The answer is "yes it is possible, but it would be unpredictable".
This has a hard science tag, so this draft might need a few revisions, including references, which I will attend to as time permits. Be aware that the best hard science out there is still very speculative, just marginally better than reading tea leaves
Memories are mediated by the hippocampus and mapped into other areas of the brain.
For a long time, the hippocampus was believed to be the seat of long term memory, because damage or removal of the hippocampus would result in instant and irreversible amnesia. However, by adding the use of functional brain scans to what has been determined through decade of experiment and modelling, we now know there is much more to the story.
Be aware the the best functional brain scan technology only has a 3 mm resolution today (2019). That is pretty coarse for a structure that is only 3 cm³ in volume, so our understanding is still very blurry.
Functional brain scans show a physical mapping process occurring during both memory encoding and reconstruction. The word reconstruction is used deliberately rather than recall because the process of recall is as important as the process of encoding.
Sidebar ...be patient and you'll see how his fits.
During execution of spatial tasks, regions of the hippocampus activate in response to
achieving decision points in the task. For example, if the task is a maze, then at
each decision point a different set of physical regions of the hippocampus will activate.
What is very interesting is that during execution of non-spatial tasks, like
solving a math problem, the hippocampus exhibits exactly the same kind of activity.
We have known for a long time that the hippocampus is active in long term recall. With the advent of readily available brain scan techniques, we can see that its role in long term recall is active rather than passive, and essentially identical to its activity in solving spatial problems.
We also know that the hippocampus is crucial to forming new memories. When new data is first introduced to the brain, the hippocampus is recruited, and exhibits the same sort of activity that it does in solving spatial problems.
In effect, the hippocampus is a spatial mapping organ that has been adapted into a general problem solving tool. Think of it as being similar to the GPU in a modern graphics card, specialized for some function, but over time has been adapted to solve a broad range of problems unrelated to its original intent.
For a long time, psychologists have known about the bookend effect. People tend to remember the first time they did something, the last time they did something, but not much in between. This why the first time you drive a route it seems so much longer than subsequent times.
* often called intelligent compression or single-instance storage
* is a process that eliminates redundant copies of data and reduces storage
overhead. Data deduplication techniques ensure that only one unique instance of data
is retained on storage media, such as disk, flash or tape
As near as we can tell (at this point its an educated guess backed by experiment and modelling), during memory encoding the hippocampus mediates an aggressive deduplication process. That process compartmentalizes the new memory into fragments representing the state of the brain at that point in time and builds a map of some sort (As a database person I think of it as a key) to tie the fragments back together.
We do not keep detailed memories at all, only statistically reinforced probabilities (described as the sum of the activation functions for the involved neurons - I don't write that kind of math, sorry).
This actually introduces the specific technical issues to be resolved
to "implant" false memories, and also why the results might be unpredictable
no matter how good the tech is.
During recall, the hippocampus is recruited to find and reassemble the memory fragments into a comprehensive whole, again treating it as a spatial navigation problem.
This is why human memory is so fallible.
Examples of this mechanism in action:
My wife often wears a blue dress to church. We get separated, so when I go looking for her and have trouble finding her because she's wearing a green dress today. The reconstructed memory has been statistically reinforced to blue as part of the deduplication and reconstruction process.
"Oh, I misread that"
The well known phenomenon that no two eyewitnesses of an event recall it exactly the same.
Memory and Emotion
From experimental and clinical studies we know that impairment of emotion processing also inhibits memory encoding
It turns out that you can't really separate memory from emotion, because part of the memory encoding process includes a partial state of the amygdala.
From an experiential perspective, I remember what it felt like to really enjoy a good fireworks show. After brain damage impairing emotional processing, (1) I find it difficult to encode new memories and (2) The only "feeling" i get from the same fireworks show is the sensation of the booms rattling my internal organs, kind of like indigestion.
What is interesting (to me) is that recall causes the same areas of the brain to be activated as the original event (with the caveat that the encoding process is lossy) so there is little difference in the state of the brain during recall as opposed to the original event. The fact that I can recall (reconstruct) emotional states that I can no longer experience in the immediate says that there is a different mediator for recalled emotion versus immediate emotion.