But translating short- into longer-term memory involves more permanent structural changes in the connections (synapses) between the nerve cells. This process involves a cascade of biochemical mechanisms, triggered by the initial increases in neurotransmission resulting eventually in the creation of new proteins. These move to the synapse, modifying its structure. My own lab has spent several decades unravelling the details of this cascade in young chickens. There are good reasons to believe that similar molecular processes operate in humans.
Because a sequence of biochemical steps is required to reshape the synapses in the hours after an important event or experience, it is possible to use drugs or inhibitors to disrupt the process. Indeed it is through the use of such specific drugs that many of the steps have been explained.
Say you find out your partner is having an affair. Would it be possible to excise the moment of discovery from your brain? Disrupting neurotransmission in the minutes after the experience, or blocking protein synthesis in the hours that follow does prevent long-term memory being formed, and the result is amnesia of that event. However, blocking protein synthesis outside this critical period leaves memory unimpaired. So the conventional wisdom, at least since the 1960s, has been that once a memory is "fixed" biochemically, it is permanent and cannot be erased.