[Neuroscience] Re: Wherefore art thou Neuron Code?

Glen M. Sizemore via neur-sci%40net.bio.net (by gmsizemore2 from yahoo.com)
Fri Apr 13 05:48:31 EST 2007


"Benjamin" <Benjamin from verizon.net> wrote in message 
news:T9uTh.14637$OU1.10430 from trndny04...
> "Benjamin" <Benjamin from verizon.net> wrote in message
> news:zCGSh.6251$FC5.2247 from trndny06...
> | "Glen M. Sizemore" <gmsizemore2 from yahoo.com> wrote in message
> | news:4618d1a3$0$22574$ed362ca5 from nr2.newsreader.com...
> ||| [...]
> || [...]
> | [...]
>
> || Q1. Which keys does the animal peck?
> |
> | [Short Answer]
> |
> | 'assuming' that the pigeon is complete-
> | ly-naive, it will have no key-preference.
> |
> | As the pigeon interacts with the appar-
> | atus, the built-in contingencies will
> | automatically train the bird to con-
> | verge-upon preferring the left key -- 
> | contingency 1, above -- because:
> | [...]
>
> Dr. Sizemore wrote otherwise in his
> original posting of his 'test', but, as I
> worked, I 'lost-track' of the fact that
> he said that the "green" key can oc-
> cur on the right =or= the left in the
> '2nd' main contingency.
>
> So the pigeon will tend toward peck-
> ing 'the' "green" key, regardless of
> whether it's left or right.

This is probably what most people would say and you should have, thus, 
suspected that it was not correct. Indeed, animals (including humans) tend 
to be extremely sensitive to stimuli correlated with short delays to 
reinforcement, even when they wind up obtaining far less food per unit 
time.Thus, in the simplified problem I gave you, the pigeon would always 
peck the key leading to the short-delay small reinforcer. When the amount on 
the alternative is very, very large, the animal will choose the long-delay 
large reinforcer. The behavior is, of course, sensitive to manipulation of 
most of the parameters of the procedure.

>
> I haven't been able to see that "t"
> makes any difference -- unless
> it's correlated with the pre-feeding
> "wait" 'times' -- and, when t is relatively-
> long, the pigeon pecks "red" 'because'
> its TD E/I increases in proportion to
> the pre-feeding "wait" -- so the pigeon
> will(?) peck "red" 'because' its primed-
> to-act as a result of the "wait"-induced
> TD E/I(up) that's occurring 'within' its
> nervous system -- even though peck-
> ing "red" =always= allows the pigeon
> to eat less than does pecking "green".

No matter what t is, if the pigeon is confronted by the choice between red 
and green it pecks red (unless the amount of food obtained by pecking green 
is enormous). This is true when the pigeon chooses the "initial link" that 
led to the red-green choice OR when it chooses the link that precludes 
exposure to red - but you present the choice anyway (as a "probe trial"). 
So, in the full problem that I gave you, the answer is that at small t 
values, the animal pecks the right key in the "initial links" and then 
chooses the short delay, small reinforcer. As t grows, however, it begins to 
choose the left key in the initial link, thus precluding itself from 
choosing red because red is not an option. The results are generally 
interpreted in terms of so-called "hyperbolic discounting."

My motivation should be clear. You harangue this group with your 
"explanations" of complex human behavior based on your, ummm, obscure 
notions. But can your notions really be used to account for behavior? The 
answer appears to be "no."


>
> I experienced an analogous thing
> 'yesterday'.
>
> Went out to look for a car.
>
> Much walking, which left me weary.
>
> Saw some cars I would've liked to have,
> but, since I didn't want to take out a
> loan, I ended up paying ~4 times
> what it was worth for an 11-'year'-old
> car with a banged-in passenger door
> and rusting fender  relatively-low milage.
>
> In my insurance costs, I'll be paying
> what the cars worth every 'year'.
>
> It's not entirely-analogous, but I real-
> ized that I was 'pecking-thered-key :-]
>
> "Oh well."
>
> Taking out a loan was not a viable option
> for me.
>
> k. p. collins
>
> 




More information about the Neur-sci mailing list