Likely many of the changes to a neuron can be reversed, so it removes receptors from the dendrites for example and thus becomes less responsive to excitation. So it is certainly possible that sometimes we forget simply because the connection strengths reduce below a level from which you can filter a signal from the noise.

Since memories are stored in a network of neurons, the configuration could indeed come to mean something else over time, which would at least confuse if not always erase the traces of patterns stored there.

A very large amount of the answer is likely to be failure to retrieve rather than full on forgetting though I think. You have to be able to get enough prompts to excite the pattern enough that it’ll fire fully. Aiming to talk more (and indeed learn more) about ways you may be able to help this in the next lap around next year when I write about Memory Retrieval.

The question that came up for me is if the neuron is changed permanently by having information pushed into long term memory then how do we forget things? Do they get overwritten by more recent, more intense experiences? Which in turn change the neuron permanently?