The study of the mind and of its biological
basis is one of the greatest scientific endeavors of all times. It is the key
to the definitive understanding of the very nature of human beings. It is regarded by many as one of the "final
frontiers" of big science.
However, the sheer
complexity of the nervous system and the many methodological barriers which
exist on the way to the objective study of its structure and function require
extensive collaboration among the scientific disciplines. Collaboration and
integration leads to new knowledge, which would not be possible without this
integration. It may assume several forms: in one of them, multidisciplinarity, it
occurs when disciplines work side by side in distinct problems of aspects of a
single problem. Interdisciplinarity, on the other hand, occurs when disciplines
intermesh, integrate and collaborate among themselves (Nissani, 1995). Modern neuroscience recruits collaboration among
molecular and cellular biology, developmental biology, genetics, biochemistry,
biophysics, pharmacology, electronics, information technology, biomedical
engineering, mathematics, statistics, physics, cognitive sciences, psychology, linguistics
and many other disciplines. They converge and intermesh in what is probably the
most multidisciplinary and interdisciplinary of all sciences.
For example, the
membrane of a neuron can be studied separately by chemistry or physics as a
complex phase of organic molecules with distinctive electrical properties;
abstracting entirely the fact that it is part of a living cell and the result
of organic evolution. This is multidisciplinary research. However, when the
membrane's structure, properties and functions are studied using a approach
combining the contributions of several disciplines working together, we have
interdisciplinarity. For example, in a research project described as "the
influence of second messenger systems upon the molecular conformation of ionic
channels and its consequence on integration of input information by dendritic
fields of genetically defective neurons in the visual area, using electronic
imaging and micromachined electrodes" we note the intrinsic
collaboration of many disciplines from many different areas of knowledge. The outcome of this seemingly important
research would not be possible without their integration.
The Roots of
Interdisciplinarity
The fascination of many philosophers
and natural scientists with the workings of the brain and of the mind has made
neuroscience a strong multi- and interdisciplinary field of research and
speculation centuries before the very concept of interdisciplinarity was
invented.
The baffling and unfathomable nature
of nervous activity has bred for hundreds of years all sorts of
"explanatory models", which, in turn, were highly influenced by the
cultural and scientific paradigms of the day (Kuhn, 1970). Since the detailed
structural study of the human body could be effectively carried out many
centuries before any meaningful physiological studies were possible, and due to
the fact that Renaissance science was fascinated with everything mechanical, it
is not surprising that the first models of function of the nervous system were
mechanical ones (Sabbatini, 1998). The real philosophical breakthrough was to
regard the body as a kind of machine, albeit suffused with "vital
fluids" which seemed to differentiate it from the mere artifacts created
by man. This stimulated a lot of interdisciplinary thought and collaboration.
For example, René Descartes proposed
around 1660 that nerves conducted the stimulus to the brain thanks to a wave of
propulsion of the fluid in its hollow interior. Pumped full of this fluid, the
muscles would expand, leading to contraction. Descartes delighted in comparing
what he called the "nervous machine" with the elaborate water
fountains which were in vogue at the time. This hydraulic model, using internal
fluids, was universally accepted for more than a century by anatomists such as
Thomas Willis and Albrecht von Haller, as well as by physicists such as Sir
Isaac Newton (see Brazier, 1959, for a compact and highly readable description
of the historical development of neurophysiology).
In the last decade of the 18th
century, however electricity substituted mechanics as the main interest of
physical science, and also, as the most likely explanation for the actions of
the nervous system. The new paradigm in
neurophysiology started by simple observations by naturalists of electric
fishes, which proved that animals could generate electricity in their bodies.
The Italian anatomist and physician Luigi Galvani (1737-1798) was one of the
first to investigate experimentally the phenomenon of what came to be named
"bioelectrogenesis". In a series of experiments started around
1780, Galvani, working in the University of Bologna, found that the electric
current delivered by a Leyden jar or a rotating static electricity generator
would cause the contraction of the muscles in the leg of a frog and many other
animals, either by applying the charge to the
muscle or to the nerve. His conclusions were published on 1791, with an essay
titled "De Viribus Electricitatis in Motu Musculari Commentarius"
(Commentary on the Effect of Electricity on Muscular Motion). A famous polemic
with physicist Alessandro Volta ensued, but eventually Galvani was proven
right, thus originating two centuries of collaboration between physicists and
biologists in the study of nervous action at the cellular level.
The paradigm shift was complete:
nerves were not water pipes or channels, as Descartes and his contemporaries
thought, but electrical conductors. Information within the nervous system was
carried by electricity generated directly by the organic tissue. This became an
enormously powerful paradigm for the experimental investigation in
neurosciences. It was the first time in the
history of science that a close cooperation between physics and biology was
necessary, not only because biology was looking upon the phenomena of neural
function under the light of the recent knowledge provided by physics, but also
because they were using the same measuring instruments, devices and apparatuses
(for a review, see the section on "Tools of the Trade" in my
1998 paper on the history of the discovery of bioelectricity).
The applied knowledge amassed by
experimental physicists in the fields of electricity, optics and mechanics was
starting to make a great contribution to physiology. Many times, however, the
physiologists who were carrying out pioneering experiments were forced to
invent or to adapt existing instruments in order to fit the demanding
requirements of working with living tissue, capable of generating extremely
feeble currents and voltages, never before studied by physicists. This was the
origin of the interdisciplinary professional in neuroscience. Electrical
stimulation and recording, and visualization of neural tissue at the cellular
level demanded the most sophisticated and advanced physical equipment of the
time, such as foil and string galvanometers, electrical piles and generators,
mechanical time recorders, optical microscopes, etc. Later on, as both the
instruments and neurophysiology evolved, one field boosted the other. High gain
electronic amplifiers, cathode ray tube oscilloscopes, precision square-wave stimulators, wave
recorders, etc., were essential for electrophysiological investigation. New
diagnostic tools for medicine such as the electroencephalographic recorder,
were also born as the children of the marriage between the physical and
biological disciplines. The same happened in the area of neurochemistry,
leading to the discovery and study of neurotransmitters and many other nervous
phenomena.
The Emergence of Modern
Neuroscience
The emergence of neuroscience
as a distinct discipline is relatively recent. It occurred in the early 1960s,
as a result of the convergence of more traditional fields, such as
neuroanatomy, neurophysiology, neurochemistry, behavioural psychology,
ethology, etc., and its birth was greatly influenced by interdisciplinary
research programmes, such as those developed by Stephen W. Kuffler and Francis
O. Schmitt, among others. The concept of integrated and interdisciplinary
neurosciences rapidly evolved as the consequence of a series of symposia,
conferences, publications (such as the influential "Study Programs"
series, edited by Schmitt for the Rockefeller University Press), and, finally,
the launching of the Society for Neuroscience, in the USA, and of the
International Brain Research Organization (IBRO), both in the 1970s. Many
prominent neuroscientists, such as H.H. Jasper, Albe Fessard, W.J.H. Nauta,
Mary A. Brazier, were involved in this.
Schmitt's early efforts,
beginning in 1962, is a good example of the role of interdisciplinarity in the
birth of modern neuroscience. He started the Neurosciences Research Program
(NRP), affiliated with the Massachusetts Institute of Technology and the
American Academy of Arts and Sciences, and later with the Rockefeller
University. From the very beginning, NRP was revolutionary because it involved
an interdisciplinary, inter-university programme joining the exact, the
biomedical and the behavioural sciences to "investigate the
physicochemical and biophysical bases of mental processes, such as memory,
learning and consciousness". A disproportionately large number of past and
current scientific leaders in the neurosciences were participants in the NRP.
More recently, the strong
development of molecular and cellular biology, genetic mapping and engineering,
in great part aided by spectacular technical advances in neuroinformatics,
bioinformatics, laboratory robotics, etc. have greatly accelerated the
neuroscientific knowledge on the basic functions and development of the nervous
system, as well as the causes of many neurological and psychiatric disorders.
Due to these developments, three noted neuroscientists have recently presaged
the "emergence in the coming decades of a new nosology, certainly in
neurology and perhaps also in psychiatry, based not on symptomatology but on
the dysfunction of specific genes, molecules, neuronal organelles and
particular neural systems." (Cowan, Harter & Kandel, 2000).
New impulses to the
interdisciplinary integration of the neurosciences were given by the announcement
of the Decade of the Brain, in 1990, by the American government, and the
formation of the Human Brain Project (HBP), an internationally funded drive to
increase dramatically our knowledge about our brains and minds. The result of
all this has been nothing short of spectacular. According to the Society for
Neuroscience (SFN), " In 1970, neuroscience barely existed as a
separate discipline. Today, more than 300 training programs exist in
neuroscience alone, and neuroscience is one of the most exciting areas of
biomedical research.". A clear indicator of this growth is SFN itself,
which has now more than 29,000 members and more than 100 local chapters.
New Interdisciplines with
Neuroscience in Mind
The last decade of the last
century has witnessed an extraordinary increase in the number of bridges
between disciplines, having neuroscience as the focus. If we imagine that each
discipline of knowledge is a vertex in a polygon, the lines connecting these
vertices in double, triple and n-tuple configurations have been proliferating
at a phenomenal rate. In fact, it has been difficult to keep pace, and the
naming conventions for new interdisciplinary and transdisciplinary fields are
becoming sometimes hard to recognize.
Just to give the reader an
idea about this proliferation, see below a very limited list of
interdisciplinary neuroscience areas which have appeared recently:
The degree of
interdisciplinarity in any of these new areas may vary, of course. Nissani
(1995) proposes to characterize the degree of interdisciplinary integration
according to four criteria:
Interdisciplinarity
evolves, too. What was interdisciplinary in the past, such as the integration
between biology and physics, became a discipline (e.g., biophysics), with its
own set of experts, conferences, journals, professional societies, committees,
graduate courses and students, methods and even established professions with
undergraduate curricula (Klein, 1990).
In the following sections we
will examine the dynamics and rationale of creation of new
"interconnecting lines" between disciplines and levels of analysis,
and two case studies which we will use to exemplify the powerful results of
these interactions in two areas: technology helping brain research and
applications, and the approach between computer and neurosciences.
Interdisciplinary
Integration: The Concept of Levels of Analysis
Most of the advances which
occurred in the history of the basic and clinical neurosciences have been
limited to research within a single level of organization; i.e., either at the
molecular, cellular, organ, psychological/behavioral or social/environmental
levels. This is the traditional intradisciplinary research. For example,
biochemists have studied intraneuronal proteins at the molecular level, cell
biologists have studied neural organelles and processes at the cellular level,
neurophysiologists have studied the properties, structure and functions of
larger neural networks in sensorial and motor systems, psychologists have
studied psychological and behavioral phenomena such as learning, conditioning,
emotions, etc.; and sociologists, biologists and other professionals have
studied social organizations, the relationships of human societies to the
environment, and so on.
Interdisciplinary,
cross-disciplinary and multidisciplinary research requires that some degree of
interaction occur between these levels of analysis. Cacciopo and Berntson, in
1992, have proposed the concept of multilevel analysis as a way of breaking the
barriers between traditional disciplines and to achieve interdisciplinary
collaboration. Multilevel analysis is not synonymous with interdisciplinarity,
because interdisciplinarity may occur even when restricted to a single level.
However, it is essential for a wider collaboration between disciplines which
have specialized themselves in distinct levels of this hierarchy.
Two processes are necessary
for multilevel analysis to take place: the first one is to use discoveries,
findings and concepts taken from one level of analysis to inform, to refine,
and to constrain inferences obtained at another level of analysis. Evolutionary
psychology is a new interdisciplinary field which shows this: the knowledge
about evolutionary genetics is used to inform, clarify and derive new knowledge
about human social behavior, language, cognitive development, etc. which were
hitherto not seen under this new light. The second one is to study
simultaneously a phenomenon using two or more levels of analysis, or, in other
words, by disciplines which are specialized at these levels. In the same vein
of the previous example, a phenomenon such as the human laughter can be studied
by comparative ethologists, developmental human biologists, neurophysiologists,
clinical neurologists, social psychologists and other professionals to
understand what is laughter, why we laugh, how it appeared in the human record,
what is its function in human evolution and social organization, how it is
organised by the brain, how much of it is innate and how much is socially
learned, and so on (Cardoso, 2000). The difference here is that the research
hypotheses themselves are multilevel in nature. The resulting explanations can
coexist at several levels, simultaneously, and also generate new explanations
at any single level.
The needs for a multilevel
approach and its advantages are clearly spelled out by Andersen (1998):
"Although the success of single-discipline research is evident,
this approach may also be seen as somewhat limiting. The limitations exist
because, while the disciplines concerned with health research may be separated
conceptually, methodologically, and administratively, the processes about which
they are concerned are inextricably linked. In other words, the social,
behavioral, and biological processes that affect health are interdependent.
Failure to conduct research across disciplinary lines precludes the discovery
of these interdependent processes. […] Research that integrates the various
levels of analysis represents one of the next great frontiers in the health
sciences, with the potential to accelerate advances in both basic and clinical
research and in public health. The hallmark of such integrated, multilevel
research is interdisciplinary collaborations, which use the expertise of
several disciplines to address complex health issues. "
Multilevel analysis provides an
excellent paradigm to understand, explain and initiate new interdisciplinary
research in the neurosciences. The several types of interactions between the
levels of analysis have been systematized and are described in the review paper
by Andersen (1998). In our view, this systematic knowledge about how and where
to carry out multilevel interdisciplinary research should be a required subject
of all courses on the methodology of science for future scientists.
The reason is that there are
still wide distances between several levels of analysis which remain to be
bridged over. Noted neuroscientist Jaak Panksepp has recently written an essay
for a new journal devoted to cover the gap between neuroscience and
psychoanalysis, a giant and influential field which has been often criticised
for its lack of scientific support:
"The failure of too many in the various psychological,
psychoanalytic and neuroscience communities to concurrently embrace the full
hierarchical complexity of the human brain-mind with a full devotion to
conceptual flexibility and scientific rigor, has saddened many generations of
students who wanted a deep and realistic understanding of the human condition.
[…] Because of the recent revolutions in neuroscience and molecular biology
along with emerging psycho-behavioral and dynamic perspectives, we may finally
be in a position to deal with some of those long neglected issues credibly
using rather standard, albeit somewhat more theoretically flexible, scientific
approaches. The only thing that should matter in this scientific game is the
capacity to make predictions that can be empirically verified or falsified
[…]."
The same applies to
philosophy and other areas of the human sciences which evolved much earlier and
separately from the neurosciences, in regard to the study of the human mind and
its manifestations, under totally different paradigms. It remains to be seen
whether the gaping chasm between them and the "hard" natural sciences
eventually will be filled by novel interdisciplinary approaches.
Interdisciplinary
Collaboration: The Case of Functional Neuroimaging
A good example on how a
totally new field of research and application can be created by
interdisciplinary and multidisciplinary collaboration, and how this approach
can revolutionize a traditional science, is functional neuroimaging.
Powerful neuroimaging
techniques developed in the last decade have allowed neuroscientists to map and
to study in detail the brain regions which are associated with many motor,
sensory, cognitive and language actions and tasks in humans. Their most
important contribution so far has been the establishment of clearer
relationships and influences between the social/behavioral and organ levels of
analysis we have described above, by providing novel insights into the function
of several regions of the brain that underlie normal and abnormal behavior
(Nemeroff, 1999). The so-called "big four" functional imaging
techniques are SPECT (Single-Photon Emission Computed Tomography), PET
(Positron Emission Tomography), MRS (Magnetic Resonance Spectroscopy) and fMRI
(functional Magnetic Resonance Imaging).
The first two are based on radioligand tracing with scintillography,
using short lived radiopharmaceuticals which can be used to produce two and
three dimensional images of the brain depicting in color the increased activity
in metabolism, glucose and oxygen consumption by neural cells, regional blood
flow, distribution and density of receptors and transporters, neurotransmitter
synthesis activity, etc. The last two use the differential paramagnetic
properties of substances found in the body, such as deoxy-hemoglobin, to obtain
images related to neural activity, concentration and distribution, etc. Externally administered pharmaceuticals can
also be used to trace specific activities.
In addition to these four
imaging methods, we have also two topographical signal analysis techniques
which are also able to produce bi- and three-dimensional dynamic maps of neural
activity, which are Colour Electroencephalographic Topography and Magnetic
Electroencephalography.
The impressive fact is that
all these techniques have been developed in a relatively short period by virtue
of a wide and unprecedented interdisciplinary effort, involving electrical and
mechanical engineers, physicists, mathematicians, computer programmers,
neurologists, psychiatrists, and many other professionals. The main fuels of
this effort have been scientific, medical and economic interests. The
functional imaging industry is nowadays a major segment of the biomedical
equipment market, as it can be attested by the size of the technical exhibit
during the famous annual meeting of the Radiological Society of North America
(RSNA) in Chicago, reputedly the largest medical meeting in the world. The
field continues to evolve at a furious pace, recruiting ever more technical
talents and aggregating more and more advanced concepts and inventions, such as
virtual reality, telepresence, robotics, etc. The industry is already
developing, manufacturing and marketing powerful combined devices, such as
those which put together in a single apparatus a PET and a fMRI imaging
systems, or a PET and a CT scanner, or a CT scanner and a
magnetoencephalograph. The resulting images are cross-registered or aligned, in
order to produce, for example, a picture of metabolic neural activation using
deoxyfluoroglucose (DFG) superimposed on a MRI anatomical slice of the brain at
exactly the same level. Complex algorithms and mathematical techniques can be
used to extract novel information from these images in real time and depict
them in the form of dynamical maps in two, three and four dimensions.
Astounding examples of the
power of these combined techniques can be seen in many research projects
carried out by the International Consortium for Brain Mapping, such as a
probabilistic four dimensional map of the human brain which is in development
under the leadership of UCLA's Dr. John Maziotta (Maziotta et al., 2001).
Microscopic and macroscopic, functional and anatomic images of the brains of
ca. 7,000 individuals are being analysed. Detailed demographic, clinical and
behavioral information is also collected for each subject, and, in addition, 5,800 subjects will contribute
DNA for the purpose of determining genotype– phenotype–behavioral correlations.
This and other contributions proposed by the Human Brain Project can be seen in
the book by Pechura & Martin (1991).
Among the many
interdisciplinary fields which were richly fertilized with the tools of
functional brain imaging, cognitive neuroscience stands high.
According to one of the pioneers and leaders in this area, Marcus Raichle,
cognitive neuroscience "combines the experimental strategies of
cognitive psychology with various techniques to actually examine how
brain function supports mental activities" (Raichle, 1998).
Our current knowledge about the brain substrate for learning, memory, visual
and auditory cognition, etc. has been impacted tremendously in the last ten
years by the advent of neuroimaging. The same has happened to the clinical
neurosciences, were the understanding about the causes and consequences of
several brain and mental diseases have profited from the availability of
functional neuroimaging in patients (Honey et al., 2002).
The Emergence of
Interdisciplines: The Approach Between Computing and Neurosciences
The recognition, by
philosophers and scientists alike, that the brain is a kind of computing device
is quite old, and has spawned many new interdisciplines, of which, the most
recent ones are Computational Neurosciences and Neuroinformatics.
Charles Babbage, the 19th
century prodigious inventor of the first programmable computer, laid out the
basis for artificial machines capable of imitating some aspects of human
intelligence, such as numerical calculations, symbolic processing, etc. But it
was only in the second half of the 20th century that pioneers such
as Alan M. Turing, Norbert Wiesner and John von Neumann argued that the brain
could be seen as a kind of computer, and that making a machine capable of
duplicating human intelligence was a desirable and achievable goal. Von
Neumann, one of the most brilliant mathematicians of all times, and the
inventor of the stored-programme concept for digital computers, expressed the
view, in 1956, in his now famous Yale lecture "The Computer and the
Brain", that the brain seemed to operate in part using digital (binary)
computing, and in part using analogical computing, and that it really was quite
different from man-made computers, including the fact that it was really a statistical
computing device.
At the same time, however,
logical calculus using the binary
system was developed for electrical circuits and it was shown to be amenable to
implementation in automatic mechanical or electrical machines, thus providing
the basis for achieving "machine thinking" using deterministic logic,
an old desire (and fear) of the human race since the times of the Golem legend.
Thereafter, advances in
digital automatic stored-programme computing technology enabled in the 1950s
and 1960s the development of a new
branch of the computer sciences, called Artificial Intelligence (A.I.). It was
created by the work of pioneers such as Marvin Minsky, John McCarthy, Edward
Feigenbaum, Alan Newell, Herbert Simon and others. It tried to model human
intelligence in a machine without actually understanding how it works, an
approach which is called "heuristics". For example, a superior chess
playing program can be programmed to defeat the world champion of chess (such
as IBM's Deep Blue) using only algorithmic resources which have nothing to do
with the way the human brain works when playing chess! However, A.I. has
influenced directly and indirectly the development of the cognitive sciences,
mainly because, as it was put by Herbert A. Simon, one of its founders,
"Cognitive science is the study of intelligence and intelligent systems,
with particular reference to intelligent behavior as computation". This is
apparent in the writings of modern thinkers such as Daniel C. Demett, when
musing on the basis of the consciousness.
Although it essentially
failed its grand promises of "fifth generation computing", of
achieving the level of human intelligence by the year 2000, and of producing of
a new breed of symbolic processing machines, the top-down A.I. approach
produced a number of useful domain-restricted expert functions for computers,
which are widely utilised in many areas, such as credit rating in banks, the
interpretation of medical lab tests, the discovery of new knowledge hidden in
statistical databases, etc.; but it uses little of what neuroscience knows
about the brain, because it was unfeasible to develop hardware and software
working similarly to the nervous system at the time that A.I. began to unfold.
This bottom-up approach
started to have results only recently, and in large part due to
interdisciplinary collaboration between neuroscientists and computer scientists
(see Churchland & Sejnowski, 1999, for a comprehensive and didactic review
of the field). It all began very early with a MIT neurophysiologist, Warren
McCulloch, who in the 1950s and 1960s tried to model intelligence based on the
networked action of individual neuron, an approach which has been called
"emergent", or "connectivist computing", because computing
functions emerge from the collective action of interconnected simpler and
stereotyped elements ("neurodes"). Parallel distributed processing
(PDP) is also another name given to this approach, because it is a key
component of any kind of biological intelligence.
In 1950, Frank Rosenblatt, an
American psychologist, developed the first artificial neural network capable of
learning useful computations, which he called the "perceptron", which
connected artificial neurones in two layers (input and output). Geoffrey Hinton
and others later extended the model to multilayer perceptrons (MLPs), which
were then mathematically proven to be capable of performing complex non-linear
calculations. After some initial mishaps, this gave origin in the 1990s to a
rapidly developing and fully-bloomed interdisciplinary area called Neural
Computing, or Neuroinformatics. Many of
the artificial neural networks that were subsequently invented, such as those
by Stephen Grossberg, rely heavily on what neuroscience knows about sensory
systems, attention, learning, memory, etc. They were also shown to be capable
of amazing feats which were impossible to implement efficiently by using
conventional Artificial Intelligence, such as the recognition of human faces,
of handwritten letters and of spoken words, etc. Among other applications, we
have used artificial neural networks (ANNs) to carry out complex medical
prognoses and diagnoses, such as the accurate prediction of mortality for
cardiac patients based on clinical and echocardiographic data (Sabbatini & Ortiz,
1993) and the recommendation of brain surgery for patients with head trauma
(Botelho, Araújo & Sabbatini, 1997).
Neuroinformatics combines
neuroscience research with computer sciences, mathematics, physics, engineering
and related sciences. Its aim is to produce artificial devices which work very
much as the complex perceptual, motor and cognitive functions of human brains.
In addition, according to the NIH, it aims at developing new conceptual
approaches to basic and/or clinical neuroscientific research and analysis and
to acquire, store, retrieve, organize, manage, analyze, visualize, manipulate,
integrate, synthesize, disseminate, and share data about the brain and
behavior. Therefore, there is clearly
here a circular chain of influence: Neuroinformatics builds its artificial
devices based on knowledge about the brain, and as a result provides tools that
can enhance our understanding about the brain. Other interdisciplines have
arisen as a result, such as neurobionics, the invention of intelligent neural prostheses (for instance,
artificial retinas and cochleas, neurally controlled hands and limbs, etc.),
and neuro-robotics (robots controlled by artificial neural networks). This is
one of the most powerful results of interdisciplinary integration that are
known.
From the initial interests of
neurophysiologists in the mathematical modeling of the electrical activity of
neurons, another interdisciplinary branch soon grew out: computational
neurosciences. The fathers of this approach were British biologists Adrian
Huxley and Alan Hodgkin (both of whom were later awarded the Nobel Prize and
life peerages). In the 1950s, they described a set of differential equations
which modeled the behaviour of ion transport and the dynamic electrical and
chemical equilibrium of membrane polarization in active, transmitting neurones
("action potential"). By simulating in a computer the numerical integration of these equations,
they obtained a remarkably accurate reproduction of naturally observed
recordings in the squid giant axon and the HH model, as it is called, remained
for almost three decades the theoretical basis of the electrochemistry of
action potentials. Other pioneers in the field, such as Wilfred Rall, Michael
Hines and Art Vance later continued the interdisciplinary work which advanced
computational neurosciences at the single cell level.
It was only a matter of time
that the computer simulation of more complex neural networks was feasible by
joining excitatory an inhibitory HH neurones to each other and by observing the
resulting activity. Jim Bower, at Caltech, was one of the first to demonstrate
that neural networks with up to 40,000 elements could be simulated in
supercomputers and render remarkable accurate models of the naturally-observed
olfactory cortex activity, for example (Eeckman & Bower, 1993). He also
showed how the progress of Computational Neurosciences could be spurred by
having tools for realistic neural simulation using special software tools, like
GENESIS, developed by him and his group at Caltech, and which is widely used by
researchers in the field.
Nowadays, Computational
Neuroscience is developing at full speed and, in the USA, it is a priority area
for the National Sciences Foundation and of the National Institute of Mental
Health, which have established a special grants programme for it. Many academic
institutions have established research programmes in Computational
Neurosciences, particularly in the interface with the cognitive neurosciences
and the training and recruiting of new talents proceed at a brisk pace. There
are at least a major annual conference (Computational and Neural Sciences),
an international journal (Journal of Computational Neurosciences),
several national and international summer courses and all the accoutrements of
established science. Computational Neuroscience is also firmly entrenched in
the Human Brain Project, because the emergence of supercomputers and novel
simulation techniques have provided a hope that the neurosciences will be one
day the first "complex object" biological science to have a strong
mathematical basis, such as it happens with other "hard sciences",
such as physics and chemistry. The value of this cannot be underestimated, and,
in my opinion, it is a future development that will eventually be able to
integrate molecular and genetic aspects, too. In other words, the powerful
field of Bioinformatics, which deals with the processing and representation of
molecular genetic information on the genome, proteome, etc., will join hands
with computational neuroscience, leading to an amazing powerful tool for
understanding all of the brain.
The ability that computer
databases and algorithms will have on the modeling and description of entire
nervous systems is already evident with simple organisms, such as the soil
nematode Caenorhabditis elegans, which has all of its 302 neural cells
described and catalogued, as well as its entire genome, with 97 million bases,
therefore making it the most completely characterized metazoan. A total
computer model of this organism, incorporating all the knowledge accumulated
about it, from its basic molecular systems to its behavioral repertoire, is not
far away. WORMATLAS, a database comprising its behavioural and structural
anatomy is already available. It is entirely possible that other organisms,
more complex that C. elegans, will be described in the same way in the
future. The consequences for biological research are simply mind boggling.
It is easy to see
that the advancement of the neurosciences will rely more and more on
interdisciplinary approaches on the next years. This poses three great
challenge to the neuroscientific establishment: first, the way neuroscientists
are trained; second, the way departments and laboratories are organized in the
academic institutions, and third, how to foster interdisciplinarity in the
neurosciences.
The first of
these challenges, in my opinion, is the most serious one. It is marked by the
fact that many biologists usually do not like the exact sciences and are wary
of the methods of engineering, computing and mathematics. Their professional
profile is still too "soft", too devoid of the knowledge and
interactions which are required to be a modern and effective professional. In
consequence, it is difficult to find and to recruit interdisciplinary talents,
because they are rare, difficult and slow to train by conventional educational
methods. This must be changed, and a new kind of neurobiologist must arise: one
who is able to work comfortably and to interact simultaneously with the
specialists of many of the disciplines and levels which are being integrated. Someone in the research team must
have the overarching vision of what the project is about, and how and when the different
disciplines are going to interact. More resources must also be expended toward
finding and training the new interdisciplinary neuroscientist. Experts
agree that the best point in time to do this is during the undergraduate
formative years. Talent for interdisciplinary work already appears at college
age and must be cultivated by means of special educational programs which will
take this student out of the conventional "mill" and put him or her
in alternate tracks which emphasize team work, multidisciplinary and
interdisciplinary work and which encourage the study and specialization in two
or more disciplines. For instance, many neuroinformatics labs accept only
graduate students who have previous practical experience and good background in
neurobiology and informatics, and who have talent for advanced
mathematics.
A particular area
appears as challenging enough to require its own kind of interdisciplinary
effort and dedicated professionals: how we are going to collect, systematize
and disseminate the huge amount of scientific information about the brain and
the mind. Informatics and computer networks such as the Internet are now
essential for this endeavor (Cardoso, 1998). For example, the GeneBank, which
holds all the known genetic sequences of all species, is so large and changes
so frequently that the only way of maintaining it updated and to carry out
quick universal distribution is via the Internet. The USA and other countries
have established exemplary huge computational and networked structures to serve
this purpose, such as the National Center for Biotechnology Information. A
similar centralized effort is due for neuroscience data. The Human Brain
Project is one of the most impressive interdisciplinary programs which were
proposed to face this challenge. A strong collaboration between informatics and
neurosciences has been a vital component for its brain mapping efforts (Pechura
& Martin, 1991)
What it's in the
future? The complete sequencing of the human genome, and subsequent efforts to
decode the entire set of human genes and its expression, which are underway, is
also a major stepping stone for a new interdisciplinary neuroscience.
Everything in neural biology will be referenced to this framework, and
unfortunately it is also the case that the practical and theoretical knowledge
of neuroscientists in this area must be increased if the integration to other
levels of analysis is to happen.
Creativity in
science depends on the establishment of links between previously unrelated
ideas. The growth of neuroscience will ultimately depend on this creativity, as
more and more complex problems are tackled.
References
Internet Resources
The Authors
Renato M.E. Sabbatini is a neuroscientist
with a PhD in neurophysiology of behavior by the University of São Paulo,
Brasil, and a post-doctoral fellow in the Department of Behavioral Physiology
of the Max-Planck Institute of Psychiatry, Munich, Germany. Currently, Dr.
Sabbatini is the director of the Center for Biomedical Informatics and Chairman
of Medical Informatics of the Medical School of the State University of
Campinas, Campinas, Brazil. He is also the associate editor of "Brain
& Mind" magazine, and editor-in-chief of Intermedic, a journal on
Internet and Medicine. Email: renato@sabbatini.com
Silvia Helena Cardoso, PhD.
Psychobiologist, master and doctor in Sciences by the University of São Paulo
and post doctoral fellowship by the University of California, Los Angeles.
Invited Professor and Associate Researcher of the Center for Biomedical
Informatics, State University of Campinas (Unicamp), Brazil. She is also
editor-in-chief of "Brain & Mind" magazine, and associate
editor of Intermedic, a journal on Internet and Medicine. Email:
cardoso@nib.unicamp.br