Open in TheBrain

Consciousness Framework Crick Koch

<p class="default">06:51 PM, August 10, 2007

&#160;

Christof Koch and Francis Crick A framework for consciousness

Nature Neuroscience vol6 #2 feb 2003

&#160;

Summary of their work on NCC (neural correlates of consciousness)

page 119 (first page)

&#160;

General Strategy: NCC - neural correlates of consciousness.

&#160;

Hard problem is too hard. Instead we look at " &#160;<b>what are the NCC of consciousness</b> -

the minimal set of neuronal events that give rise to a conscious percept?"

&#160;

how is consciousness of color, shape, or movement produced?

&#160;

VISUAL system of primates is their focus. (cannot do emotion or self-consciousness)

&#160;

ALERT Macaque monkey. must investigate single or small groups of neurons on very fast time scale.

&#160;

Humans can report contents of consc. can can do fMRI; monkeys ... good for neuronal experiments.

Do both in parallel.

&#160;

Framework

&#160;

Biology does not have rigid laws, as physics does. Evolution produces mechanisms and submechanisms:

but these have occasional exceptions. Example of DNA=&#62;RNA as central mechanism but with occasional exceptions.

&#160;

Cerebral Cortex (preamble)

p119: By cortical system we mean cortex plus all the subcortical systems connected to it.

&#160;

We refer to the "FRONT" of the cortex (frontal lobe/prefrontal etc) and the "BACK" of the cortex

Dividing line is the CENTRAL SULCUS.

&#160;

Visual system is likely to be specified EPIGENETICALLY in considerable detail and can learn a large amt

from experience.

&#160;

The main function of the sensory cortex is to construct and use highly specific feature detectors such as

those for orientation, motion, or faces.

page 120

&#160;

The set of features that any neuron responds to is its "receptive field" apt to be highly specific and multi-dimensional.

&#160;

An important aspect of the firing of a neuron is its PROJECTIVE FIELD (what it projects to) its perceptual and behaviioral consequences.

&#160;

Mechanism of learning is probably Hebbian.

&#160;

Is cortical action LOCAL??? NO. Extensive cortico-cortical routes and cortico-thalamo-cortical routes.

&#160;

For predators and prey both: FAST is best. FAST (but rough and ready, imperfect) response is good

followed by a more prolonged response that produces a more perfect result.

&#160;

Also, use several rough and ready methods in parallel to reach a conclusion .. mya be how people see DEPTH.

&#160;

Brain FILLS-IN is there is incomplete info.

&#160;

Filling in happends in many places in the brain. Important guide to human behavior: JUMPING to CONCLUSIONS.

&#160;

Our present framework:

&#160;

Not interested in laying down new memories (time scale is too long)

We are interested in time periods of a few hundred milliseconds.

&#160;

120: People are NOT directly conscious of neural activity in primary visual cortex (V1).

&#160;

In addition to &#160;<b>a slower, all-purpose conscious mode</b> , the brain has many &#160;<b>"zombie modes</b> : which are

characterized &#160;<u>by rapid and somewhat stereotypical responses</u> .

&#160;

  1. The (unconscious?) &#160;<b>homunculus</b>

<b>the FRONT of the brain is "looking at" the BACK of the brain</b> .

&#160;

Is the neural activity in the front of the brain largely unconscious?

One proposal, for example, is that humans are directly conscious of their thoughts, but

only of sensory representations of them in their imagination. At the moment,

there is no consensus about this.

&#160;

The hypothesis of the homunculus is very much out of fashion, BUT this is how everyone thinks of

themselves. It would be surprising if this overwhelming illusion did not reflect in some way the genl

organizaiton of the brain.

&#160;

<b>2. Zombie modes and consciousness</b>

&#160;

Many actions in response to sensory inputs are rapid, transient, stereotyped and unconscious.

They could be thought of as cortical reflexes.

&#160;

Consciousness deals more slowly with broader, less stereotyped aspects of the sensory inputs (or a reflection

of these in imagery) and takes time to decide on appropriate thoughts and responses.

&#160;

It is a great evolutionary advantage to have fast zombie modes that respond stereotypically AND

a slower system that allows thinking and planning more complex behavior.

&#160;

In zombie mode, the main flow of info is feed-forward.

In conscious mode, it seems likely that flow is in both directions: resembling a standing net-wave.

&#160;

Page 121:

&#160;<b>3. Coaltions of Neurons</b>

Coalitions are assemblies of neurons.

The various neurons in a coalition support one another by increasing activity of fellow members.

The dynamics are not simple.

In general, at any moment the winning coaltion is somewhat sustained, and embdoies what we are conscious of.

&#160;

Consider a political analogy. The primaries and early elections correspond to the preliminary

unconscious processing. The winning coalition associated with an object or event

would correspond to the winning party, which would remain in power for some time and

would attempt to influence and control future events.

"Attention" would correspond to journalists and pollsters focusing on certain issues rather than others to bias the electorate

in their favor.

For example those large pyramidal cells in cortical layer 5 that project to super. colliculus and thalamus

(both involved in attention) would correspond to electoral polls.

&#160;

<b>4. Explicit representations</b>

&#160;

Explicit rep means that a small set of neurons exists that responds

as a detector for a particular feature, without further complex processing.

Can a single layer of neurons deliver the correct anser.

&#160;

Single layer neural network fed from retina cannot recognize a face.

BUT a single layer fed from the relevant parts of the interior temporal cortex,

might reliably signal face or no face.

&#160;

There is much evidence from humans and mokeys that if there are no such

explicit neurons or they get wiped out, then the subject CANNOT consciously

perceive that aspect directly.

&#160;

Well known examples are &#160;<b>achromatopsia</b> (loss of color perception);

<b>prosopagnosia</b> (loss of face recognition); or &#160;<b>akinetopsia</b> (loss of motion perception).

&#160;

A node in the cortical network is needed to express one aspect of one percept or another.

An aspect cannot become conscious unless there is an essential node for it.

For consciousness, there may be other necessary conditions, such as

PROJECTING TO THE FRONT of the BRAIN.

&#160;

A node, all by itself, cannot produce consciousness. Even if the neurons

in the node were firing appropriately, this would produce little effect, if their

OUTPUT synapses were inactivated. A node is a node, not a network.

&#160;

<b>5. Higher levels first</b> .

&#160;

For a new visual input, the neural activity first travels rapidly UP the visual hierarchy to a

high level, possibly in the fron of the brain. Signlas then move backward down.

So... gist of the scene first then successive lower levels provide more detail.

&#160;

<b>6. Driving and modulating connections</b> .

&#160;

It is possible that connections from the back of the brain to the front are larely driving, whereas reverse pathways are largely

modulatory.

&#160;

  1. Snapshots

&#160;

We propose that conscious awareness (for vision) is a series of static snapshots,

with motion "painted" on them. picture of girl running.

&#160;

What could be special about this activity that reaches above the consciousness threshold.

it could be some particular way of firing: sustained, high rate or synchronized firing or

in bursts. Or it might be firing of special types of neurons such as those pyramidal cells that project to

the front of the brain. This may seem unlikely but, if true, would greatly simplify the problem.

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;

&#160;</p>

Attachments