Yes, I completely agree with this. This is my neuroscience explanation (always useful as technobabble, if nothing else): A biomorph has the ego implemented as a neural network running on its biological brain. This network consists of the synaptic connections between the neurons and their strengths. The cortical stack monitors this and stores a copy. Uploading via an egobridge also does this (if there is no stack or if there is a reason to think the stack data is not perfect), using very invasive probes to scan every part of the brain. This process then turns into a "formatting" of the brain so that it is ready to receive another ego imprint. The formatting involves removing the ego-specific neural connectivity and leaving/replacing it with some default blank mindstate. Blanks are essentially passive subhumans, with only a few life support functions and simple behaviours (very creepy to be around). During downloading with an egobridge a new neural connectivity pattern is imposed and - if everything works - the ego wakes up in a new body. Now, how do we format brains? Overall we want to change the synaptic strengths to remove all traces of the previous ego. So the first step is to reset the protein phosophorylation state of all synapses: CaMKII, PKAII and especially PKMz (plus all the proteins they phosphorylate) must be returned to the right levels, activin ß-A, synaptotagmins and CREB needs to be up- or downregulated. This would be a nightmare to do by nanotech (lots of proteins sloshing around in a small, fragile space), so I think the solution is to have bioengineered in a reset function from the start: there is a genetic program that can be activated with the right code chemical (something not normally occurring in humans, like the insect hormone ecdysone). This triggers a set of genes that in turn change the regulatory state of CREB et al., plus produces some short-acting phosphatases that resets the synapses. If we do this, we reset every synapse affected. But we don't want to mess up the brainstem and other key systems, so the above genetic program is only expressed in the forebrain. This means that anything that has been learned in the midbrain and downwards (OK, must add an exception for the cerebellum) will remain - this shouldn't be any personally relevant memories. The egobridge can also locally chemically overrule the reset or send a signal that forces the chemical levels in particular synapses to be at certain levels - this way the blank persona can be installed. This is also at least partially how synaptic states are put into a brain during downloading: a chemical signal sets all synapses to "write mode", the branches of the egobridge and stack nanofibres then signal for each synapse what level it has to be at (either chemically or electrically), and the right levels of persistent kinases and gene expression is triggered. OK, what are the game consequences of this model? 1. There exists a ego erasure chemical. Inject it into the brain of somebody, and he will fade out over the next few minutes. It might very well be different from morph model to model, but it will be documented since egobridges need to be able to send it. Probably not super-efficient as a poison since the substance likely needs to be injected straight into the brain rather than ingested, but a cool plot device. Have an NPC stumble into the PCs room, rapidly fading but trying to tell them the important things before his forgets and starts singing 'Daisy, Daisy'. (Another version would be to make it an optogenetic gene cascade. Shine the right laser frequencies on the brain to erase, read or write. Might be neater) 2. Egobridges usually have the two options 'move' and 'copy'. Users are not supposed to be able to do anything in between, except of course that hacked egobridges or bad software can cause it anyway. So you could in theory make a partial erasure, leaving some traits behind. And a bad egobridge (or morph!) might not be erasing as it should. Leading to egos being downloaded into brains that retain some memories, reflexes, skills or habits (ah, that nasty smoking habit is hiding down there in the midbrain!) Since memories are distributed networks and fairly resilient, they could remain hidden until triggered by the right stimuli. Like a smell (the classic case). 3. Most such errors will not make sense. The concepts and actions involved are not similar across the egos, so they will just be surreal messes (whenever you smell ivory soap you get a vision of Mount Matterhorn and a zebra in your left visual field). This might be like having the negative trait "Neural damage (morph)". 4. Memories in the midbrain and deeper might actually matter. They *should* be impersonal, but might show up as odd reflexes. The tectum has visual and auditory reflexes - maybe something interesting or creepy remains in this "frog brain". The substantia nigra might have a bit of a personality biasing effect ("Why am I so hedonistic in this morph? I'm not normally so excited by gambling?") These levels are also directly tied into direct perception and action, so memories left here are not necessarily chaotic messes, they actually work. Game idea: an agent for some group realizes that he is about the be caught and wiped, so he quickly uses psychosurgery to dump what he knows in the form of pontine neural networks. He is wiped, and a PC is sleeved into his morph. Soon the PC is getting a bit annoyed about the strange facial tics the morph does whenever it sees certain things - why does it always grimace when Cognite is mentioned? After a while somebody notices that the eye blinks actually are Morse code... The agent ego is still dead and wiped, but its ghost is inhabiting the body, trying to report back the findings from its infiltration of Cognite's research lab. Meanwhile some Cognite low level security AI doing a routine check that all loose threads of Operation Emerald Kitten are tied up notices a quality control AI report that the sleeve of that spy is acting up and begins to wonder...