| The brain of mammals is composed of up to 10 |
10 |
brain cells, or neurons. |
Each neuron is a complicated device in itself. The goal of this research
unit is to investigate the function of individual brain cells. For that
purpose, we are using a combination of experimental and theoretical approaches.
The experimental approaches are whole-cell patch clamp recordings in slices
of the mouse frontal cortex. The whole-cell patch-clamp technique is a
method of recording the electrical activity of neurons by pressing a relatively
blunt glass-electrode on the neuron's surface. The piece of membrane underneath
the electrode is then sucked away and the neuron's membrane potential can
be recorded. This method allows stable low-noise recordings. A slice is
a thin piece of brain tissue that can be kept alive in a dish ("in
vitro") for a few hours. Recording in slices, removed from the intact
brain, allows us to look at the properties of single neurons in relative
isolation.
Using these methods, we are studying several aspects of the signal processing
of neurons, such as their precision and regularity with which they fire
action potentials and their response to neuromodulatory substances, such
as acetylcholine and dopamine.
The theoretical approaches are biophysical simulations of neurons and the
use of genetic algorithms.
One approach are the biophysical simulations of single- and multicompartmental
neuron models. In these simulations, we use a computer to solve ordinary
differential equations that represent the different ionic currents in the
different parts of a neuron. With this we aim to reproduce the experimental
results and thus confirm that we understand the components giving rise
to the electrophysiological dynamics we observed.
Another direction we are taking is the construction of multi-scale models
of the brain. The brain operates at a number of interacting spatial and
temporal scales, from sub cellular protein signaling networks (nm and ms)
to whole brains (cm, and years for long lasting memories). Constructing
models which incorporate more than one of these scales will be a major
challenge for theoretical neurobiology in the coming years and we hope
to contribute to this endeavor.
We are also very interested in the correlation between dendritic shape
and function. To investigate this relationship, we have developed a method
of finding models of neurons with denritic trees optimized for a given
type of computation. This method utilizes genetic algorithms as a search
method and L-systems as a method for generating dendritic trees.
Please visit our lab website.
|
|
