Re: Retrieving & Recreating Stored Memories

Anders Sandberg (
Mon, 21 Oct 1996 12:32:33 +0200 (MET DST)

On Sun, 20 Oct 1996, Chris Hind wrote:

> Alright. Anyone here seen the movie Strange Days? How possible is it that we
> could record all of our sensory input and store it as a file and how large
> would say a 1 minute file be?

A rough estimate (I don't have my books nearby) of the sensory input is a
a few gigabits per second. [The human visual nerve has around a million
axons, firing at around 1000 Hz, and the spinal cord is on the same
order]. That would give an estimate of around 10-100 gigabits for one
minute storage. However, I am sure this can be decreased quite a bit, both
because there are inherent correlations that can be compressed, and
because other estimates give lower values (they are in the megabit range,
and have support in the literature). Of this information, most would be
touch and vision, with smaller amounts of auditory information and a
trickle of taste, smell and other senses.

> Also, would it be possible to go back into the
> brain and retrieve all our old memories in sequence and then be able to play
> them back and reexperience them?

Hard. Episodic memory doesn't seem to be stored in any nice, sequential
manner in any particular region of the brain (although some regions are
of course more important than others); it is likely that our memories are
distriobuted over large parts of our neural nets, mixed with each other.

Besides, we do not remember sensory information. What we remember is
rather our reactions to them, our elaborations, associations and actions,
not exactly what happens. Memory is actually very unreliable, and
continually recreated.

[Yes, I'm studying cognitive science right now. I'm planning to write an
"Owners Guide to the Human Nervous System" when I get the time]

> Also what about the possibility of being
> able to communicate realtime feelings or stored memories direct to another
> individual with a brain-computer-brain interface?

Feelings might be transferrable, at least the basic data of arousal,
activity in the mesolimbic regions etc; if combined with higher-level
means of interaction I think we could transfer quite a bit of our
feelings this way.

Memories are much harder, as explained above, and the same goes for
cognitive structures. Language or verbalized thought might be a
bottleneck, but it is easy to transfer. Our personal mental languages are
so different that they would be incomprehensible (Can anybody guess what a
certain scaly, golden scorpion-like fractal that the viewer flies along
represents to me?).

Anders Sandberg Towards Ascension!
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y