Graphics and Supercomputing
Research by the
centre using dynamical systems to model the brain requires
a lot of computational resources. Experiments are constructed to explore
the parameter space of the dynamical systems that have been modelled.
Sophisticated presentation techniques are then used to show the results
of these experiments in both static and dynamic forms.
The experiments
and graphics techniques are performed with the assistance and resources
of the Swinburne Centre
for Astrophysics and Supercomputing which is hosted by the
Faculty of Information
and Communication Technologies at Swinburne.
Below we present
some of the images and movies that are the result of this interesting
work. For further information please contact either Dr
David Liley or Mathew Dafilis.
There is also a list of publications
that present some of this work.
Movie: Simulation of continuous neural field equations
 [pnt_gr.mov
~33.4MB, 640x480, plays for 2:13, 2000 frames]
Here we can see a
one dimensional array of neurons initialised to random conditions that
evolve over time to exhibit organised (or "wave" like) behaviour.
Height (vertical axes) represents activity while colour represents the
gradient or difference in activity between neighbouring neurons.
Movie: Variation of a Theme
 [rpnt_gr.mov
~27.5MB, 640x480, plays for 2:13, 2344 frames]
This is a radial presentation of the same neural field equation simulation
data. Colour still represents the gradient between points, however activity
is represented as a radial distance from the centre of the screen (rather
than just vertical). Provides an interesting perspective of activity.
Movie: Simulated grid of neuron activity
 [dtl_3s.mov
~11.2MB, 300x225, plays for 1:36]
80 by 80 discrete array of neurons where the height and colour of the
array represents neuron activity (firing rate). It is interesting to
note that the system starts from a random initialisation of conditions
and over time begins to display strong self-organised and propogating
activity.
Image: Chaotic attractor system
[attractor.gif, ~50KB, single image, jpeg
format]
This image is a 2D rendering of a three dimensional representation of
a chaotic attractor system that has been found during exploration of
the models parameter space. Many calculations had to be performed to
determine each point of this system, and the collection of points is
shown here as a series (or path). Colour represents the amount of transition
that occurs between each point in the series (similar to "speed"
if you were travelling along the path).
Movie: Changing attractor forms by parameter variation
 [attr_seq.mov,
~31MB, 500x500, plays for 0:33]
This movie shows us the effect that changing one of the system parameters
has on the form of the chaotic attractor that results. As you can see
from the movie, this provides a very interesting representation of the
effect changing a parameter has on such a complex system. A feature
of interest are the representations of "limit cycles" where
we can see the image forms a solid continuous band.
Image: Complexity
[hlc28.tri_1.jpg, ~129KB, 1000x1000 pixels]
This image
illustrates the complicated nature of the behaviour of a mathematical
model of brain electrical activity developed by the researchers, and
is an example of a "fat fractal". The colours give an indication
of how this complexity varies when input to the brain changes.
|
