WORKING NOTES
We do not intend to enter here into such a complex philosophical discussion; it
is however evident that, when one focuses his/her attention on the signal
fluxes and elaboration within the brain, the problem of avoiding an implicit
use of the homunculus hypothesis represents a serious task.
An enormous amount of signals are transmitted between different and distributed
parts of the brain even when relatively simple tasks are performed [4 ]. A
short list can be found here. One is
then lead to the conclusion that a great quantity of information processing and
transmission is carried out also for relatively simple inputs. Here the term
"processing" does not imply in any sense that instructions (in the
computational sense) are performed to send and interpret all the signals that
we can detect in the central nervous system.
In a number of papers [5,6,7] Tononi et al. have introduced a new
interpretation of the basic neuronal mechanisms of information processing
within the brain that differs substantially from the instructivist hypothesis
of the most radical computational theories of brain. Posner et al, in a series of pioneering works [8,9] has advanced a
theoretical model of the brain, where the action of independent, interacting
modules exchanging information was somehow suggested as a basic scenario for
brain activity interpretation.
Of particular interest, from this point of view, are the recent results by
LeDoux [10] on the information fluxes between the environment, emotional brain,
and neocortex. Even in its more schematic form, the neuronal circuit put into
evidence by LeDoux represents an interesting model to focus on, as a possible
prototype of non-instructional processing of information, based on the
interaction of different but strictly co-ordinated modules operating in the
brain. In the following we discuss such a circuit, having in mind however the more
general problem of brain information processing as possibly implemented by a
system of nested, interconnected modules implementing a non-instructional
information process, as physically identified in references [5,6,7]. If our hypothesis
is correct, LeDoux circuit can be identified as a particularly evident
prototype of the more general structure suggested there.
As far as the resulting emotional-cortical scenario is concerned, the basics
are captured by the synergetic interaction between a first-level emotional
alarm system, operating on very rough data set about the external environment,
and a more sophisticated cortical system, which operates on the basis of a more
detailed knowledge about the external world.
What seems to be the most interesting consequence of our result is the role of
the emotional module to shape the evolution of the conceptual system in its
cortical counterpart. Taking into account the most obvious needs of a
realistic, evolving, surviving system in a given environment, we show that some
important conditions must hold. Then it is shown that those conditions can be
obtained in terms of an information theory based principle. The results are
capable of somehow mirroring the results obtained [5,6,7], enlightening the
"rational" of the brain structure, as observed in the above mentioned
references, as well as the psychological analysis proposed by LeDoux.
[2]
Searle, John. 1992. The Rediscovery of the Mind, Cambridge, MA: MIT Press.
[4]
Edelman M. E., Tononi G., (2000) "A universe of consciousness", Basic
Books, New York
[10]
LeDoux J. (1996) "The Emotional Brain", Touchstone Book, New York
3-Considerations about the relation between risk and probability
4-Basics of the proposed Model
5-Gell-Mann-Lloyd definition of effective and total information
7-Discussion of the Theory in terms of Gell-Mann, Lloyd concept of Total Information
8-Evolving Agents Structures and Information Fluxes
9-Robustness to errors of the two-modules structure
What is a state of affairs in our model