Neural Networks

As a student in physics and biophysics I worked on a biologically oriented neural network model. Together with J.Zimányi and P. Vecsernyés we demonstrated optical illusion effects in a two-layer model of excitatory and inhibitory neurons giving an evidence this way for their non - psychological origin. This work won a first prize on the National Student Competition, OTDK, in 1978.

The model is based on two complementary sets of equations: the neuron equation describing the fast reaction (chargation and signal generation) and the memory equation describing long term synaptic changes. The excitatory neurons receive their inputs from a smaller enviroment than the inhibitory ones -- a condition for spontaneous pattern formation in the interconnected layer.

The illusion we investigated is based on two light points approaching and aparting each other and the reaction of the layer with a given hysteresis: The excitement patterns are longer separated from each other if started with distinguishably separated stimuli.

Lipid Membranes

My diploma thesis [1] was worked out with the guidance of I. Sugár at the Institute for Biophysics of the Semmelweiss Medical University, Budapest in 1980. I developed a mean field model in order to describe phase transitions at around body temperature (37 - 42 C) in two molecular layer lipid membranes - which were being developed those times as artifical cell membranes for packing some drugs.

The model describes the lipid molecules with a compact hydrofil head and a multi-carbon-chain tail as a high spin system based on the consecutive cis- and trans- configurations of the tail elements. Calculating with the Van der Waals forces between the neighbouring lipid tails a phase transition could be described between a more elongated (densely packed) and a more curled, losely packed membrane structure. The driving force of the phase transition is the change of the balance between the intra- and intertail free energies.

The fact that two-layer lipid membranes can suddenly change their surface structure at about body temperature supports the idea that in the early evolution of the life on Earth cell membranes were able for active transport: They kept ions inside the cell even against concentration gradients, without the help of highly developed, complex protein structures which function as gate openers in living cells.

Brownian Motion in the Ear

In a collaboration with Wolfgang A. Svrcek-Seiler, Ilse C. Gegenhuber, Frank Rattay, and Harald Markum at the University of Technology in Vienna we have studied a micromechanical model of the hair cell stereocilia in the inner ear of amphibia [ 56 ]. Brownian motion of the hairs of such cells has been measured in experiments to be in the range of some nm.

Our model consisting of coupled harmonic oscillators with mechanical properties of the stereocilia shows similar displacements during a Brownian motion generated by the thermal noise. According to the computer simulations stohastic fluctuations enhance the encoding of low level acoustic signals. Stohastic resonance lowers the detection threshold for auditory signals to amplitudes one order of magnitude lower than that of the Brownian motion.