PubMed Health⌕ Search

Biomedical subjects

H G Schuster

Publications and source records attributed to H G Schuster.

6 recordsLinked to original sources
PubMed2000

Beyond Hebb: exclusive-OR and biological learning.

A learning algorithm for multilayer neural networks based on biologically plausible mechanisms is studied. Motivated by findings in experimental neurobiology, we consider synaptic averaging in the induction of plasticity changes, which happen on a slower time scale than firing dynamics. This mechanism is shown to enable learning of the exclusive-OR (XOR) problem without the aid of error backpropagation, as well as to increase robustness of learning in the presence of noise.

Brain↗
PubMed1999

Self-organized criticality in a nutshell.

In order to gain insight into the nature of self-organized criticality (SOC), we present a minimal model exhibiting this phenomenon. In this analytically solvable model, the state of the system is fully described by a single-integer variable. The system organizes in its critical state without external tuning. We derive analytically the probability distribution of durations of disturbances propagating through the system. As required by SOC, this distribution is scale invariant and follows a power law over several orders of magnitude. Our solution also reproduces the exponential tail of the distribution due to finite size effects. Moreover, we show that large avalanches are suppressed when stabilizing the system in its critical state. Interestingly, avalanches are affected in a similar way when driving the system away from the critical state. With this model, we have reduced SOC dynamics to a leveling process as described by Ehrenfest's famous flea model.

Journal Article↗
PubMed1990

A model for neuronal oscillations in the visual cortex. 1. Mean-field theory and derivation of the phase equations.

We study a neural network consisting of model neurons whose efferent synapses are either excitatory or inhibitory. They are densely interconnected on a local scale, but only sparsely on a larger scale. The local clusters are described by the mean activities of excitatory and inhibitory neurons. The equations for these activities define a neuronal oscillator, which can be switched between an active and a passive state by an external input. Investigating the coupling of two of these oscillators we found their coupling behaviour to be activity-dependent. They are tightly coupled and almost synchronized if both oscillators are active, but weakly coupled if one or both oscillators are passive. This activity-dependent coupling is independent of the underlying connectivities, which are fixed. Finally, for coupled active oscillators we derive a simplified description by disregarding the amplitudes of the oscillators and working with their phases. We use this simplified description in a compagnion article to model the oscillations in the visual cortex.

Animals↗
PubMed1990

A model for neuronal oscillations in the visual cortex. 2. Phase description of the feature dependent synchronization.

In a previous paper we have shown, that it is possible to model the oscillations observed in an orientation specific column in the visual cortex by coupling excitatory and inhibitory subpopulations of neurons which compose the column, and that these oscillations can be described by the phases of the corresponding limit cycle oscillators. By coupling different columns via long but finite range sparse interactions, we generate in the phase description stimulus dependent multiplicative couplings which explain experimentally observed synchronization effects.

Animals↗