A. Summary of the Module: Dealing with time in the nervous system:
Theme Brain Region: Cerebral Hemispheres.
Sensory surfaces to maps; coincidence detection as a general mode; examples from neuroethology; clocks in the nervous system; nanoseconds to microseconds to milliseconds to hours
Integrative Power of the Nervous System:
How many neurons?
"There are 10 to the power10 neurones in the human brain, of which
10 to the11th power lie in the cerebellum"
Walle Nauta, ~1970
Nauta's joke hinges on the fact that von Economo, who personally counted the number of neurons in the human brain in the 19th century, came up with a figure of 10 to the10th, but omitted the tiny granule cells of the cerebellum which do not look like typical neurons with the stains used by von Economo. There are about 10 to the 11th cerebellar granule cells . There are also granule cells in the hippocampus, so the true total number of neurons in the human brain probably lies somewhere between 10 to the11 and 10 to the12.
Nauta's joke reminds us to keep an open mind when it comes to dogma about the brain.
How many synapses?
Human cerebellar Purkinje cells have close to 1 million spine synapses
each. Neocortical pyramidal cells in frontal and pre-motor cortex can have
10 to the 5th spine synapses each . Small neurons like granule cells may
have less than one hundred synapses.
How Many Connections?
As a rough exercise we can conservatively estimate the mean number
of connections per neuron at 1000, noting the big margin of error here
(5 orders of magnitude!) . This gives a total connectivity per brain of
around 10 to the 15th power (10 to the power 3 X 10 to the power
11.5).
How many bits/synapse?
It is usually been assumed that one spine synapse is a processing unit,
with no integrative power of its own. This could be wrong.
A spine has the same dimensions as a bacterium, which performs complex
calculations, has memory, shows sophisticated chemotaxis, adaptation, signal
processing etc.
A spine is much more complex morphologically than a bacterium, so it
will not be a surprise to learn that it can process its own inputs (using
learned temporal filters, for example, as suggested by Horace Barlow).
If we assume that each spine can recognise 10 different inputs, the
number of possible interconnections in the human brain is ~10 to the 16.
How many different brain states?
If one permutes 10 to the16th different inputs, remembering that they
are each excitable and can have different states, the numbers are staggering.
For example, taking just 10 inputs at a time, the number of possibilities
is 10 to the powers of 16+15+14+13+12+11+10+9+8+7, 10 to the 115 power!
If you think that it is inappropriate to take individual synapses one-by-one
in this way, and that it would be better to look at the mass action of
millions of them, the numbers are still huge. The activity of small numbers
of neurons can also make a difference, so the calculation is not wildly
off. To take just one example of the great sensitivity and specificity
of the brain, individual retinal neurons can detect single photons! A human
observer can see and count extra quanta. Really.
There are also some motor phenomena that seem to require the concerted
action of small groups of neurons. Newsome and colleagues
can preditably change the perception and behaviour of a behaving monkey
by tickling a few neurons in the middle temporal visual area with
electric current from a microelectrode.
The Most Complex Object in the Known Universe:
One can easily imagine many more neurons being drawn into play than
the 10 in the example above, so we our brain easily enters into the magic
land of Dirac's big numbers, where 10 to the 80th power is the total number
of protons in the universe (something like that, I think). Whether
I have these numbers exactly right or not, the point is that the human
brain is by far the most complex object in the known universe. Even our
galaxy pales by comparison with the brain, if we take its ~10 to the 9th,
largely non-communicating, components on a star-by-star basis (and ignore
the possibility of other life and other brains out there!).
click
here for a view of a galaxy
click here for
a different view of the complexity of the human brain
Introspection is OK:
Introspective observations about the brain and mind are considered
suspiciously by some, not being as apparently objective as hard physiological
observations made on nervous tissue with recording apparatus.
By contrast, this course will make frequent use of introspection. Hard-nosed
sceptics may therefore consider the whole thing too soft, too subjective
and "not physiological enough".
Two points for such sceptics:-
1. All science ultimately depends on subjective data we gather with
our senses and analyse with our brains. No one can escape from this philosophic
problem of subjectivity! To take one example, you might be at home
in bed dreaming about reading this point, rather than actually reading
it!
An eloquent account of this problem is given in What Is Life? by
Erwin Schrodinger, whose quantum wave equation is perhaps the most used
equation in science and who influenced a large number of brilliant physicists
to change fields and create the new field of molecular biology.
2. In the twenty-first century, introspection is becoming fashionable
because of the advent of sophisticated brain scanners, such as functional
magnetic resonance imaging (fMRI) that reveal the patterns of activity
underlying brain states. This enables a new check on the validity of a
brain state......... for the suspicious, ............as well as a picture
of its underlying neural mechanisms, ..................for the curious.
I predict that in the next decade we will be able to validate and understand
the neural basis of the many meditative brain states discovered by masters
over the centuries, but mostly lying outside of present hard-nosed medical
disciplines.
B. Contrast Static (Anatomical) with Dynamic (Plastic) Properties
of Brain
The emerging view of brain organisation is much more dynamic than the
traditional, rather anatomical view. I will give a number of striking examples
of reorganisation. These involve synaptic plasticity and changes in representations
over short time frames.
All text books give the static picture of the nervous system, which
mostly amounts to neuroanatomy, some details of which you will nevertheless
have to learn to ground the dynamic view.
The dynamic view of brain function is elusive and hard to capture,
although new techniques such as optical recording of activity of large
numbers of nerve cells and functional magnetic resonance imaging (fMRI)
are bringing it closer. A readable account of non-linear dynamics of the
cerebral cortex can be found in Walter Freeman's "How the Brain
Makes Up Its Mind" (1999
C. Constructing sensory representations:
a. Retinal photoreceptors vs. hair cell mechanoreceptors.
b. Topographical maps
c. Computational Maps
d. Constructing a map
1. Genetic algorithms: Sperry's chemoaffinity molecules
now cloned as molecules of the ephrin class (F.
Bonhoeffer).
2. Self-assembling networks: activity-dependent connections
D. Coincidence Detection in Cerebral cortex:
Somatotopic map vs. Somatotopic plasticity:
All the textbooks show a homunculus in the somatosensory cortex that has very large lips and fingers on a tiny body, to show the fact that the size of the representation in cortex is related to the size of the innervation: highly innervated areas like the fingertips have more neurons in cortex devoted to their processing. This is all relatively easy to understand and fits well with the anatomical picture of the brain's wiring, like a telephone exchange making specific connections between regions. This is also the prevailing view.....so be prepared to change it with the new information showing that the map is actually much more dynamic, with considerable abilities to readjust quickly to changes.
1. Amputees and phantom limbs: Stimulated by experiments on flying foxes here at UQ, Ramachandran did a simple test on recent arm amputees, using a cotton bud. Stroking the amputee's face (remember that the face representation is adjacent to the arm representation in the somatotopic map of cortex), Rama found that subjects experienced a precisely localised sensation in their phantom that moved around the phantom arm as the location on the face was changed! I saw this experiment carried out in front of an audience at a pub, completely impromptu, on a one-armed sailor who had the same strange remapping between face and his phantom as the other amputees. Using modern brain scanning methods, it can be shown that the new map in cortex appears immediately after the amputation.
2. Animal Studies: (Calford, Kaas, Merzenich)
a. Denervating a thumb does not lead to a an unreponsive "hole" in
the map as might be predicted from the static, anatomical viewpoint. Instead,
the map instantly reorganises itself so that the neurons that would otherwise
be responsive to stimulation of the thumb are now responsive to adjacent
digits.
b. If a monkey is trained to use one digit continually, at the expense
of the others, the region of the map corrresponding to that digit becomes
larger.
click here for diagram of topographic remapping after amputation
There are a number of party tricks that depend upon this phenomenon of cortical plasticity and reorganisation. Try them yourself, bearing in mind that some people at some times are rigidly unable to experience these changes. It helps to be a positive frame of mind rather than a negative, sceptical one.
Self-illusions caused by "suspicious coincidences":
a. Hand example; tap irregularly, but in synchrony, both
on the subject's hand out of sight under the table and at the same time
on someone else's hand, (or a dummy hand, or even the table itself) that
is visible to teh subject. When this works (as it does on most happy party
goers) one gets the impression that one's hand under the table has moved......
to occupy the other hand, or the dummy, or the table itself!
b. Nose example; Arrange a ring so that each person can stroke
the nose of the person adjacent with two finger tips. Do this with closed
eyes, with everyone in exactly the same rhythm so that your hand movements
are synchronous with the stroking by your neighbour. When this works, your
nose moves off to occupy a new position under your fingers many centimetres
away from your face.
c. Head example; This requires some props: A life size dummy,
2 desk lamps and a half-silvered mirror (some smokey plexiglass works).
Arrange yourself and the dummy on opposite sides of the 1/2 mirror so that
the dummy's reflection is exactly superimposed on your own. Illuminate
the lower half of your own face with one lamp and illuminate the upper
half of the dummy's face with the other, so that the reflection is a composite
of both you and the dummy, but has a stronger contribution from you mouth
region. Now move your mouth and lips in an exaggerated fashion, through
a wide range of emotions and postures. If this works (it shocked me when
I did this), you brain will abandon its model of your body and instead
adopt one where you are decapitated, with your detached head staring at
you from across the table!
All these demonstrations emphasise the dynamic nature of the brain's model of the world and its surprising readiness to change that model when the evidence demands it.
d. Ketamine:
This drug is commonly used as an anaesthetic for both humans and animals,
in tranquilliser darts, and illicitly for "astral", out-of-body experiences.
The mode of action is NMDA receptor blockade. NMDA receptors are important
for synaptic plasticity (they provide a substrate for Hebbian synapses
as the receptor is active only when the whole neuron depolarises in an
action spike and ejects the Mg2+ ion that otherwise blocks the receptor.
This link between receptor activation at one synapse and depolarisation
of the whole neuron caused by the action of many other synapses, is an
essential feature of associative learning mediated by Hebb-type synapses).
The role of plasticity in continually re-updating our cortical model of
the world explains why an NMDA receptor blocker such as Ketamine, in modifying
plasticity, should bring about such a dissociative, out-of-body experience
("Yes! That needle is sharp, but it is not troubling me, as I am up here
in the corner of the ceiling watching you stick the pin into my body down
there").
One trial learning: Konrad Lorenz' Banded Iguana:
It would not be possible to do this experiment today, given that the
banded iguana has all but disappeared from Fiji,. Nevertheless, we can
still learn from Lorenz' account of the contrived meeting between a territorial
male and his female mate, painted to look like an encroaching male.
If a lizard has an enduring memory of a single emotional encounter,
humans can be sure that one-trial learning will be a feature of our own
nervous sytem.
click here for a pic of the now-endangered, Banded Iguana, from Fiji.
One-off memories are the forte of the cerebral cortex, whose declarative kind of memory contrasts with procedural memory connected to motor processes and structures more posterior in the brain such as the cerebellum. (see list of differences between these two kinds of memory).
Constructing self: Continually Updated:
Hebb synapses: foreshadow cerebellar anti-Hebb
Contrasting forms of Coincidence detection: LTP vs LTD
Hippocampus Declarative memory
Cerebellum Procedural memory
| DECLARATIVE MEMORY | PROCEDURAL MEMORY | |
| 1. Trials needed to "lay down" memory trace | Single trial memory possible (cf. iguana experiment) | Many repeat trials needed to secure memory |
| 2. Creativity and Sensitivity of memory system | Both Creative and Sensitive: The complete memory can be recalled using only a tiny fraction of the original inputs | "Stupid" and Insensitive: Needs most of the original inputs to recall the memory (cf. riding a bike) |
| 3. Principal mechanism of synaptic plasticity and coincidence detection | Long Term Potentiation.................LTP | Long Term Depression.....................LTD |
| 4. Second Messenger system | cAMP | cGMP |
| 5. Errors | Error-prone (because of 1 & 2). | Error-free (because of 1 & 2) |
| 6. Error correction | Difficult, especially for one trial events.
? during REM sleep |
Can occur regularly, during actions whose fidelity of execution can be judged. |
| 7. Associated brain structure | Bilateral;; Cerebral Hemispheres (incl. hippocampus) | Midline: Includes cerebellum |
Further Reading:
VS Ramachandran and Sandra Blakeslee
Phantoms in the Brain
