inputs from the axons of cochlear nucleus of the same side.
And the pther side, the dendrites of the other side of the neuron get input that
come on axons, that come from the other side from the other ear.
How does this work? The point to notice in this diagram, the
crucial point to notice in this diagram is that the axon from this ear is a bit
shorter than the axon from the other ear. The reason for that is simply distance.
The, this superior olive sits closer to its cochlear nucleus, than it sits to the
cochlear nucleus from the other side. So we have slightly longer axons.
Next we know that action potential propagation along an axon takes time.
There's a finite velocity, so the longer the axon the longer it will take for the
spike to go from the cell body to the axon terminus.
In consequence, a spike that occurs in, in this, on this side of the brain, will
reach this, the neuron in the MSO slightly earlier than its, its spike that
occurs slightly faster than it spikes that occurs on this side.
Okay, now assume that a sound a occurs a on the right side.
This sound will reach the right ear fills.
And the left ear a little bit later, 300 microseconds later, let's say.
The spikes that are locked to the peaks of the sine wave will occur 300
microseconds earlier here, than here. However, this axon is longer.
And if its length is such that the travel time from this cochlear nucleus to MSO is
exactly 300 microseconds longer, than the travel time from this cochlear nucleus to
the MSO. The spikes that originated from this, it
will reach the MSO at exactly the same time as the spikes that originated from
that ear. In consequence, the spikes reaching the
MSO neurons from both sides will arrive at the same time and this MSO neurons
have a special property, their coincidence detectors.
The fire mostly when spikes occur simultaneously on both sides.
So, the MSO will see the spikes at the same time, and the reason that it will
see the spikes at the same time is because of this internal delay caused by
the axon length and it will fire. On the other hand, for this specific
neuron if the sums come exactly from the middle, the spikes will occur at the same
time, locked to the peaks that occur at the same time in both ears.
The delay due to the longer pass lengths here, the axon lengths here, will cause
the spike to reach the MSO neuron at slightly different times on the two
dendrites. And then the MSO neuron will not see the
spikes at the same time and so it will not fire.
In summary we got here a circuit that a insensitive to the time of arrival, to
the, to the interval or time difference. It fires in this case, this neuron fires
when they. sounds arrives at the right ear first
then its the left ear later and it doesn't first when the sound reaches the
two ears at the same time. So this neuron can tell us that the sound
came from the right rather than from the center or, or obviously from the left.
This circuit requires extreme specialization of the bio physics of the
neuron that participate in it. The reason is that it needs to keep the
timing of the spikes extremely, precisely all the all the way from the auditory
nerve until they reach the MSO. And this is done by neuron in the
cochlear nucleus with the specialized not to do anything to the input for relying
it to the output. You learned with a done a lot about what
neurons can do. That did calculations, temporal
summation, partial summation and so on and so forth.
The neuron in the cochlear nucleus participate in this eh, in this silqeek
are designed so that they don't do anything of eh, eh, eh, any of these
things. The only thing that they do is to produce
a spike with very short, very small jitter for every spike that reaches them
from the auditory nerve. Secondly, the, the neurons that do the
coincidence detection in the medial superior olive also very special in terms
of, most of the structure and of the bio physical mechanisms.
And the reason that again, is that they needed a, this coincidence detection
process that they participate in, requires sensitivity to time differences
of a tens of micro seconds. You learned with [INAUDIBLE], that spike
wave is order of one millisecond, that's 1000th micro second.
So, we have spikes whose width is 100's of micro seconds, maybe 1000th micro
second. And with these extremely wide things, we
need to be able to discriminate into all time differences of a few tens of
microseconds. So we cannot allow our self any jitter,
we need a new one who's temple, eh, eh, temporal summation properties work with
time concerns that are extremely short. There are other problems, other issues
that these neurons have to solve as well. And in fact it turns out that some of
these a problems that a are due to the to, to, to the attempt to the
discriminate tens of micro seconds we say [INAUDIBLE] life will whose time
constants are much longer, some of these problems.
probably solved by using tricks that that have to do with the properties
[INAUDIBLE]. Ehh, the, this is as far as I want to go
into the description of the, eh, ITD interaural time difference all-time
difference calculation. But I want to remark that similar
mechanisms can be observed throughout the sensory systems.
Sensory systems have, each has its own specialized needs and specialized
calculations that needs to be done. I showed you here, the comp, the
comparison of timing between the two ears.
There's a parallel pathway that calculates the inter aural level
differences, the differences in the amplitude of the two sound that reaches,
reach the two ears. And in other sensory systems you find
other such eh, computational tasks that are done by neurons.
For example, the retina contains neurons that calculates motion, that they're
sensitive to motion and that they act to motion in one direction but not in other
directions. Eh, so this is a kind of calculation
that's done by the, by the retina. Eh, early processing of smells also has
its own share of eh, specialized eh, eh, calculations and needs to be done, and
all of these specialized needs are catered in different sensory systems by
different specialized circuits. So, for example, in the auditory system,
we have the specialized circuit in the auditory brain stem.
They're specialized in terms of the neurons that live there.
They're specialized in terms of the connectivity between them, and they'll
specialize in terms of the computations that they perform.
Similarly, in the retina, there are a lot of different [UNKNOWN], which are now
starting get eh, eh, eh, dissected out. And that perform all of these specialized
calculations that are specific to the visual system.
And similarly in the eh, eh, for the smell eh, sense, there's eh, a structure
called the olfactory bulb, which gets the input from the nose.
And eh, again, has a specialized circuitry that eh, performs the early
processing that's eh, that is specific the olfactory system.
And all of these, again, come with a lot of important fascinating details, and I
urge you to go and look for them. So this is what I wanted to discuss with
the spectral early processing of sensory information.
Explorer notre catalogue
Rejoignez-nous gratuitement et obtenez des recommendations, des mises à jour et des offres personnalisées.
Coursera propose un accès universel à la meilleure formation au monde,
en partenariat avec des universités et des organisations du plus haut niveau, pour proposer des cours en ligne.
© 2018 Coursera Inc. Tous droits réservés.
