But whenever I read these, I always think they might be getting it largely wrong. In fact, they themselves admit to it being partially speculative information.
I am by no means an expert, but I was neuroscience major in undergrad, and back then (~5-7 years ago) it wasn't known exactly for sure how it works besides the fact that a part of hippocampus interacts with other parts of the brain and that certain things are reinforced with more learning.
Most of the well known and still relevant studies in the field look at people that have a particular part of the brain damaged and how it affected them. Obviously, there aren't too many of such patients.
There are also studies in mice and other animals, but their brains work differently sometimes and other times there are a lot of unrelated variables.
To give you an idea of where the field was a few years ago: I was a research assistant in a neuro-psych lab and we were giving animals drugs in large amounts and then checking whether the cells in a specific region of the brain are damaged. If the CA1 area of hipoccampus was damaged, we could say that it potentially affects memory formation. That was checked by the ability of a rat to solve puzzles it previously solved. Very crude, unfortunately.
http://www.ncbi.nlm.nih.gov/pubmed/15746236
Or consider the fact that the cerebellum is a lot more crucial in global brain function than was previously taught. It's really exciting, cutting-edge work:
https://en.wikipedia.org/wiki/Cerebellar_cognitive_affective...
Somewhat tangentially (but I can't resist), this excellent story by Ted Chiang explores the transience of memory, and the role technology might play in sustaining it: https://subterraneanpress.com/magazine/fall_2013/the_truth_o...
Brains don't store "knowledge" like SSDs by capturing discrete electrons then poking them to read the stored data. Brains encode data physically as the arrangement of synapses between neurons. Brain's aren't just electrical—they are electrochemical.
Important point: knowledge/memory is stored as synapse configurations, not by shoving information inside neurons themselves.
As for suggesting "bizarre quaternary structure of DNA to store the memory ..." that's probably more for wishful new age crystal healing websites than anything based in reality.
Back to your original query: let's not forget The Source Of All Knowledge happily fields such questions: https://www.google.com/search?q=a+good+article+on+how+memory...
[0] https://medium.com/precision-medicine/how-big-is-the-human-g...