The brain is not a passive system that sits there waiting to respond to events in the world. If anything, it is living in the near future, always predicting what is about to happen, what other people are about to do or say.
Its predictions are built up from previous experience: the actions that have been produced in different situations, the sights and sounds that tend to be perceived, the way the world has responded to previous actions, the events that have tended to follow each other, the rewards and otherwise that have tended to accrue from different behaviours, how people tend to behave in certain situations, likely narratives, what goals other people have when they produce certain behaviours.
The predictions span different scales of time: from the level of ‘Jonny is likely to choose a strawberry ice cream, because I’ve had ice cream with Jonny before’; to what word is likely to end a sentence, and where the football will be a second from now. Multiple parts of the brain are tracking what just happened and what that means for what might happen next.
A developed brain has thousands of hours of experience imprinted on its connections. Each situation will re-activate scripts and prompt goals (‘I’m in a restaurant, I need to look at the menu!’). Perceived objects will prompt actions based on their previous use: light switches will prompt the action of turning them on, coffee mugs prompt being lifted to the lips. It’s the role of the modulatory system to decide between these options, to select among them according to current goals. When the modulatory system doesn’t take command, these behaviours can happen on their own. Idle hands fiddle. Lights get switched on. Coffee mugs lifted.
Why would the brain work this way? There are two reasons.
First, it makes the brain faster to respond. Time is often of the essence in biological systems. The brain doesn’t work fast from a standing start – like when you awake and open your eyes in the morning, it takes a moment to orient to the room around you. Instead, the brain is always in the game, primed for the most likely next move, what it’s likely to see next, getting relevant circuits ready to fire.
Second, predictions optimise the brain to learn from when its predictions are wrong. Surprise is a powerful teacher. Curiosity and exploration are about playing with the world, poking it, seeing if it responds in the way you expect, and learning when it doesn’t.
What in the brain’s structure allows it to use prediction in this powerful way? The answer is the way it is connected. When there is a connection from A to B, there is often a connection back from B to A. Where there is forth, there is back, where this is up there is down. Where there is hokey there is cokey. Recall we talked of the sensory systems as towers, with lower floors seeing detail, and each higher floor seeing larger patterns in the detail, patterns within patterns, until at the top, the system is recognising objects and actions in the world. In the tower, information from the senses is passed up the tower. But as the higher floors begin to get a sense of what may be out there in the world, they pass this information back down again, to help the lower floors make more sense of the basic sensory signal. It is these ‘top-down’ connections that allow expectations and previous knowledge to influence perception and action.
Having these back and forth connections is not without risk. It allows for the possibility of uncontrolled build up of activation. A excites B, B excites A, which gets B more excited, which excites A, and so on, until neurons burn out. The brain has to use a lot of damping down to stop such activation build up. But even then, there are risks of uncontrolled activation flow, such as that observed in epilepsy. The risks are highest in childhood, when the levels of brain connectivity are at their greatest.
The flow of information up and down the towers also requires a balancing act. On the one hand, if your expectations rule what you perceive, you can miss subtle things changing in the world. (For example, watch this video and see if you can work out who committed the murder. Is your attention up to it?)
If you don’t get the balance right, at worst, you’ll suffer delusions and hallucinations: you’ll see your ideas and not reality. On the other hand, if your perceptions are ruled by the sensory input and are dismissive of previous experience, the world will look strange and perpetually new, nothing like you expect, all trees and no forest – this is an experience that has sometimes been used to describe perception in autism.
The price of speed, the price of learning from surprise, is to keep the brain in a perpetual balancing act.