ITB logo
WS 2004/05

Computational Neuroscience: Oberseminar (31178)

Dr. Laurenz Wiskott, Prof. Andreas V. M. Herz, and Dr. Richard Kempter
HU logo


Oberseminar: Monday 11-13 Uhr in the ITB seminar room (Invalidenstraße 43)

In this seminar various topics of current research in computational neuroscience are presented. Attendees should have basic knowledge in neuroscience and computational neuroscience, e.g. from the courses "Computational Neuroscience I-IV".


Talks

06.09.2004 Jan Benda
"Report on the 7th International Congress of Neuroethology"

Abstract: I present a short summary of a few interesting talks from this years neuroetholgy conference:

James Poulet found corollary discharge in the auditory system of crickets that inhibited sensory neurons during singing (Poulet 2002, 2003a, 2003b).

Eve Marder showed that several very different parameter sets (conductances and synaptic strength) for a model of the central pattern generator in crustaceans lead to very similar spiking behavior that all match the experimentally observed one (Prinz 2003, Golowasch 2002).

In weakly electric fish Sternopygus the frequency of its electric organ discharge (EOD) and the width of single discharges can be quite variable. Harlod Zakon showed that the different time scales are correlated with different expressions of potassium channel and sodium channel subunits.

In leech the coupling of two motorneurons is modulated by the NS neuron via rectifying gap junctions as Lidia Szczupak pointed out (Rela 2003).

Frederic Theunissen presented some data on coding of sound identity in zebra finch. I want to discuss his use of information theory...
Literature:
Conference homepage
Conference program and abstracts
Golowasch J, Goldman MS, Abbott LF, Marder E. (2002): "Failure of averaging in the construction of a conductance-based neuron model" J Neurophysiol., Vol. 87, pages 1129-1131
Poulet JF, Hedwig B (2003): "Corollary discharge inhibition of ascending auditory neurons in the stridulating cricket" J Neurosci., Vol. 23, pages 4717-4725
Poulet JF, Hedwig B (2003): "A corollary discharge mechanism modulates central auditory processing in singing crickets" J Neurophysiol., Vol. 89, pages 1528-1540
Poulet JF, Hedwig B (2002): "A corollary discharge maintains auditory sensitivity during sound production" Nature, Vol. 418, pages 872-876
Prinz AA, Billimoria CP, Marder E. (2003): "Alternative to hand-tuning conductance-based models: construction and analysis of databases of model neurons" J Neurophysiol., Vol. 90, pages 3998-4015
Rela L, Szczupak L. (2003): "Coactivation of motoneurons regulated by a network combining electrical and chemical synapses" J Neurosci., Vol. 23, pages 682-692

13.09.2004 Inés Samengo (Centro Atómico Bariloche, Argentina)
"Defining a subjective distance: quantifying the metric structure of confusion""

20.09.2004 Christian Michaelis
"Shaping STDP learning by the timecourse of pre- and postsynaptic signals"
Abstract: The talk will introduce and discuss a paper by Saudargiene et al. that recently was published in Neural Computation. The authors develop a model for STDP learning and show that the resulting shape of the learning window changes if the timecourse of pre- or postsynaptic signals is altered.
Literature:
Ausra Saudargiene, Bernd Porr, and Florentin Wörgötter (2004): "How the Shape of Pre- and Postsynaptic Signals Can Influence STDP: A Biophysical Model" Neural Comp. 16: 595-625

27.09.2004 Marion Moisel
"Principal Component Analysis of Hippocampal Mossy Fiber EPSPs"
Abstract: Synapses exhibit different forms of plasticity over a wide range of time scales. The mossy fiber synapse plays a decisive role in information transmission in the hippocampus, since it holds a bottleneck position by providing the main feedforward input into the CA3-region. Earlier studies on the mossy fiber synapse have shown its specific properties such as a large mean unitary postsynaptic potential size and pronounced frequency facilitation.Extracellular postsynaptic field potentials (fEPSPs) were recorded from mouse hippocampal slices. These field potential recordings exhibit a good signal-to-noise ratio because they provide an average over many mossy fiber synapses. A disadvantage of extracellularly recorded data is that responses of mossy fibers are superimposed by other processes, such as fiber excitability, activation of associational/commissural synapses and action potential generation of CA3 pyramidal cells. We have analyzed fEPSPs by means of Principal Component Analysis (PCA) in order to separate pure mossy fiber signals from underlying processes. PCA indicates the presence of at least two different processes which could not be separated due to their interdependence. Spikes of CA3 pyramidal cells for instance occur only when synapses are strongly facilitated and evoke large fEPSP amplitudes. Application of PCA to simulated fEPSPs that consisted of mossy fiber EPSPs and CA3 pyramidal cell spikes leads to similar results as PCA applied to neurophysiological data. Analysis of data subsets leads to the identification of the two dependent processes. We conclude that in experiments the mossy fiber contribution to fEPSP amplitudes is overestimated due to action potentials of CA3 pyramidal cells. Intracellular patch clamp recordings, which do not contain superimposed spikes, support this hypothesis.
Literature:
AE Dityatev, RS Altinbaev, AV Astrelin, LL Voronin (2003): "Combining principal component and spectral analyses with the method of moments in studies of quantal transmission" J Neursoci Meth (2003) 130:173-199
Salin PA, Scanziani M, Malenka RC, Nicoll RA (1966) "Distinct short-term plasticity at two excitatory synapses in the hippocampus" Proc Natl Acad Sci U S A. 1996 Nov 12;93(23):13304-9
Zucker RS, Regehr WG (2002) "Short-term synaptic plasticity" Annu Rev Physiol. 2002;64:355-405. Review

04.10.2004 Susanne Schreiber

11.10.2004 SFB-Autumn School "Learning and Memory"

18.10.2004 Thomas Strösslin (Laboratory of Computational Neuroscience, EPFL, Lausanne, Switzerland)
"A Connectionist Model of Spatial Learning in the Rat"

Abstract: When animals explore an environment, they store useful spatial information in their brains. In subsequent visits, they can recall this information and thus avoid dangerous places or find again a food location. This ability, which may be crucial for the animal's survival, is termed ``spatial learning''.

In the late 1940s, theoretical considerations have led researchers to the conclusion that rats establish a ``cognitive map'' of their environment. This spatial representation can then be used by the animal in order to navigate towards a rewarding location. In 1971, researchers have for the first time found direct evidence that the \emph{hippocampus}, a brain area in the limbic system, may contain such a cognitive map. The activity of neurons in the hippocampus of rats tends to be highly correlated with the animal's position within the environment. These ``place cells'' have since been the target of a large body of research.

Apart from spatial learning, the hippocampus seems to be involved in a more general type of learning, namely in the formation of so-called ``episodic memories''. Models of hippocampal function could thus provide valuable insights for the understanding of memory processes in general. Insights from animal navigation could also prove beneficial for the design of autonomous mobile robots. constructing a consistent map of the environment from experience, and using it for solving navigation problems are difficult tasks. Incorporating principles borrowed from animal navigation may help building more robust and autonomous robots.

The main objective of this thesis is to develop a neural network model of spatial learning in the rat. The system should be capable of learning how to navigate to a hidden reward location based on realistic sensory input. The system is validated on a mobile robot.

Our model consists of several interconnected brain regions, each represented by a population of neurons. The model closely follows experimental results on functional, anatomical and neurophysiological properties of these regions. One population, for instance, models the hippocampal place cells. A head-direction system closely interacts with the place cells and endows the robot with a sense of direction. A population of motor-related cells codes for the direction of the next movement. Associations are learnt between place cells and motor cells in order to navigate towards a goal location.

This study allows us to make experimental predictions on functional and neurophysiological properties of the modelled brain regions. In our validation experiments, the robot successfully establishes a spatial representation. The robot can localise itself in the environment and quickly learns to navigate to the hidden goal location.


25.10.2004 Kit Longden (School of Informatics, The University of Edinburgh, Scotland)
"The role of the cortical input to CA1 in hippocampal information processing"

Abstract: CA1 is the principal source of outputs from the hippocampus, however, the function of both CA1 and its cortical input remains unclear. Brun et al. (2002, Science, 296(5576):2243-6) recently reported normal CA1 place field activity driven by the cortical input alone, after CA3 had been lesioned. This was a surprising result, since the response of the cortical input pathway to stimulation is almost exclusively inhibitory. A leaky integrate-and-fire model of CA1 place field formation in the cortical pathway is presented, demonstrating the mechanisms by which CA1 place fields can form. The results are discussed in the light of current papers dissociating recordings of neuronal ensemble activity in CA3 and CA1 on the basis of environmental similarity and experience.


01.11.2004 Roland Schätte

08.11.2004 Martin Stemmler
"The Return of Resonance (Part 451)"
Abstract: In the subthreshold regime, the membrane potential of some nerve cells exhibits a resonance in the frequency of the driving input, even in the absence of spontaneous membrane potential oscillations. A recent claim in the literature attributes the difference between the resonance frequency and oscillation frequency to noise shaping. This talk will present a critical reappraisal of this claim.
Literature:
Richardson, Brunel, and Hakim (2003): "From subthreshold to firing rate resonance." J Neurophysiol. Vol. 89, pages 2538-2554

15.11.2004 Tim Oppermann
"How are resonance and subthreshold oscillations connected?
Work in progress"


It has been well established that Stellate cells in the entorhinal cortex
exhibit subthreshold resonance in the region of the theta rhythm. In this
context, a resonance describes the phenomenon, that a sinusoidal input
current of fixed strength, leads to a sinusoidal membrane potential, where
special frequencies are enhanced, i.e. the amplitude of the potential is
amplified.

In our biophsical stellate cell model both, the hyperpolarization activated
mixed cation current (h-current) and the persistent sodium current,
are important. We analyze the subthreshold properties of a reduced stellate cell integrate and fire
model, and hypothesize that there is a close connection between resonance
phenomena and intrinsic subthreshold oscillations.

Literature:
Hutcheon, Yarom, TINS, 2003
        http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Abstract&list_uids=10782127
Richardson et al., J.Neurophys., 2003
        http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Abstract&list_uids=12611957
Dickson et al., J. Neurophys.,2000
        http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=10805658

22.11.2004 Raphael Ritz
"Software Development - Advanced Concepts and Techniques"

29.11.2004 Felix Creutzig
"The Ganglion Neuron AN12 Encodes Relevant Song Features"
Abstract: Grasshoppers communicate via species specific songs. From behavioural experiments it is known that a syllable-pause pattern is crucial for successful communication. One specific ganglion neuron - the so called AN12 - encodes the relevant information in the spike count. We were able to model the system with a simple Integrate & Fire neuron and a fast excitatory and a slow inhibitory channel.
Literature:
Ronacher B. and Krahe R. (1998): "Song recognition in the grasshopper Chorthippus biguttulus is not impaired by shortening song signals: implications for neural encoding" J Comp Physiol A Vol. 183, pages 729-735
Lesica, N. and Stanley G. (2004): "Encoding of Natural Scene Movies by Tonic and Burst Spikes in the Lateral Geniculate Nucleus" The Journal of Neuroscience, Vol. 24, pages 10731-10740

06.12.2004 Alexander Wolf
"Discriminating between Nonlinearities in Auditory Signal Transduction Chains"
Abstract: The transformation of an acoustic signal into a neural response can be modeled as cascade containing two linear filters and two nonlinearities (LNLN-cascade). Using the Iso-Response-Approach, Tim Gollisch succeeded in extracting the dynamical features of the filter functions under the assumption that the nonlinear coupling is a square. The question arises, how this functionaly can be interpreted. A statistical approach is proposed that leads to new nonlinear coupling. This coupling reproduces previous results for small sound amplitudes, but show deviations for higher sound pressure. Experiments are suggested, that should enable to discriminate between different models.
Literature:
A.J.Hudspeth, Y.Choe, A.D.Mehta, and P.Martin (2000): "Putting ion channels to work: Mechanoelectrical transduction, adaptation, and amplification by hair cells" PNAS Colloqium, vol.97, no.22, pages 11765-11772
T.Gollisch (2002): "Sound Intensity Coding in an Insect Auditory System" J. Neurosci., Vol. 22(23), pages 10434-10448

13.12.2004 Samuel Glauser
"Examining spike precision in locust auditory receptor neurons"
Abstract: The focus of this talk is on spike precision in locust auditory receptor neurons and how this precision can be measured. Some data from electrophysiological experiments is already available and the method proposed by Schreiber et al. will be used to examine the precision of those spikes.
Literature:
S. Schreiber, J.M. Fellous, D. Whitmer, P. Tiesinga, T.J. Sejnowski: "A new correlation-based measure of spike timing reliability" Neurocomputing 52-54 (2003) 925-931

20.12.2004 Christian Leibold
"Effects of LTP on the Dynamics of the Mossy Fiber Synapse"
Abstract: The hippocampal mossy fiber synapse connects granule cells of the dentate gyrus to pyramidal cells of the CA3 region. The synapse exhibits strong facilitation with an amplitude increase of the postsynaptic response up to 10 fold, at least in vitro. Based on field potential recordings from Dietmar Schmitz's lab, we have developed a model that can describe synaptic short-term dynamics by means of seven parameters. The model has two time constants; a shorter one accounting for paired-pulse facilitation and a longer one describing frequence facilitation. By means of our decription we can quantify the changes of the short-term dynamics upon temperature increase and tetanically induced long-term potentiation. The former leads to a reduction of the long time-constant, whereas the latter, in addition to an enhanced base-line amplitude, also reduces the paired-pulse time constant.
Literature:
Salin PA, Scanziani M, Malenka RC, Nicoll RA. Distinct short-term plasticity at two excitatory synapses in the hippocampus. Proc Natl Acad Sci U S A. 1996 Nov 12;93(23):13304-9.
Varela JA, Sen K, Gibson J, Fost J, Abbott LF, Nelson SB. A quantitative description of short-term plasticity at excitatory synapses in layer 2/3 of rat primary visual cortex. J Neurosci. 1997 Oct 15;17(20):7926-40.


05.01.2005 Wednesday 16:00! Babette Dellen (Washinton University, St. Louis, USA)
"Computing relative motion with complex cells"

Abstract: Contextual influences shape our perception of local visual stimuli. Relative-motion stimuli represent an important contextual influence, yet the mechanism subserving relative-motion computation remains largely unknown. In the present work, we investigated the responses of an established model for simple and complex cells to relative-motion stimuli. A straightforward mathematical analysis showed that relative-motion computation is inherent in the nonlinear transformation of the complex-cell model. Tuning to relative velocity is achieved by applying a temporal filter to the complex-cell response.The mathematical inference is supported by simulations that quantitatively reproduce measured complex-cell responses in both cat and monkey to a variety of relative-motion stimuli. Importantly, the posited mechanism for cortical computation of relative motion does not require an intermediate neural representation of local velocities and does not require lateral or feedback interactions within a network.

10.01.2005 Pietro Berkes
Analysis of inhomogeneous quadratic forms for physiological and theoretical studies

Abstract: This is my second talk about quadratic forms at the Neuro Talks. This time I'm going to skip most of the maths and concentrate on some new results and their relation with other physiological and theoretical studies.

Literature:
Berkes, P. and Wiskott, L. (2005). Analysis of inhomogeneous quadratic forms for physiological and theoretical studies. (Manuscript in preparation).
Hashimoto, W. (2003). Quadratic forms in natural images. Network: Computation in Neural Systems, 14(4):765-788.
Rust, N.C., Schwartz, O., Moxshon, J.A., and Simoncelli, E. (2004). Spike-triggered characterization of excitatory and suppressive stimulus dimensions in monkey v1. Neurocomputing, (58-60):793-799.
Hyvaerinen, A. and Hoyer, P. (2001). A two-layer sparse coding model learns simple and complex cell receptive fields and topography from natural images. Vision Research, 41(18):2413-2423.
Koerding, K., Kayser, C., Einhaeuser, W., and Koenig, P. (2004). How are complex cell properties adapted to the statistics of natural scenes? Journal of Neurophysiology, 91(1):206-212.

17.01.2005 Claudia Wolf (Max-Delbrück Centrum, Berlin)
On the final number of dentate granule cells: Lifelong turnover or net increase?

Abstract: Adult hippocampal neurogenesis is an exception to the rule that adult brain cannot produce new neurons. Several recent reports suggested that these new neurons would provide cells for a substantial turnover of the dentate gyrus granule cells and functional interpretations have been proposed based on this lifelong cellular replacement (e.g. Deisseroth et al., 2004). Despite a large number of studies on adult hippocampal neurogenesis, no data were available to allow estimating the net contribution of adult hippocampal neurogenesis to the final neuron numbers and the size of the dentate gyrus and whether such a turnover might exist.

The aim of this investigation was to find out which consequences would derive for the net cell numbers in the granule cell layer from adult hippocampal neurogenesis during aging. Proliferation of precursor cells as a measure of the potential for new granule neurons was assessed and the absolute granule cell numbers, the granular layer volume and the granule cell density in the granule cell layer were investigated for several postnatal ages in mice.

The absolute numbers of BrdU-positive cells and the absolute numbers of BrdU-positive cells that were also positive for neuronal markers like Doublecortin (DCX) and Neuronal Nuclei (NeuN) were determined. In accordance with earlier investigations (e.g. Kempermann et al., 1998; Kuhn et al., 2003; Cameron et al, 1999), the obtained results confirm the decrease of neurogenesis with increasing age. Because, the absolute numbers of proliferating cells as well as the absolute numbers of newborn potential neurons declined significantly towards old age.

The absolute granule cell numbers per granular layer were estimated applying the optical fractionator to Hoechst stained hippocampal slices of 8, 36, 49, 270, 365 and 730 day-old mice.

More than 50% of granule cells were already present in the dentate gyrus of only 8 day-old mice. Adding of granule cells continued on a lower level reaching a significant increase at the age of 49 days and the maximum of absolute granule cell numbers at the age of 270 days. There was an addition of approximately 216000 granule cells in the dentate gyrus from postnatal day 8 to the age of 9 months in mice. Afterwards, numbers did not change significantly anymore.

The net gain of absolute granule cells is a consequence of the continuous granule cell generation and the maintenance of integrated granule cells. The acquisition of new granule cells and the persistence of once incorporated granule cells continues throughout life.

Early in life, more granule cells are generated leading to higher increases in absolute numbers. In old age, fewer cells are produced, thus reducing the increase of granule cells. The maintenance of once acquired cells is responsible for the high absolute granule cell numbers in old age.

In other words despite the attenuation of neurogenesis at old ages, the absolute granule cell numbers not change significantly in old age. These facts contradict the assumption of a massive replacement of granule cells or a so-called “turn-over”.

There was also a volume increase of the granular layer and a density decrease until the age of 270 days. Then the volume decreased significantly and the density increased significantly whereas the absolute numbers of granule cells remained nearly unchanged.

The adding and guarding of granule cells throughout life and strong coupling of neurogenesis and network activity suggests a functional role for these newly integrated cells for structural brain plasticity and to cope with network demands.

24.01.2004 Kay Thurley
"Hippocampal phase precession through synaptic facilitation"

Abstract: During spatial behaviour, the phase of pyramidal cell spikes in the hippocampus is strongly correlated to the rat's location within the place field. This phenomenon is referred to as hippocampal phase precession. I'm going to present a model, which hypothesizes that the short term dynamics of the mossy fiber synapse, in particular its facilitation, is the biophysical mechanism for phase precession.

Literature:
J. O'Keefe and M.L. Recce (1993): "Phase relationship between hippocampal place units and {EEG} theta rhythm" Hippocampus, Vol. 3, pages 317-330
W.E. Skaggs, B.L. McNaughton, M.A. Wilson and C.A. Barnes (1996): "Theta phase precession in hippocampal neuronal populations and the compression of temporal sequences" Hippocampus, Vol. 6, pages 149-172
M.R. Mehta, A.K. Lee and M.A. Wilson (2002): "Role of experience and oscillations in transforming a rate code into a temporal code" Nature, Vol. 417, pages 741-746
K.D. Harris, D.A. Henze, H. Hirase, X. Leinekugel, G. Dragoi, A. Czurko and Buzsaki, G. (2002): "Spike train dynamics predicts theta-related phase precession in hippocampal pyramidal cells" Nature, Vol. 417, pages 738-741

31.01.2004 Jan Benda
"Dynamics of intensity invariance in an auditory interneuron of the cricket."

Abstract: The AN1 is a first order interneuron of the cricket's auditory system. I present data (recorded by Matthias Hennig) demonstrating that the response of the AN1 is independent of the mean intensity of the stimulus. It achieves this intensity invariance by shifting its intensity-response curve to the appropriate intensity. We can succesfully explain the AN1 response by a variant of a general model for spike-frequency adaptation (Benda & Herz, 2003). For a wide range of intensities (about 30 dB) the adaptation results in a mean firing rate of the response of about 100 Hz. This firing rate sets a fixed temporal resolution for changes of the stimulus around its mean intensity.

Literature:
J. Benda and Andreas V. M. Herz (2003): "A universal model for spike-frequency adaptation" Neural Computation, Vol 15, pages 2523-2564

07.02.2005 Mathias Franzius
"Review: models of visual object recognition"

21.02.2005 Participants of the Göttingen Neurobiology Meeting
"Report of Meeting Highlights"

28.02.2005 Tiziano Zito
"Nonlinear Blind Source Separation: a review"

Abstract: In this talk, I briefly review the state of the art in blind source separation (BSS) foro nonlinear mixing models. After a general introduction to the nonlinear BSS and ICA (independent component analysis) problems, I will discuss in more detail existence and uniqueness issues for solutions. I will then give an overview of various separation techniques and present the work that has been done in our group.

Literature:
A. Hyvärinen, J. Karhunen, E. Oja (2001): "Independent Component Analysis", J. Wiley.
C. Jutten, J. Karhunen (2003): "Advances in Nonlinear Blind Source Separation", Proc. of the 4th Int. Symp. on Independent Component Analysis and Blind Signal Separation (ICA2003), pages 245-256.
T. Blaschke, L. Wiskott (2004): "Independent Slow Feature Analysis and Nonlinear Blind Source Separation", Proc. of the 5th Int. Conf. on Independent Component Analysis and Blind Signal Separation (ICA2004), pages 742-749.

07.03.2005 Andreas Engel (Institute of Neurophysiology and Pathophysiology, University Hospital Hamburg-Eppendorf)
"Fast neuronal oscillations in sensory processing: influence of bottom-up and top-down factors "

Abstract: Cognitive functions like perception, attention, memory or language are based on highly parallel and distributed information processing by the brain. One of the major unresolved questions is how information can be integrated and how coherent representational states can be established in the distributed neuronal systems subserving these functions. It has been suggested that this so-called "binding problem" may be solved in the temporal domain. The hypothesis is that synchronization of neuronal discharges can serve for the integration of distributed neurons into cell assemblies and that this process may underlie the selection of perceptually and behaviourally relevant information. Moreover, it has been suggested that fast oscillations at frequencies in the so-called gamma-range (>30 Hz) may help to entrain spatially separate neurons into synchrony and, thus, may indirectly promote the dynamic binding of neuronal populations. When transient stimuli are used, such oscillations typically manifest themselves as an 'early' and a 'late' gamma-band response, occurring around 100ms and around 300-500ms post-stimulus, respectively. In this talk, experimental data obtained in animal as well as human measurements will be discussed that demonstrate a relevance of temporal binding mechanisms for sensory integration and perceptual selection. These data suggest that gamma-band synchrony in sensory pathways is influenced by both bottom-up (stimulus coherence, stimulus saliency) and top-down factors (attention, memory contents). Based on these data, we propose a new framework that relates gamma-band synchrony to two underlying processes, namely, the comparison of memory contents with stimulus-related information and the utilization of signals derived from this comparison. This model attempts to explain 'early' gamma-band responses in terms of the match between bottom-up and top-down information. Furthermore, it assumes that 'late' gamma-band activity reflects the readout and utilization of the information resulting from this match.

14.03.2005 Susanne Schreiber
"Literature review: Experimentally measured phase-response curves and synchronization in hybrid neuronal networks"

Abstract: The paper of Netoff et al. (2005) presents experimentally measured phase-response curves for entorhinal stellate cells. These changes in spike timing induced by artificial synaptic inputs characterize the dynamics of individual neurons. The authors show that the single cell experiments can predict network synchronization behavior, which they measure with the dynamic clamp technique in hybrid neuronal networks of the hippocampal formation.

Literature:
T.I. Netoff et al. (2005): "Synchronization in Hybrid Neuronal Networks of the Hippocampal Formation." J Neurophysiol, Vol. 93, pages 1197-1208

21.03.2005
16:00!
Ralf Wessel (Physics Department, Washington University)


Created by Laurenz Wiskott, http://itb.biologie.hu-berlin.de/~wiskott/