[Comp-neuro] The Journal of Integrative Neuroscience
roman.poznanski at um.edu.my
roman.poznanski at um.edu.my
Fri Jul 10 05:21:14 CEST 2009
The Editorial appended below may be of interest to readers of this
list. It will appear in JIN, 8(3),2009.
This Journal epitomizes a trend to promote the theoretical
investigation of brain function by elucidating its functional
organization. Such theory of the functional organization of the brain
requires consideration of the spatiotemporal continuity between its
many functional levels. Theoretical foundations for explicit
hierarchical and functional integration in the brain are the essential
first steps of relevance for a synthesis to emerge.
Currently, there are clear divisions within the neuroscience theory
community-of-interest with regard to how this objective of
understanding complex integrative features of brain organization is to
Integrative Neuroscience is a first step towards the construction of a
theoretical foundation of brain science. It limits the use of
computation to the final last step in the modeling process by assuming
that symbolic representations of functional traits can be
mathematically integrated across adjacent hierarchical levels, thereby
allowing for an explicit integration of more realistically plausible
models fueled by mathematical (structural or organizational) insight
and constrained by experiment. The theoretical modeling takes into
account the fact that the brain is endogenous (Kercel) from the
synapses to the circuits governing phenotypic behavour.
Neuroinformatics applies an information science infrastructure to
neuroscience databases. While such cataloging provides useful access
to a variety of distinct facts, it necessarily compartmentalizes data
into ever-more subdivided similarity classes. Thus, while
neuroinformatics provides taxonomy for neuroscience data, integrative
analysis must be sought elsewhere (analogous to the Linnaeus-to-Darwin
transition in biology’s history).
Neural Engineering is not strictly “neural" without explicit
relationship to neuroscience concepts. Specifically, current
engineering techniques are fortuitous applications without delineating
as a bare minimum a theoretical foundation, which can surmise the
principles of integrative neuroscience. So it is important to
rationalize whether technology used today is capable of linking such
methodologies with notion of integration in the brain. Synthetic
neural modelling and brain-based devices pioneered by Edelman can be
considered as heuristic tools demonstrating a given brain theory.
Unfortunately at present such a theory is nonexistent.
Computational Neuroscience suffers from an epistemological limitation:
it assumes a discrete functionality as an objective and seeks
“integrative mechanisms” which reconstruct aspects of the given
function. Such reductionist directionality is intuitively incompatible
with the multidimensional parametric space of complex systems, which
must obtain in the highly interconnected and massively parallel
architecture of the brain. Therefore, the use of computation a priori
invokes discretization of the hierarchical structure that encompasses
a superfluous aurora of integration through multi-scale modeling,
tailored to different questions that require different levels of
structural realism, independent of each other and not integrated.
Although functionionality of the brain undermines the necessity for
lynch pins bridging across hierarchical scales, simply ignoring the
linkages between hierarchical levels in order to get a computed handle
on a single-level process represents an unrealistic approach, changing
neuroscience into a collection of irreconcilably arbitrary empirical
The Journal will aim to publish efforts in search of “integrative
mechanisms”, especially from the perspective of neuro-robotics and
artificial intelligence. In this perspective, conceptual circuit
models will need to be developed that can explain human behavior
through the use of adaptive, biological algorithms. Constructing a
neurobiological theory of the mind from the bottom up is
unquestionably daunting. However, by carefully selecting appropriate
“integrative mechanisms” connecting the various hierarchical systems
representing entire brain regions, the modeling effort becomes
conceivable and comprehensible, thus nurturing the ‘science’ in
Roman R. Poznanski
Kercel,S.W. (2004) The endogenous brain. J. Integr. Neurosci. 3, 61-84.
Edelman,G.M.(2006) Synthetic neural modeling and brain-based
devices.Biol. Theory 1, 8-9.
Professor Roman R. Poznanski
Department of Artificial Intelligence
Faculty of Computer Science and Information Technology
University of Malaya
50603 Kuala Lumpur, Malaysia
Office: (+603) 7967-6418
E-mail: roman.poznanski at um.edu.my
Journal of Integrative Neuroscience
UNIVERSITY OF MALAYA - " The Leader in Research & Innovation "
More information about the Comp-neuro