The Bioengineering Lab - Neuromodeling Section is a collaboration between the
Bioengineering Group of the University of Trieste and Interuniversity Consortium
CINECA (Bologna, Italy) to conduct research and development in advanced modelling,
simulation, and visualization methods for solving bioelectric field problems.
Modern medical imaging technologies such as Magnetic Resonance Imaging (MRI),
Computed Tomography (CT), and Positron Emission Tomography (PET), provide a
wealth of anatomical information to doctors and researchers. Measurements of
the electric and magnetic fields from the body, such as electroencephalography
(EEG) and magnetoencephalogrphy (MEG), reflect the underlying bioelectrical
activity of the tissues and organs. However, without equally advanced modelling
and visualization technologies, much of the potential value of this information
is lost. Our goal is to couple advanced medical imaging technology with state
of the art computer simulation and modelling techniques to produce new methods
and tools, which will allow doctors and researchers to tackle immediately important
medical problems.
To accomplish this goal, we have created an integrated software tool for bioelectric
field problems called "BrainOne".
Bioelectricity occurs in all living tissue and has been the subject of investigation
since Swammerdam, in 1658, and later Luigi Galvani, in 1786, stimulated muscle
contractions by mechanical and electrical means, respectively. The origins of
bioelectricity lie within cell membranes, which maintain a small potential difference
between the interior and exterior of each cell. Fluctuation of this potential
acts as a signalling mechanism that permits nerves to interact, muscles to contract,
and communication to occur over the whole body. The rapid regulation of body
functions from vision to walking occurs by means of bioelectricity.
Of special importance in research and clinical practice are the bioelectric
fields produced by the heart, brain, and nervous system. Detection and analysis
of bioelectric fields lie at the core of research and clinical medicine in the
areas of electroencephalography, electrocardiography, and basic neural and cardiac
electrophysiology.
A large number of diseases can be diagnosed and sometimes treated on the basis
of their electrical activity. The measurement of electrical brain activity provides
important insights into the functioning of the brain by revealing the location
and sequence of neural activities, and thereby pinpointing the origins of certain
neurological disorders, such as epilepsy, sleep disturbances, psychiatric illness
and brain tumors.
Despite this high level of interest and research activity, many aspects of the
body's bioelectric fields still elude understanding. Although driven by very
basic laws of physics and chemistry, the complexity of in vivo bioelectric systems
continue to defy easy answers. Hence there is a steady demand to apply larger
and larger resources to the problem of bioelectric fields.
The activity of the Bioengineering Lab - Neuromodeling Section is built upon
the diverse research background of its components, which includes many aspects
of bioengineering (neural systems, electroencephalography, bioelectricity) and
scientific computing (modelling, numerical analysis, large-scale computing,
and scientific visualization). The group's goal was to combine these strengths
to build software tools for solving computational problems in science, engineering,
and medicine.
From the experience gained in computational medicine, it was noted that the
process of running and visualizing large-scale simulations usually required
hours or even days of a researcher's time. To remedy this, instead of using
the predominant batch-mode, off-line, non-interactive computational science
pipeline, they decided to create a system in which all the computational components
are linked -- to "close the loop," so that all aspects of the modelling,
simulation, and visualization process could be controlled graphically within
the context of a single application. This effort resulted in the computational
science problem solving environment called BrainOne.
Once a prototype problem solving environment was developed, the group pursued
and joined the collaboration with the Supercomputing Centre of the Interuniversity
Consortium CINECA.
BrainOne is designed to solve bioelectric field inverse problems allowing researchers
to localize electrical sources from the electric potentials detected outside
the body.
The Bioengineering Lab - Neuromodeling Section consists of researchers and
developers in bioengineering, computer science, and neuroelectrophysiology from
the University of Trieste. We have organized the Lab to mirror our Technological
Research and Development goals, which consist of four research thrust areas
to develop methods/techniques and implement them in software, for use within
the integrated software problem solving environment BrainOne. The problem solving
environment consists of core technology elements that link specific modules
dedicated to geometric modeling, simulations, and visualization.
With the widespread availability of computers for research has come a large
demand for software that is specialized for bioelectric phenomena. Compared
to other areas of engineering, the pace of development of software for bioelectric
field problems has been modest, especially in four aspects - flexibility, complexity,
interactivity and integration. Most programs developed for other areas of application
are not flexible enough to accommodate bioelectric field problems.
The level of inherent complexity in biological systems also exceeds that of
many other engineering applications. For instance, analysis of experimental
results from the ionic currents from cardiac cell membranes suggest that there
are perhaps tens of different types of channels all carrying potassium in and
out of the cell, each with different kinetic behavior. The heart is an aggregate
made up of billions of cells with such membranes including some degree of differentiation
in cell structure and function, however, it is a relatively simple organ compared
to the brain with its millions of dynamically changing connections. Software
to deal with problems of this complexity will always lag behind a complete description
of reality, but coming as close as possible requires specialized algorithms
and code.
Thus far, most software for bioelectric field modeling, simulation, and visualization
has been oriented towards independent, sequential processing. Each step in a
sequence of necessary computations is done separately, often by a independent
program with little or no interaction allowed between the different programs
or between the software and the user. In contrast, recent developments in software
development are aimed at providing more integration and interactivity within
the software system, allowing communication between elements of the system and
a high degree of user control over the function of the program. It is the application
of such modern software developments to the computational needs that arise in
bioelectric field problems that is our goal.
BrainOne is an integrated software tool for solving bioelectric field problems. It brings together and allows interaction between the modeling, computation, and visualization phases of a bioelectric field simulation. Though it is robust enough to support expert-users and novices alike, it is easy to move around in. This software environment empowers researchers of bioelectric fields to analyze their data, their methods, and the full range of their problem space in ways they had never before considered. When such tools integrate seamlessly, they fade into the background, the user is freed to concentrate on the problem at hand rather than on the software tools.
Modeling is the geometric counterpart to simulation in that the goal is not to describe function, but to quantitatively capture anatomy and physical locations of objects in space. From the locations of points in space, modeling seeks to define connections between these points in order to define areas, surfaces or volumes. Models in biomedical applications define anatomy of tissues and organs in the body by means of discrete points joined to form polygonal elements. There is a natural synergy between modeling and simulations in that many simulations require a geometric description of the tissue whose function is to be simulated. Examples of models that arise in bioelectric field simulations include models of the head and brain for localizing neural sources and models of the thorax and heart for simulating cardiac defibrillation.
 |
 |
|
| Plate1: 3D surfaces of model compartments. Visible compartments: scalp and muscles (pink), bone (light yellow), cerebro-spinal fluid (light green-blue), grey matter (dark grey), white matter (light grey), ventricles (deep blue), internal air pockets (light blue). Segmentation: 3D Slicer 1.3 (MIT). Visualization: VTK 4.1 (Kitware Inc.). |
||
Medical simulation is the quantitative description of biophysical behavior in terms of mathematical equations. The reasons for performing simulations include the desire to replicate the function of living organisms, both as a test of our understanding, and as a tool to investigate conditions that are difficult or even impossible to create experimentally. Examples of simulation include external fields from discrete and distributed neural sources, and the relationship between brain bioelectric sources and body surface potentials.
 |
 |
|
| Plate 2: Topographical modifications visible in the EEG produced by identical neural sources (located in the primary visual cortex) simulated with two different head models. The simulated potentials are visualized in rainbow pseudo-color map (with the same scalar range). | ||
Another category of computer applications is scientific visualization. Visualization is an essential component of virtually every bioelectric field problem and provides a means for viewing geometric models, experimental results, simulation results, and clinical observations. For example, visualizing a three-dimensional head model along with the MRI scans from the patient and the results from a source localization simulation requires the integration of many different types of visualization techniques - visualization of the geometrical mesh, visualization of the MRI data, visualization of the potentials and currents from the simulation using surface shading - all integrated into a single frame.
 |
 |