Neural Coding

Collins, Kenneth KPCollins at ameshome.com
Sat Oct 28 00:03:40 EST 2000


do you see that, unless what you say is 'embedded' in the 3-D neural
architecture, it's of no use?

to understand, think of rush hour in Grand Central Station.

if one chooses a floor tile, and counts, either by 'rate' or by 'time', the
number of feet that step upon that floor tile, the information one gets is
'useless', except in the grossest ways.

one cannot use such information, for instance, to describe the directions in
which folks are traveling through the station, and one, therefore, cannot
derive any knowledge re., say, how full any particular train is likely to
be. etc.

[which is just I/O... Sensory-motor correlations, etc., all of which are of
extreme importance, not to be 'waved-off'.]

but, if one embeds the floor-tile-stepping stuff into a 3-D rendering of
Grand Central Station, then one can do all of the above, with increasing
accuracy as more and more floor tiles are integrated.

concepts such as 'rate' and 'time' 'coding' are 'worthless' if their stuff
isn't embedded in the 3-D neural architecture.

please don't say 'of course, but it's trivial'.

it's not 'trivial' because embedding experimental results within the 3-D
neural architecture is virtually never done in single papers.

everyone leaves the 3-D synthesis 'for later'.

as a result, the 3-D synthesis simply never happens because everyone's been
'leaving it for later'.

forgive me, please.

'rate' and 'time' 'coding, sans 3-D, is nothing.

K. P. Collins


---






More information about the Neur-sci mailing list