The Visual Front End

The Model

[Presented at the Society for Computers in Psychology's 30th annual meeting]

There seems a great rush to understand the nature of consciousness in the last 20 to 30 years.  Some of the selections include Hofstadter's (1980) original and fascinating Godel, Escher, Bach and Dennett's (1991) challenging and provocative Consciousness Explained.  What both these authors have in common is an argument that consciousness is a product of brain.  It is a mechanized phenomenon.  That is perhaps not quite fair to their positions but it does put them in their camp.  On the other side of the field is Searle (1980) and his controversial and challenging Chinese Room and Penrose's (1989) The Emporer's New Mind which uses many of the same issues raised by Hofstader.  These people believe that human intelligence and consciousness is not mechanizable in the sense that our mind is a fancy computer.  It cannot be reduced to a calculating machine.  Their argument vary and are regularly colorful.  However, for me it is a bit like putting the cart before the horse.

For me, I want desperately to know material our mind, whatever that is, works on.  In particular I want to know how we know there is an outside world and the nature of that information, for want of a better word.  Ultimately our perceptions are what we experience, but the early stages look current ideas about the output of the eye and ear and other senses presents a very strange picture relative to the nice view that our perceptions are directly what we experience.  So how do we get from these early stages to our perceptions?  To begin to answer this questions, I want to look more closely at the nature of the these early stages.  What is really is the output of these early stages.  Is the information recognizable or is it so distorted that we would never make sense of it?  What is the best way to represent and consider these output.

In fact, I am most interested in starting at the retinal output.  This one aspect of the visual system fascinates me.  It is a bottle neck and just as the bottlenecks of cognition, such as attention and short-term or working memory (Solso, 1998), represent key functions to understand so I think this bottleneck is very important.  Think about it.  We have over 100 million receptors in each eye, over 7 million cones alone.  We have about that many cells in our visual cortex, but only just over 1 million ganglion cells making up the optic nerve.  There is an obvious bottleneck.  A very effective bottle neck or we would not get around so easily but it is there non-the-less.  The information in the image in the eye is compressed but that compression does not come without cost.  Whatever shortcuts are taken here have to impact all of perception.  The mistakes at this level cannot be undone. 

These pages contain my speculations on the nature of this front end of the visual system where compression is king.  By front end I am referring to what comes out of the eye.  I don't intend these pages to be a comprehensive review of the existing literature but a synthesis that brings together ideas to yield a synthesis of what is known.  I doubt there will be a novel idea here but you never know.  I just needed a place to collect my thoughts.   So at first the ideas will be random and uncoordintated but hopefully will assume an organization over time.

Currently some of the issues I want to work on are:

bulletThe background for my ideas, including my assumptions 
bulletThe questions that I want to address as part of this project 
bulletSome of the methods I hope to or am using to address these questions
bulletA quantitative model of the fovea which has a graphic depiction of the output of the eye

Those links that are active have some content.  If you have comments or suggestions, I would love to hear themYou can see comments by others here, if there are any.

Last update: November 19, 2003

Hit Counter