Micrograph of neurons in rat neocortex stained with a dye. These neurons will randomly connect to each other at the presentation of a stimulus and then prune out the weaker connections to set up a working circuit that will activate to the stimulus in the future. A great example of learning at the circuit level in the brain. Neurons are constantly forming new connections and pruning out the weaker ones to be able to more efficiently send messages along and encode new things. This is a case where more is not better- efficiency relies on having the right number, which is why the strongest connections are maintained and the weaker ones destroyed. If connections are not used for a long time, they can become weaker and this is why abilities or facts that you used to know really well can be lost. For instance, for classes you’ll memorize something- say, the periodic table- but not use it again for a long time and the memory will become fuzzy or lost (yeah, I can’t recall all the elements anymore). Note: “muscle memory” is different, which is the root of the always knowing how to ride a bike saying.
[Image Source]