You are currently browsing the thoughts on thoughts weblog archives for February, 2009.

February 2009
« Jan   Mar »

Archive for February 2009

Analog computers

An analog computer is a model of what is to be computed. It is built of mechanical, hydraulic or electrical devices to form a simulation of a problem. A mechanical example is the astrolabe which was built by the ancients to predict the heavens. A famous hydraulic example was MONIAC which modeled the UK economy. More usual is the general electronic analog computer that can have different problems setup using a patch board to connect various devices to create the model of each individual problem.


Analog computers are very fast, next to instantaneous, and can record dynamic processes as they are modeled, usually faster than the real processes that are being investigated. This can be thought of as solving a problem that consists of a large number of constraints. What is the one answer that is available while satisfying all the constraints? If all the constraints are physically modeled and exist simultaneously then the answer is simply there – for any input there is an output. In this speed they are very similar to digital computers that have a lot of parallel processors. In fact if there are enough processors to model all the constraints at the same time, then the digital computer has became essentially the same as an analog one.


A type of computing system that is like an analog one or a massively parallel digital one is a neural net. In a neural net there are a large number of processors and they are each connected to all the others. But each connection has a weight that can be varied from effectively no connection to fully connected. These individual strengths of connection between each pair of processors are like a set of constraints. Each particular pattern of input gives a particular pattern of output. If weights are changed, a different output results. Neural nets and digital programs that mimic neural nets are used on problems like speech and pattern recognition and simulations of brain functions. Not only are neural nets fast and parallel but they can be made so that weighting changes are automated to give a learning computer.


Stepwise algorithms, sequential programs, on conventional computers are nothing like a brain. We have to look at analog, parallel and networked computers to see something that resembles the brain.

A different type of computer



Once, long ago, I encountered a discussion on different ways of computing – I cannot remember who wrote the article or when/where it was. Two images have stayed in my mind.


If you have a lot of broken spaghetti and you want to find the longest piece, you could do three things. One, you could measure each piece with a ruler, number each piece and tabulate the lengths as you go. After this you would scan the values you recorded for the largest value, note its number and retrieve the piece. Two, you could pick up a piece and compare it to another piece, discard the shortest and continue comparing to another piece. When all pieces but the one you are currently holding are in the discard pile, you have the longest piece. Or three, you could gather up the whole lot of spaghetti in your hands and hold them in an upright bundle. You would then tap the bundle lightly on the table so that all the pieces rested on the table. The longest piece would stand out as the tallest and could be picked out easily. Method three is very fast and easy when compared to the others. It is like an analog calculation rather than a digital one.


The other example I remember was finding the centre of three points – the point which is equal distant from all three points or the centre of the circle that passes through all three points. It is possible to construct, with compass and straight edge, the line that is equally distant between each pair of points. There will be a point where the three lines cross and that is the point you want. Another way is to attach an identical spring to each point and then bring the other ends of the three springs together and attach them all to the same little vertical stake. Where the stake rests is the centre point as the three springs will be identically extended. Unless you have to go out and buy the springs etc. but do not have to buy a compass, the spring method is instantaneous. Again it is like the different between an analog calculation and a digital one.


When we think of the brain as a computer, we have to be careful about what type of computer we have in mind, as well as remembering that this is just a rough metaphor.

A framework

We have a framework for our conscious experiences. This framework is innate so we don’t have much choice about its nature.

There is three dimensional space in our consciousness. No one seems to live in a flat two dimensional world or a four dimensional one. We all seem to locate points in space using the coordinates up-down, left-right, and forward-backward. Physicists can tell us there are more than three dimensions (four or eleven or ten and a quarter); they can show the mathematical proofs, but no one can actually experience anything but three dimensions.

The organs that we use to trace orientation and acceleration are the semi-circular channels of the ears and there are three of them at 90 degrees to one another. This may be why we are incapable of any other number of dimensions.

There is an arrow of time in our consciousness. We seem to have a continuum of time that consists of the past and the future which are connected by a now. Our experiences are located along this continuum. Usually we see the continuum as similar to the forward-backward dimension of space. Depending on culture and situation, time either flows past us while we stand still or we move through time while it remains stationary.

Time is very important to the sequencing of actions, the notion of cause and effect, and the understanding of processes.

There are discrete objects in our consciousness. According to physics, things are not so self-contained and separate from other things as we experience them. We are attuned to edges, surfaces, complete shapes and what moves together as one. Our experience is filled with objects. Not just sight and touch give us objects – even hearing, in the blind, can give us a sense of the space we are in and at least some of the objects in that space. It is probable that we come into this world with some rough categories for objects. Faces, for example, are a special category of object.

I am sure there are other interesting aspects of an innate framework for our consciousness. The point I am trying to make here is that such a framework exists and it may or may not correspond to reality. But if it is not absolutely accurate, it was good enough to allow our fore-bearers to live successfully. 

On the way to artificial conscousness

Discovery magazine had an interview by Susan Kruglinski with Gerald Edelman, What Makes You Uniquely ‘You’? (here). Here is another part of Edelman’s ideas – how to make artificial consciousness.

“The cortex is responsible for a good degree of the contents of consciousness, and if I take out an awful lot of cortex, there gets to be a point where it’s debatable as to whether you’re conscious or not.

For example, there are some people who claim that babies born without much cortex—a condition called hydran­encephaly—are still conscious because they have their midbrain. It doesn’t seem very likely. There’s a special interaction between the cortex and the thalamus, this walnut-size relay system that maps all senses except smell into the cortex. If certain parts of the thalamo­cortical system are destroyed, you are in a chronic vegetative state; you don’t have consciousness. That does not mean consciousness is in the thalamus, though.

If you touch a hot stove, you pull your finger away, and then you become conscious of pain, right? So the problem is this: No one is saying that consciousness is what causes you to instantly pull your finger away. That’s a set of reflexes. But consciousness sure gives you a lesson, doesn’t it? You’re not going to go near a stove again. As William James pointed out, consciousness is a process, not a thing.”

I like the reference to the thalamocortical system!!


“If we ever create a conscious artifact, it won’t be living. That might horrify some people. How can you have consciousness in something that isn’t alive? There are people who are dualists, who think that to be conscious is to have some kind of special immaterial agency that is outside of science. The soul, floating free—all of that.

There might be people who say, “If you make it conscious, you just increase the amount of suffering in this world.” They think that consciousness is what differentiates you or allows you to have a specific set of beliefs and values. You have to remind yourself that the body and brain of this artifact will not be a human being. It will have a unique body and brain, and it will be quite different from us…

We construct what we call brain-based devices, or BBDs, which will be increasingly useful in understanding how the brain works and modeling the brain. They may also be the beginning of the design of truly intelligent machines.

It looks like maybe a robot, R2-D2 almost. But it isn’t a robot, because it’s not run by an artificial intelligence [AI] program of logic. It’s run by an artificial brain modeled on the vertebrate or mammalian brain. Where it differs from a real brain, aside from being simulated in a computer, is in the number of neurons. Compared with, let’s say, 30 billion neurons and a million billion connections in the human cortex alone, the most complex brain-based devices presently have less than a million neurons and maybe up to 10 million or so synapses, the space across which nerve impulses pass from one neuron to another. Our brain-based device learned to pick up a ball and kick it back to a human colleague. It did not just execute algorithms…

An artificial intelligence program is algorithmic: You write a series of instructions that are based on conditionals, and you anticipate what the problems might be. AI robot soccer players make mistakes because you can’t possibly anticipate every possible scenario on a field. Instead of writing algorithms, we have our BBDs play sample games and learn, just the way you train your dog to do tricks.

At the invitation of the Defense Advanced Research Projects Agency, we incorporated a brain of the kind that we were just talking about into a Segway transporter. And we played a match of soccer against Carnegie Mellon University, which worked with an AI-based Segway. We won five games out of five. That’s because our device learned to pick up a ball and kick it back to a human colleague. It learned the colors of its teammates. It did not just execute algorithms.

It seems to work better to build in a sort of consciousness.


“The brain can be simulated on a computer, but when you interface a BBD with the real world, it has the same old problem: The input is ambiguous and complex. What is the best way for the BBD to respond? Neural Darwinism explains how to solve the problem. On our computers we can trace all of the simulated neuronal connections during anything the BBD does. Every 200 milliseconds after the behavior, we ask: What was firing? What was connected? Using mathematical techniques we can actually see the whole thing converge to an output. Of course we are not working with a real brain, but it’s a hint as to what we might need to do to understand real brains.

Eugene Izhikevitch [a mathematician at the Neurosciences Institute] and I have made a model with a million simulated neurons and almost half a billion synapses, all connected through neuronal anatomy equivalent to that of a cat brain. What we find, to our delight, is that it has intrinsic activity. Up until now our BBDs had activity only when they confronted the world, when they saw input signals. In between signals, they went dark. But this damn thing now fires on its own continually. The second thing is, it has beta waves and gamma waves just like the regular cortex—what you would see if you did an electroencephalogram. Third of all, it has a rest state. That is, when you don’t stimulate it, the whole population of neurons stray back and forth, as has been described by scientists in human beings who aren’t thinking of anything.

In other words, our device has some lovely properties that are necessary to the idea of a conscious artifact. It has that property of indwelling activity. So the brain is already speaking to itself. That’s a very important concept for consciousness.”

If you can make something – you can understand it. This work is on the way to understanding consciousness.

Other definitions

Discovery magazine had an interview by Susan Kruglinski with Gerald Edelman, What Makes You Uniquely ‘You’? (here). There are some interesting definitions in it that I comment on below.


What is consciousness? He gives William James, “It is a process, and it involves awareness. It’s what you lose when you fall into a deep, dreamless slumber and what you regain when you wake up. It is continuous and changing. Finally, consciousness is modulated or modified by attention, so it’s not exhaustive.” This seems a very clear and fair definition.

What is the evolutionary advantage of consciousness? “Consciousness allows you the capacity to plan.” Well yes, planning is maybe the most important, but also it allows you to remember in a way that allows particular types of learning.

Are animals conscious? “There is every indirect indication that a dog is conscious—its anatomy and its nervous system organization are very similar to ours. It sleeps and its eyelids flutter during REM sleep. It acts as if it’s conscious, right?… It’s the experience of a unitary scene in a period of seconds, at most, which I call the remembered present.” He then goes on to make a distinction between primary consciousness, the ‘remembered present’ above, and consciousness of consciousness. He feels animals have only primary consciousness and therefore has no narrative history of the part or projected future plans. To my mind, this is too simplistic and all-or-nothing. There is a lot of evidence that animals vary in the amount and sophistication of their projected plans and remembered history. Some animals would have the sort of consciousness that Edelman describes, some more, some less.

How does this primary consciousness contrast with the self-consciousness that seems to define people? “Humans are conscious of being conscious, and our memories, strung together into past and future narratives, use semantics and syntax, a true language. We are the only species with true language, and we have this higher-order consciousness in its greatest form.” I think he is on the mark in the importance of language to human consciousness – it is the significant difference between our awareness and that of other animals. But we have a chicken and egg problem here. If we are looking at the evolution of human consciousness, language is important. But if we are looking at the evolution of language, then many skills that we association with language have to pre-date it and evolved with their own functionality. Otherwise either the language or the thought processes have to be a huge leap. Animals can use symbolism, or sequential structures, or actor-action-object categorization or other aspects we think of as parts of language as basic cognitive tools.


Edelman has very interesting things to say about artificial consciousness as opposed to programmed AI in robots and I will probably return to this in a future post.

Explicit and implicit memory

Memories can be divided into two types: implicit and explicit. Only the explicit memories require conscious attention when they are formed and produce conscious memories when they are retrieved. Implicit memories are formed without our awareness and either not retrieved in the normal sense or retrieved as ‘guesses’. Implicit memories may come in various types: a procedural memory like the skill of riding a bicycle; a perceptual memory that allows us to guess at something we have perceived when we are primed with a small part of the perception given in the same sensory modality as the original encounter; and finally, emotional conditioning.


Perceptual implicit memory and explicit memory differ in a number of ways:

1.      They are different types of events in the brain. Implicit storage occurs 200-450 msec. after an event with a negative-going potential in the centroparietal region. Explicit storage occurs 900-1200 msec. after an event with a positive-going potential in the right frontal region.

2.      Damage to the hippocampus interferes with explicit but not implicit memory.

3.      Explicit memory retrieval is accompanied by feelings of remembering, recognition or familiarity and these feeling are absent when implicit memories are retrieved.

4.      Implicit memory is not affected by the depth of processing of the original event whereas explicit memory is stronger if the original event is attended to in more depth.

5.      Implicit memories can not be experienced or described in language. Explicit memories are experienced consciously, can be reported and are holistic with additional aspects that were not part of the retrieval specification.

6.      Implicit memories using priming can be more accurate then explicit memories. However implicit memories are more prone to the illusion-of-truth type of errors.


Explicit memory is semantic or episodic. The picture appears clear that in order to have a conscious memory, the original experience must have been part of consciousness. The formation of these memories probably is an important function of consciousness.

Memories in time

A recent item in Science Daily (here) discusses work on how the hippocampus time-stamps memories.  

“Ironically, Gage and his team had not set out to explain how the brain stores temporal information. Instead they were interested in why adult brains continually spawn new brain cells in the dentate gyrus, the entryway to the hippocampus. The hippocampus, a small seahorse-shaped area of the brain, distributes memory to appropriate storage sections in the brain after readying the information for efficient recall… Each of these newborn neurons undergoes a prolonged maturation process, during which it changes from hyper-excitable to composed and reaches out to mature brain cells that are already well-connected within the established circuitry. Exercise, learning, and environmental enrichment increase proliferation and survival of new neurons, while pathological (chronic) stress and age send their numbers plummeting. Despite an increasing understanding of how new neurons become part of the existing dentate gyrus network, it is still unclear what their exact function is… It quickly became clear that overly excitable youngsters respond indiscriminately to incoming information… But nothing lasts forever. Even the most highly strung nerve cells that used to get excited by just about anything will eventually quiet down. As they mature into fully functional granule cells, they take their place in the existing circuitry while the next generation of newborn neurons takes their place firing away at new events.

Yet, independent events that had nothing in common but the fact that they occurred around the same time will now be connected forever in our minds—explaining why discussing the movie we saw a couple of months ago might bring back the name of the café we visited afterward but whose name has been eluding us.

“Current thinking holds that when we bring up a certain memory, it passes back to the dentate gyrus, which pulls all related bits of information from their offsite storage,” says Gage. “Our hypothesis suggests that cells that were easily excitable bystanders when the memory was formed are engaged as well, providing a hyperlink between all events that happened during their hyperactive youth.”


This would be the source of the life-long narrative that is our ‘life’.


I try very hard to be open to ideas rather than fix on a good one to the exclusion of others. Most people try to do this to a certain extent, of course. I am not saying that I am particularly good at it even. Nor is an open mind always the most effective tool – sometimes people who are in stubborn, blind opposition to one another’s ideas can get further, faster.

But there are ideas that I find it hard to keep at arm’s length. One is that consciousness is the leading edge of memory. It is what becomes, in a fraction of a second, sensory memory, then working memory, then short-term memory, then long-term memories that become progressively more consolidated.

How do we know we have been unconscious? We know because there is a discontinuity in our memory. I was standing and now I am on the floor. It was evening and now it is morning. I was in the house and now am in the garden. I was watching the news and now I’m watching a movie. What happened? I must have lost consciousness – fainted or slept or something else but I definitely was not conscious.

What is the difference between my conscious experience and my memory of an experience? Not much seems different. There is the same feeling of space and time. There is the same feeling of self. There are the same colors and movement and sound. The only difference seems to be that the real thing is more vivid and compelling than the memory. Also the one seems to carry a label saying ‘now’ and the other a label saying ‘then’.

How do you make sure you are going to remember something? When we want to improve our memory we focus our conscious attention on the thing we want to remember. Also, we keep returning the thing to remember to our consciousness. We say a phone number over and over to ourselves. We look again at someone’s face and remember their name to try and make the connection firm. The route to memory is conscious attention.

If we understood more about memory than perhaps we would understand more about consciousness. 

Towards understanding working memory

A paper has recently been published that has everyone buzzing. A Texas group has found a mechanism for short-term memory (here).


There are types of memory - working memory, long-term memory and one or more intermediate memories. Until recently, it was thought that the working memory was in the form of circular networks of cells that could sustain activity in their circuits for short periods of time. This theory was problematic. The new research has shown that certain cells can maintain activity for a minute or more after a short stimulation. This is the sort of time frame that is needed for working memory. It has been compared to the RAM of a computer as opposed to the hard disk. The cells are where short-term memory is expected to reside – the prefrontal cortex.

In short-term memory tasks, individual prefrontal cortical (PFC) neurons can maintain persistent action potential output during delay periods between informative cues and behavioral responses. … Here we used patch-clamp recording of layer 5 PFC pyramidal neurons to identify a postsynaptic depolarization that was evoked by action potential bursts and mediated by metabotropic glutamate receptor 5 (mGluR5). This depolarization occurred in the absence of recurrent synaptic activity… We propose that burst-evoked intrinsic depolarization is a form of short-term cellular memory.


Working memory appears to be involved in the production of consciousness. The identification of these memory cells may soon clarify some aspects of awareness.

Visual imagery

In an item on learning in MindHacks and (here) in a discussion of learning styles is the following observation.

“I have a dear friend, Cat, who doesn’t have visual imagery. When she thinks of a dog, for example, she doesn’t see one in her mind’s eye. She doesn’t see anything. When she dreams she rarely has pictures — she just knows what is happening in the dream. People often don’t believe this. They think that everyone must experience their inner world in pictures, the way they do. Sorry. People are just different. Some always see things when they imagine them, some don’t. Some people have a sense of pitch, some don’t. So it goes…. For example visual imagery: it is not that some people are visual thinkers, it is that most people have some visual imagery and a few have very strong imagery and a few, like my friend Cat, have less than average.”


I take this description out of the context of learning and into the area of consciousness. I think it is clear that the friend is not blind in any sense. She can see a dog and does not walk into doors. There is no reason to think that she cannot tell which things differ in colour and so on. We assume that she is aware of her perception of visual objects but cannot imagine them. The lack of visual imagery in dreams would imply to me (unless of course she reported differently) that her memories were weak on visualization. This is because I assume dreaming is part or a by-product of consolidating memories.


If the same ‘spaces’ and ‘processes’ are used for perception of the present, imagination of the future, and memory of the past then I find the conscious awareness of perceptions/imaginings/memories must be somewhat similar. This is a real puzzle that requires much more information and thought.