A5016986345- Neural dynamics and brain function
- Neural Networks and Applications
- stochastic dynamics and bifurcation
- Advanced Memory and Neural Computing
- Marine Invertebrate Physiology and Ecology
- Neural Networks and Reservoir Computing
- Nonlinear Dynamics and Pattern Formation
- Zebrafish Biomedical Research Applications
- Model Reduction and Neural Networks
- Advanced Computing and Algorithms
- EEG and Brain-Computer Interfaces
- Neuroscience and Neuropharmacology Research
- Ecosystem dynamics and resilience
- Receptor Mechanisms and Signaling
- Statistical Mechanics and Entropy
- Supramolecular Self-Assembly in Materials
- Cell Image Analysis Techniques
- Diffusion and Search Dynamics
- Advanced Thermodynamics and Statistical Mechanics
University of Bonn
2016-2022
Jülich Aachen Research Alliance
2017-2018
Forschungszentrum Jülich
2017-2018
Bernstein Center for Computational Neuroscience Freiburg
2008-2009
Autonomous randomly coupled neural networks display a transition to chaos at critical coupling strength. We here investigate the effect of time-varying input on onset and resulting consequences for information processing. Dynamic mean-field theory yields statistics activity, maximum Lyapunov exponent, memory capacity network. find an exact condition that determines from stable chaotic dynamics sequential in closed form. The suppresses by dynamic mechanism, shifting significantly larger...
Significance Associative memories are thought to be represented by neuronal assemblies, ensembles of nerve cells with strong synaptic interconnectivity. Experiments have, however, shown that synapses can change spontaneously. Motivated this and experimentally observed changes representations, we propose assemblies drift freely in the brain, due noisy network activity spontaneous changes, basis associative memory. How behaviors persist despite these changes? We find simple, teacher-free...
Synchronization in feed-forward subnetworks of the brain has been proposed to explain precisely timed spike patterns observed experiments. While attractor dynamics these networks is now well understood, underlying single neuron mechanisms remain unexplained. Previous attempts have captured effects highly fluctuating membrane potential by relating intensity f(U) instantaneous voltage U generated input. This article shows that f high during rise and low decay U(t), demonstrating -dependence f,...
Experiments in various neural systems found avalanches: bursts of activity with characteristics typical for critical dynamics. A possible explanation their occurrence is an underlying network that self-organizes into a state. We propose simple spiking model developing networks, showing how these may "grow into" criticality. Avalanches generated by our correspond to clusters widely applied Hawkes processes. analytically derive the cluster size and duration distributions find they agree those...
The ability of humans and animals to quickly adapt novel tasks is difficult reconcile with the standard paradigm learning by slow synaptic weight modification. Here we show that fixed-weight neural networks can learn generate required dynamics imitation. After appropriate pretraining, dynamically new thereafter continue achieve them without further teacher feedback. We explain this illustrate it a variety target dynamics, ranging from oscillatory trajectories driven chaotic dynamical systems.
Jellyfish nerve nets provide insight into the origins of nervous systems, as both their taxonomic position and evolutionary age imply that jellyfish resemble some earliest neuron-bearing, actively-swimming animals. Here, we develop first neuronal network model for jellyfish. Specifically, focus on moon jelly Aurelia aurita control its energy-efficient swimming motion. The proposed single neuron disentangles contributions different currents to a spike. identifies factors ensuring...
Neural networks of the brain form one most complex systems we know. Many qualitative features emerging collective phenomena, such as correlated activity, stability, response to inputs, chaotic and regular behavior, can, however, be understood in simple models that are accessible a treatment statistical mechanics, or, more precisely, classical field theory. This tutorial presents fundamentals behind contemporary developments theory neural rate units based on methods from mechanics with large...
Neural circuits exhibit complex activity patterns, both spontaneously and evoked by external stimuli. Information encoding learning in neural depend on how well time-varying stimuli can control spontaneous network activity. We show that firing-rate networks the balanced state, of recurrent dynamics, i.e., suppression internally-generated chaotic variability, strongly depends correlations input. A distinctive feature is that, because common input dynamically canceled feedback, it far more...
Change is ubiquitous in living beings. In particular, the connectome and neural representations can change. Nevertheless behaviors memories often persist over long times. a standard model, are represented by assemblies of strongly interconnected neurons. For faithful storage these assumed to consist same neurons time. Here we propose contrasting memory model with complete temporal remodeling assemblies, based on experimentally observed changes connections representations. The drift freely as...
Networks in the brain consist of different types neurons. Here we investigate influence neuron diversity on dynamics, phase space structure and computational capabilities spiking neural networks. We find that already a single type can qualitatively change network dynamics mixed networks may combine ones with type. study inhibitory concave leaky (LIF) convex "anti-leaky" (XIF) integrate-and-fire neurons generalize irregularly non-chaotic LIF Endowed simple conductance-based synapses for XIF...
State-of-the-art neural network training methods depend on the gradient of function. Therefore, they cannot be applied to networks whose activation functions do not have useful derivatives, such as binary and discrete-time spiking networks. To overcome this problem, function's derivative is commonly substituted with a surrogate derivative, giving rise learning (SGL). This method works well in practice but lacks theoretical foundation. The tangent kernel (NTK) has proven successful analysis...
Synchronization of spiking activity in neuronal networks the cortex has been proposed as a mechanism underlying higher brain functions. This idea is challenged by ongoing cortical generating large fluctuations synaptic input, causing neurons to operate noisy environment. The propagation synchronized feed-forward subnetworks (synfire chains) studied demonstrate feasibility precise spike timing [1]. However, theoretical analysis even this toy model impeded intricacy calculating distribution...
Neural circuits exhibit complex activity patterns, both spontaneously and evoked by external stimuli. Information encoding learning in neural depend on how well time-varying stimuli can control spontaneous network activity. We show that firing-rate networks the balanced state, of recurrent dynamics, i.e., suppression internally-generated chaotic variability, strongly depends correlations input. A unique feature is that, because common input dynamically canceled feedback, it far easier to...
In a synfire chain [1], synchronous activity in one group of neurons can excite the next to fire synchronously themselves. If this mechanism repeats itself from group, pulse packet spiking travel down chain. For homogeneous chain, spike profile is uniquely mapped successive thereby establishing map for dynamics space pulse-shaped functions. A stable corresponds fixed point infinite-dimensional map.
Perception seems so simple.I look out of the window to see houses, trees, people walking past, sky above, grass below.I hear birds in cars going distant sound an alarm.The world is full objects that make their presence known me through my senses -what could be more simple?Yet efficacy perceptual experience hides a host questions for which we do not yet have answers.Information reaching our generally incomplete, ambiguous, distributed space and time neatly sorted according its source, key...
Abstract Jellyfish nerve nets provide insight into the origins of nervous systems, as both their taxonomic position and evolutionary age imply that jellyfish resemble some earliest neuron-bearing, actively-swimming animals. Here we develop first neuronal network model for jellyfish. Specifically, focus on moon jelly Aurelia aurita control its energy-eZcient swimming motion. The proposed single neuron disentangles contributions different currents to a spike. identifies factors ensuring...