Prediction Neurons initiate action before conscious brain decides…
Neuroscientists have recently discovered that the brain has “Prediction neurons” that learn by trial and error and can initiate action independently of the conscious brain, several seconds before the conscious brain is even aware that action is needed.
The best way to explain how Prediction neurons work is via a simple example.
A small child tries to open a cupboard door by pushing it, beating his hands against it, and finally pulling on the handle. Aha, when he pulls the handle he succeeds! This success triggers a burst of dopamine in the brain, which feels good. The Prediction neurons associate pulling the handle with a positive result and “learn” that pulling a handle will open the door.
Now, let’s say that later the child sees a handle with a latch under it. He pulls it as he did before but it won’t open. He receives no positive dopamine release and instead gets a negative feeling of frustration. He fiddles with the latch until he discovers how to open it. When that happens, the Prediction Neurons for that action get a burst of dopamine, feel good and “learn” the appropriate action for that type of device.
Prediction Neurons learn by trial and error, updating with each trial
The important thing about Prediction Neurons is that they learn by trial and error, forming schemas of how things work and updating that information with each trial. They are learning all the time.
This enables us to adapt very quickly to changing situations and to perform common tasks without the conscious brain having to decide what to do each time. That’s fast and efficient.
But here’s the problem: if an object or situation looks like something the brain has a schema for (e.g. Pull a handle to open), but it doesn’t follow that schema, the Prediction Neurons will initiate the action they predict is required, but that action will result in an error. The Prediction neurons will update their predictions based on that error, but if there is inconsistency in design (e.g. some doors with handles are pulled to open and other are not), then this action will continue to be error-prone. Putting up a sign will typically not solve the problem because signs are processed by the conscious brain and the Prediction neurons work faster than the conscious brain. It would be better to have a consistent design – e.g. doors that pull to open have large handles, doors that push to open have plates or small knobs instead of handles.
Key Point: When designing equipment, devices, work areas, packaging, paperwork and other things that humans interact with, analyze the human schemas involved. Where possible ensure the design is consistent with existing schema. If this is not possible, as a minimum ensure that an action initiated according to the existing schema would not cause adverse consequences. E.g. ensure that pulling the handle does not result in damage or injury. For more information email brainwaves@talsico.com
|
