The Guess Worker

How does a dog's brain associate the sound of the bell and food? That's easy isn't it? When the neurons of the two stimuli fire together, they wire together. The snag is in the real world it doesn't seem to work like that.

Surprise result

Suppose you are trying to condition two dogs to associate food and a bell. The first animal always hears the bell one minute before it receives food. The second animal is always presented with the two stimuli simultaneously. Which animal will you condition the fastest?

You might expect it to be the second dog because the two stimuli are activating its neurons at the same time. But, in fact, the first dog will be conditioned faster.*1

It seems that neurons which fire together don't wire together very well. Instead the experiment appears to show that neurons which fire at different times wire together the best.

Prime time

What could be going on? Of course there's little point in wiring together neurons which fire at widely separated times. The stimuli which activate these neurons are not likely to be causally related. It only makes sense for neurons to connect up if their stimuli occur at around the same time.

But how do neurons which are activated a short time apart get connected? One idea is that neurons prime each other.*2 When one neuron fires it physiologically modifies a second neuron, which prepares the second neuron to connect up to the first as soon as the second neuron fires.

Perhaps. But how would the first neuron be able to anticipate which neuron was going to fire second? It could be that when a neuron fires, it primes all other neurons, so that when one of them happens to fire, they can connect up. But it does seem wasteful of energy to prime such a lot of neurons in order to make only one linkage.

Stream of consciousness

There is a simpler, much less wasteful possibility. It ditches the idea that neurons prime each other. Only neurons which fire at exactly the same time as each other wire together. But if that's true, how could neurons which fire at different times become connected? The answer is by using other neurons as bridges to span the time gap.

Let's imagine the sequence of events as you condition the first dog in the experiment. From moment the dog hears the bell ring to the time it receives food it is conscious of a stream of information. When it hears the sound it looks towards the bell at the top of the door. A minute later the dog notices the flap at the bottom of the door opening. Then it sees food being pushed through the flap. So after hearing the bell, the dog is conscious of visual information in the following order: 1) the bell, 2) the door, 3) the closed flap, 4) the open flap, 5) the food.*3

Bridge building

All information which an animal is conscious of triggers dopaminergic neurons in the brain. If the dopaminergic neurons don't make the animal feel a more powerful emotion, they at least make it feel interest. The pleasure and pain of interest motivate the animal to search for more information.*4 When the dog hears the sound of the bell it begins a search to find out where the sound is coming from. Its consciousness switches rapidly between the sound and the sight of the bell and a linkage is formed between these stimuli.

The sight of the bell is also information which provokes interest and so when the bell stops ringing the dog doesn't immediately stop looking at the bell but switches its attention between it and 2). A linkage forms between the neurons fired by the information in 1) and 2). In the same way the dog makes linkages between 2) and 3), 3) and 4) and finally between 4) and 5). The neuron for 5) - the sight of food - is already strongly linked to the motor neuron which triggers salivation.

Once formed these linkages look like this:

Neurons have linked up to make a bridge connecting the sound of the bell and the sight of the food. *5 The next time the bell is rung some of the impulses from the bell neuron will be diverted through the chain of neurons towards 5) and onwards towards the motor neuron for salivation.

Fast track

When the impulses activate the neuron for the sight of food, the dog will feel the same pleasure as it feels when it sees food. In other words, the sound of the bell will evoke the memory of food. A memory, though, is not the same as a reflex. Once the dog is fully conditioned it will react reflexively: it will start salivating the moment it hears the bell. How does its conscious memory become an unconscious reflex?

At the beginning of the conditioning process the linkages in the bridge are weak. Most of the impulses still will be firing the dopaminergic neurons along the bridge and few will reach reach the food and motor neurons. The dog will only have a faint memory of the food and it probably won't salivate.

But because neurons which fire together wire together, each subsequent time the bell is rung and food is given, the linkages will strengthen. Once fully conditioned the dog's neurons look like this:

Now the linkages are so strong that all the impulses reach the motor neuron. The dopaminergic neurons along the bridge are not firing because none of the impulses get to them. That makes the dog's response unconscious: it no longer feels the interest that these neurons generate. The response will also be fast. It's as if an express line has been built on the tracks of a local train route: the train doesn't get delayed any more by stopping at the small stations on the way. Impulses pass quickly and unimpeded along the bridge. The dog's response has become a reflex. *6

Poor connections

Time bridges are made by simultaneous firing of neurons, not by neurons priming other neurons. But if simultaneous firing is the way neurons connect up, you might reasonably expect stimuli which are presented at the same time to become associated the most readily. So why don't they?

It has to do with salience.*7 The brain will always fix its attention on the most salient stimulus around. For a dog in an unchanging cage without food, the sound of a bell will be that most salient stimulus. The neurons activated by it connect up to much less salient stimuli in the dog's surroundings. They would connect up to more salient stimuli if there were any, but there aren't.

Food, of course, is not only much more salient than these stimuli but also much more salient than the bell. What happens when the bell and the food are presented at the same time? It's much the same as when you get two presents on your birthday: a pair of socks and a new car. You're not going to spend much time thinking about the socks, are you? Like you, the dog will spend most of its conscious time on the more salient stimulus, which in its case is the food. Its brain will focus on what the food looks like, what it smells like, on its position and how to get to it. That's so the dog can form concepts which help it eat. The dog's consciousness will spend very little time on the less salient bell and so the linkages between the bell and the food will be weak.

The weakest linkage

Salience also explains how time bridges can become very strong very quickly. If nothing much salient in the dog's surroundings happens, the weak linkages made by the dog's consciousness quickly break down, just as most linkages in so-called "short-term memory" do. *8 The salience of food, though, makes the linkages stronger and more durable. Interest in the sight of the food motivates the animal to search for other information, which is likely to be the nearby open flap. Because of the relatively intense pain and pleasure, the dog's consciousness switches many times between the stimuli 5) and 4) causing a strong linkage to form between them. And because of this strong linkage, stimulus 4) becomes more salient than it otherwise would be, which means that in time it will form a stronger linkage with 3).

The same will be true at the other end of the bridge. Although the sound of the bell is far less salient than food, it is more salient than the rest of the stimuli around. The linkage between the sound of the bell and 1) will therefore become strong relatively quickly. The weakest linkage in the bridge - and the last to strengthen - will be in the middle, in this case between the neurons for 2) and 3). If the bridge is too long, linkages in the middle won't be able to strengthen sufficiently and the bridge will not be completed. The dog therefore would find it impossible to associate the sound of the bell and food when the time interval between the stimuli is too long.

Useful narrative

Your first love, before it went wrong, was beautiful. Your partner had dark brown eyes and a special smile. Of course in those days you spent a lot of time looking at your sweetheart's face. Then one day, many years later, you meet someone else with almost exactly the same eyes and smile. You feel an instant twang - your stomach does a turn and your legs feel wobbly.

Like this, throughout our lives, our brains turn conscious experiences into instant responses. Time bridges explain why that happens.

There is little evidence for either time bridges or priming neurons. But while priming neurons leave questions unanswered, time bridges solve extra ones. Time bridges answer one of the big mysteries of science and philosophy. Why do we experience our lives like a 3D, five-sense movie? Why in our waking hours are we constantly conscious, when consciousness seems to be entirely useless? The answer is time bridges. Even when information is unimportant, it pays for us to be aware of it. Our brains use awareness of unimportant information to connect up important information separated by time. That, if nothing else, makes consciousness extremely useful.

*1 Simultaneous conditioning (where the two stimuli are presented at the same time) is a less effective procedure than forward conditioning (where the unconditioned stimulus is presented shortly before the conditioned stimulus).

*2. The phrase "neurons which fire together wire together" was probably first used by Carla J. Shatz, an American neurobiologist. The phrase neatly summarises a theory by Donald Hebb (called Hebbian Theory) from 1949 which sets out to explain how neurons enable learning. Despite the phrase, though, Hebbian Theory does not suggest that neurons which fire at exactly the same time wire together. To account for causality, the theory proposes that one neuron excites another neuron so the second neuron can later link up. In other words, Hebb was proposing that neurons primed each other.

*3. For simplicity I've assumed that all this information is visual except for the ringing of the bell.

*4. For an explanation of how this works see Interest

*5. Again for simplicity I've assumed the neurons link up linearly. In reality neurons of the sound and the food are probably connected up by a meshwork of neurons which allow impulses to travel on several alternative routes.

*6. The diagram is not perfect. In the next post Networking, I point out its flaws and try to improve it.

*7. For an explanation of salience see Learning

*8. See Memory