See Also: Book Notes, Conceptual Spaces Cheat Sheet, Swarm Intelligence, Blank Slate, Clock of the Long Now, Info Viz & Perception, Consciousness: a user's guide, Hawkins, On Intelligence <> ME: YON

Note: Christof Koch's 2012 Consciousness: Confessions of a Romantic Reductionist will be an easier place to start. The Quest requires a knife & fork to properly consume it, but it is more that worth the effort.

The Quest for Consciousness

Christof Koch, 2004 Roberts & Co. ISBN 0-9747077-0-8

Christof Koch's research goal is "to link the molecular and biophysical properties of coalitions of neurons to stimulus of awareness". (p. 100) That is to pinpoint the NCC, the neuronal correlates of consciousness. What is actually happening when we have a conscious percept. This very specific goal allows him to side step many sorts of philosophical problems of defining consciousness.

This is a fairly acedemic book with 70 pages of bibliography (mostly research papers), 18 pages in indices and 18 pages of glossary. Koch's course. I'm glad I read Zeman's Book first. Zeman is in the bibiography. Koch's book is more hard core.

Koch published several papers with Francis Crick (DNA) and dedicates the book to him. Crick should have been a co-author, but declined. Nobel Bio and WikiPedia

An election metaphor is often used to represent coalitions of neurons that sustain a conscious thought process. Our sensory processors - the visual cortex being the most well known - can be churning away full tilt, but, it will not effect our conscious thought - as long as nothing unexpected and urgent appears. It is some set of inputs that coalesces to take over the conscious uni-processor.

A fascinating supposition is that the frontal cortex is not part of our consciousness. The basic idea is that we have sensory processors which feed information (a net wave) up a recognition hierarchy until they become objects or properties that we can articulate. In addition, they pass thru to the frontal cortex where decisions get made, but we are only conscious of the inputs and the outputs of the decision process.

10 working assumptions

At the end of the book. It is hard to say which of the many assertions are actually the numbered assumptions, but who am I to quibble?

1A) Forward Projections are strong to the front of the cortex. Feedback modulates feed forward.
1)Nonconscious Homonunculus: We are not conscious of the highest levels of cognition.

2) Zombie Agents handle much of our actions. Anything we do unconsciously, things we train for. Things handled by the Dorsal pathways & "gist perception"
Consciousness handles the exceptions (See Hawkins) and it presents an executive summary to the planning stages.

3) Conscious precepts are the results of a single winning coalition of neurons with at some prefrontal parts of the network

4) Explicit representation of some stimulus is a set of neurons. Lose those and you no longer perceive that stimulus.
The cortex is a large set of nodes. A "column" is the smallest useful node.

5) Net wave from V1 up hierarchy (See Van Essen) to the prefrontal cortex. Then it travels back down the hierarchy as feedback. Visual Consciousness probably starts at teh upper stages of the ventral pathway in teh Inferior Temporal cortex. (Ventral congizes, Dorsal does)

6) Driving and modulatory loops: To understand coalitions we must understand the neural connections. Excitatory cells send driving or modulatory charges. From the back of the cortex to the front are mostly driving charges. Loops of driving charges are bad.

7) Perceptual awareness may be a series of snapshots with motion "painted on". Each snap lasts 20-200 milliseconds (ms) (it takes 250 ms to "see" something)

8A) Attention: There are two types of attention: a) bottom up saliency driven and b) top down and volitionally controlled. Bottom up would be a sensory net wave that breaks thru. For example, a ball flying at you.
There is also "gist perception" which allows us to go thru the world on auto pilot. As long as nothing salient happens, we are not really conscious of a lot of our daily activities. Like when we drive home and remember no details of the drive.
8B) Binding: Various attributes of a single object are identified by essential nodes. How these are tied together is the "binding problem". There are three types of binding: 1) lower level, V1 binds location and orientation; 2) trained binding, face, voice and hair mean Grandma; 3) novelty binding - requires top down attention

9) synchrony of action potentials discharge in the 30-60Hz range may help in forming nascent coalitions. Firing in the 4-12 Hz band may be part of snapshot processing.

10) Winning coalitions recruit from the cortex, thalmus, basal ganglia and other networks. In addition to the neurons that are explicitly part of the NCC is the penumbra which includes past associations, background and future plans. The penumbra provides meaning or the potential for meaning which might only happen if the NCC expands to those nodes.

Other notes:

Qualia are a symbolic form of representation within oceans of explicit and implicit information. Why qualia "feel" the way they do remains an enigma.

The bandwidth from the LGN to V1 is 1/10 the bandwidth from V1 to the LGN in cats p.59

Almost enough detail to totally comprehend the circuit diagram for the visual system.

Many forms of memory: associative conditioning (bell & food), procedural (hot to), episodic (recollections), semantic (knowledge). Then there is "working" memory - short term buffers and a special form of short term memory: iconic.

Massive amount of information on teh visual system and its quirks. Systematic ascension of the visual hierarchy (See circuit link below) with most of the little boxes accounted for. Much more discussion of the Ventral and Dorsal pathways.

Talk about visual search and why differences in color, size, form and motion makes things stand out. This agrees to a high degree with the Info Viz Book, which has color, form, motion and spatial positioning.

Great Book!!!

I have many, many more notes in my margins, but, they will have to wait!

Check these out: Circuit Diagram of the Macaque, from from Felleman, D. J. and Van Essen, D. C. (1991) Cerebral Cortex 1:1-47.
A similar circuit diagram exists for the auditory channel: Lloyd Watts (PDF)


draft: September 24, 2005