Need Help to Identify Neurons in Learning Theory

Philip Dorrell p at dorrell.demon.co.uk
Sun Jan 21 02:19:51 EST 1996


I am writing this posting to ask for those with neuroanatomical 
expertise to help me identify the groups of neurons in my two-level theory
of learning. The groups of neurons and the means by which they can be 
identified are as followed -

(Note -I use the term "neural pathway" to refer to paths in which information 
flows from one neuron to another via synapses as a result of neurons firing or 
not firing. This contrasts with information flow that occurs in a 
non-localised manner, e.g. the emission, diffusion and reception of chemical 
messengers.)

1) Group 1 are the neurons that receive and process perceptions.

2) Group 2 respond to rewards and punishments resulting from 
satisfaction/non-satisfaction of biological needs by neutralising them. The 
neurons are capable of learning to predict the occurrence of 
rewards/punishments in which case they neutralise them as soon as they occur
and prevent them from having any effect on the Group 3 neurons.
Neutralisation occurs by emission of chemicals representing a 
reward/punishment that is the opposite of that being neutralised. As a result,
non-occurrence of predicted reward/punishment will result in the occurrence of 
an attemped neutralisation of the non-occurring reward/punishment which will 
have a real effect on the learning of Group 3 neurons.
(In performing the predictions the Group 2 neurons would make use of 
information provided via neural pathways from Group 1.)

3) The learning of Group 3 neurons is controlled by externally received
rewards/punishments adjusted by neutralisation by Group 2 neurons. The Group 
3 neurons receive information via neural pathways from Group 1 and also 
information about the actions or intended actions of Group 4 neurons. The
output of Group 3 neurons is to control the production of chemical messengers
that apply reward or punishment to Group 4 neurons.

4) The learning of Group 4 neurons is controlled by the rewards/punishments 
applied by the Group 3 neurons. Group 4 neurons receive information via neural 
pathways from Group 1, and their final output is to cause actual observable
behaviour, e.g. muscular movements.

Various extra information pathways do exist. The learning of Group 1 neurons
is self-controlled. But learning of Group 1 neurons is generally beneficial,
so it results in the production of raw external rewards as adjusted by Group 2
and applied to Group 3. Group 4 is also able to produce internal behaviours 
whose sole purpose is to be perceived by Group 1 neurons, thus allowing 
internal processing of information.

Extra clue: In the "blindsight" phenomenon, part of the neural pathway from 
Group 1 to Group 3 is lost, so that a certain input is no longer available to 
Group 3, but the equivalent pathway from Group 1 to Group 4 is not cut off,
so that the input continues to be available to Group 4.

Note that my theory is essentially a two-tier learning theory. External 
rewards/punishments adjusted for expectation control the learning of Group 3
neurons, and Group 3 neurons produce rewards/punishments which control the 
learning of Group 4 neurons which produce actual behaviour. The learning that 
occurs in Group 3 occurs very slowly (it may take days, weeks or months for 
even extreme rewards or punishments to have an observable effect), whereas 
learning in Group 4 can be quite rapid (observable learning can occur in 
seconds or minutes).

Philip Dorrell   Email: p at dorrell.demon.co.uk
Articles on consciousness, dreams, music, evolution, anthropic principle at
http://www.xmission.com/~gastown/dorrell/



More information about the Neur-sci mailing list