[MURG] Uploading via memory implants?

[Eric Zilli]

On Mon, 29 Nov 2004 18:27:05 -0500 (EST), Thomas Weber
<aad1trailmaker at yahoo.ca> wrote:
> 
> Thomas: Isn't that external neuron activity strictly
> related to what the brain / mind is processing at a
> given time? I will give the Mona Lisa example again. I
> suppose when we stare at a specific image - a neural
> activity takes place on the vision area of the cortex
> which is specific to that image. Suppose this activity
> is recorded and observed again on the cortex.
> Emergence of the same activity patterns would indicate
> that the person is staring at Mona Lisa again. This is
> my guess - correct me if I am wrong. Another question
> to you: If my guess is right - except now we don't
> stare at the painting, but just recall and imagine it
> - will the cortex activity be any similar to when we
> stared at the actual painting?
> 

Well, similar: yes. Visual information goes through a lot of
processing at various locations. Staring at the Mona Lisa would cause
certain activity in V1 corresponding to the edges and such of the
picture and as you go up to more complex visual regions, activity of
cells corresponds to more complex constellations of visual stimulus
elements. Recording the full set of this activity, then later imposing
it back on the neurons with the right temporal pattern should indeed
give the viewer the impression they're staring at the Mona Lisa.
Of course, in order to truly capture the sensation of looking at the
painting, you'd have to record and later impose the activity of many
more neurons than the electrode array could feasably record (again:
recording from a sufficient number of neurons using this technique
would just end up destroying the tissue).

Remembering the picture later doesn't bring back this full set of
activity, instead it would trigger activity only (well, mostly only)
in higher visual regions: imagining it doesn't make you sense it
again, rather it activates high level representations of the features
you remember (and also associated features in other memory areas, like
the fact that it's called the Mona Lisa would be activated in/near
Wernicke's area).

> 
> Thomas:
> I understand that some important aspects of the mind
> may be impossible to transfer by the brain - computer
> interface I propose. But at least some aspects can be
> transfered. I envision that at some stage of the
> project the computer counterpart should contain enough
> information - that verbal or typed communication
> should become possible. That vital information which
> cannot be transfered to to the computer via the brain
> interface can be simply communicated by words.
>

I'm not sure personality aspects can be instilled with words. You
might find it interesting to read about Phineas Gage (I think I
spelled it right) who sustained major prefrontal cortex damage but
survived, only to have his personality change significantly. Were I to
be uploaded, I would want my personality intact.


> 
> Thomas: No, it is a good example. To duplicate the
> factory we would need the "contour" symbolic
> information regarding all vital components. Fragments
> of walls and doors would cover the building, other
> symbols would cover the machines and a written recepe
> would take care of the hot dogs.
> 

While a written recipe would work in this example, for the brain I
hope I convinced you above that not all important things can be
communicated by words.
For another example: I play a few musical instruments. My ability to
play them is based on motor programs in my cerebellum and possibly my
basal ganglia (depending on your choice of theories). In neither
anatomical case would your method upload my ability to play, nor would
any amount communication teach my upload to play again: I'd have to
relearn, but I don't particularly want to go through all that again.

> >Thomas:
> You misunderstood me. The upload will continue the
> mind actions and this continuation may call for a
> change. Therefore not all actions by the upload will
> be the same as before. The fact of uploading itself
> will also lead to changes in personality as does every
> experience we go thru. I tend to believe that your own
> personality has been affected (positively) by the fact
> that you went thru university.
> 
> Kind Regards,
> Thomas

Uploading will indeed change personality somewhat, but I think the
changes would be quite massive and unpredictable with this proposed
technique, possibly even to the point of losing all old aspects of
personality (which may or may not be replaced by new aspects).

In my mind, the fundamental problem with your method remains: there
are many processes that take place below the surface of the brain,
that many consider essential to the mind, that wouldn't be accounted
for.
Additionally, because you aren't fully capturing any of the circuits
at the cortical surface (you'd need to record from every neuron), the
resulting computer simulation probably wouldn't be able to maintain
the same sort of patterns the original brain was capable of. There are
highly specific connections among neurons in one region and also
between various regions that are requisite for the functioning.

Let's say you record from a neuron X. It's very unlikely that any of
the neurons to which X projects would also be recorded from. I'm not
sure what your simulation would actually do with the data you record,
but we assume it will duplicate the processing that occurs within that
neuron. Then where in the simulation does the data go? Even if you did
happen to record a neuron that receives output from X, how would you
know they're connected? (I'll answer that one: you could do a
correlation analysis on the firing of cell X and postsynaptic
potentials in the connected cell.. that *might* be able to tell you).


-- 
Eric Zilli
Hasselmo Lab - Computational Neurophysiology
Center for Memory and Brain
Boston University
2 Cummington St.
Boston MA, 02215
digfarenough at gmail.com -- www.digfarenough.com