Monday, September 27, 2010
|
| Morning |
Registration |
| |
|
Bernstein Award 2010 |
| 11:30 - 11:45 |
Welcome
H104 |
Klaus-Robert Müller |
| 11:45 - 12:00 |
Bernstein Award
H104
|
Bernstein Network Computational Neuroscience |
| 12:00 - 12:30 |
Award Keynote
H104
|
Bernstein Awardee 2010
|
| 12:30 - 13:15 |
Press Conference
H104
|
|
| 13:15 - 14:15 |
Lunch Break (Lichthof)
Fingerfood |
| |
|
brains 4 brains |
| 14:15 - 15:15 |
brains4brains Award
H104 |
Bernstein Network Computational Neuroscience
|
| 15:15 - 15:30 |
Coffee Break |
| |
Session 1a |
Chair: Jochen Triesch |
| 15:30 - 15:50 |
Talk
H104 |
Reshad Hosseini
New estimate for the redundancy of natural images |
| 15:50 - 16:10 |
Talk
H104 |
Alberto Mazzoni
Neurons in primary visual cortex encode naturalistic visual information using multiple temporal scales |
| 16:10 - 16:30 |
Talk
H104 |
Sander Bothe
Multiple timescale Continuous-time Coding with Spiking Neurons |
| 16:30 - 17:00 |
Coffee Break |
| |
Session 1b |
Chair: Matthias Bethge |
| 17:00 - 17:20 |
Talk
H104 |
Orlando Arévalo
Predicting the dynamics of dual prism adaptation with a biophysically plausible neural model |
| 17:20 - 17:40 |
Talk
H104 |
Jens Kleesiek
Object Affordances in the Context of Sensory Motor Contingencies |
| 17:40 - 18:00 |
Talk
H104 |
Sohrab Saeb
Learning Coordinated Eye and Head Movements: Unifying Principles and Architectures |
Tuesday, September 28, 2010
|
| |
Session 2a |
Chair: John-Dylan Haynes |
| 09:00 - 09:45 |
Keynote
H104 |
Lars-Kai Hansen
Machine learning strategies for fMRI analysis
Brain imaging by fMRI has become a cornerstone in neuroscience. However, statistical analysis of fMRI still faces many interesting challenges including non-linearity and multi-scale spatial and temporal dynamics. I will discuss machine learning strategies invoked for fMRI modeling, present a general framework for model evaluation and visualization, and discuss our recent progress in detection and understanding of activation networks.
|
| 09:45 - 10:05 |
Talk
H104 |
Jakob Heinzle
Cortico-cortical receptive fields: Topographic organization of intrinsic functional connectivity between V1 and V3 in the human brain |
| 10:05 - 10:35 |
Coffee Break |
| |
Session 2b |
Chair: Clemens Boucsein |
| 10:35 - 10:55 |
Talk
H104 |
Annette Witt
Dynamic effective connectivity analysis of goal-directed reaching |
| 10:55 - 11:15 |
Talk
H104 |
Matthias Treder
Towards gaze-independent visual brain-computer interfaces |
| 11:15 - 11:35 |
Talk
H104 |
Stefan Lang
Simulation of innervation and activation scenarios of morphologically detailed, large-scale neuron networks in a column of the primary somatosensory cortex with NeuroDUNE |
| 11:35 - 13:35 |
Lunch Break |
| |
Session 3a |
Chair: Klaus-Robert Müller |
| 13:35 - 14:20 |
Keynote
H104 |
Ernst Fehr
The Neuroeconomics of Social Norm Compliance
All known human societies establish social order by punishing cheaters and norm violators. In recent years, neuroeconomists have discovered important components of the neural circuitry underlying human norm obedience and norm enforcement. The lecture will document that the prefrontal cortex - a brain area particularly well developed in humans - is key in this human ability. Non-invasive down-regulation of neural activity in prefrontal cortex reduces norm compliance despite the fact that individuals are still able to distinguish between "right" and "wrong". Neuroeconomic research on young children - whose prefrontal cortex is not yet well developed - shows similar patterns. These results thus indicate a dissociation between the ability to obey social norms and the knowledge of the content of the social norms, which complicates the attribution of responsibility for norm violations.
|
| 14:20 - 14:40 |
Talk
H104 |
Emili Balaguer-Balester
Identifying Attracting Dynamics of Cortical Populations during Decision-Making Tasks |
| 14:40 - 15:10 |
Coffee Break |
| |
Session 3b |
Chair: Richard Kempter |
| 15:10 - 15:30 |
Talk
H104 |
Alexander Mathis
How Good is Grid Coding versus Place Coding for Navigation Using Noisy, Spiking Neurons? |
| 15:30 - 15:50 |
Talk
H104 |
Sen Cheng
From grid cells to place cells: a generic and robust principle accounts for multiple spatial maps |
| 15:50 - 16:20 |
Coffee Break |
| 16:20 - 19:00 |
Poster Session
Lichthof |
|
Wednesday, September 29, 2010
|
| |
Session 4a |
Chair: Gabriel Curio |
| 09:00 - 09:45 |
Keynote
H104 |
Pascal Fries
Routing and computing with neuronal synchronization
Selective attention requires the flexible communication among several brain areas. We have recorded large parts of the visual attention network, including areas V1, V4, posterior parietal, premotor and prefrontal areas simultaneously with a 252-channel subdural grid electrode. We found topographically specific gamma- and beta-band synchronization among areas. Gamma-band synchronization was stronger in the feedforward direction and beta-band synchronization in the feedback direction. Attention to a contralateral stimulus enhanced the precision of these directed rhythmic inter-areal interactions. The results suggest that rhythmic synchronization subserves the effective interaction among neurons. Such effective interaction is also crucial for the generation of stimulus selective neuronal responses. We have tested whether neuronal selectivity is modulated by the gamma rhythm. We found that the most precisely gamma-synchronized spikes are more stimulus selective than spikes occurring at any other time in the gamma cycle and more selective than the average firing rate. The optimally gamma-aligned spikes are also least corrupted by noise correlation. Thus, gamma-band synchronization creates spikes that are maximally selective, least noise-corrupted and most effectively communicated to other brain areas.
|
| 09:45 - 10:05 |
Talk
H104 |
Christian Hauptmann
Restoration of segregated, physiological neuronal connectivity by desynchronizing stimulation |
| 10:05 - 10:35 |
Coffee Break |
| |
Session 4b |
Chair: Laurenz Wiskott |
| 10:35 - 10:55 |
Talk
H104 |
Dominika Lyzwa
Spatio-temporal features of stimulus-related activity in the inferior colliculus |
| 10:55 - 11:15 |
Talk
H104 |
Adrien Jouary
Spike-Based Population Coding of Interaural Time Difference (ITD). |
| 11:15 - 11:35 |
Talk
H104 |
Thomas Wachtler
Efficient Data Management for Neurophysiology at the German Neuroinformatics Node |
| 11:35 - 13:35 |
Lunch Break |
| |
Session 5a |
Chair: Michael Brecht |
| 13:35 - 14:20 |
Keynote
H104 |
Peter Jonas
The ‘in’ and ‘out’ of
GABAergic interneurons in the hippocampus
Fast-spiking, parvalbumin-expressing basket cells (BCs) play a key role in the function of hippocampal microcircuits. However, the subcellular properties of this important class of GABAergic interneuron are incompletely understood.
To study the dendrites of BCs, we made dendritic patch-clamp recordings up to 300 µm from the soma. Recordings were performed in the dentate gyrus, the input region of the hippocampus, using slices cut from 17 – 22-day-old rats. Direct recordings revealed that the action potential was initiated near the soma, presumably in the axon, and propagated into the dendrites with marked amplitude attenuation. Analysis of conductance density demonstrated that BC dendrites showed a high K+ channel density, but a low Na+ channel density. Modeling of synaptic integration revealed that these specific dendritic properties promote coincidence detection and ensure single-spike generation following synaptic input.
To examine the properties of transmitter release from presynaptic BC terminals, we performed paired recordings between synaptically connected BCs and granule cells in hippocampal slices. Analysis of the effects of the Ca2+ chelators introduced by intracellular pipette perfusion revealed that BAPTA suppressed synaptic transmission much more efficiently than EGTA. Furthermore, the relation between transmitter release and the amplitude of the presynaptic Ca2+ transient during application of a slow blocker was more linear than that during reduction of extracellular Ca2+ concentration. Modeling of the concentration dependence of the chelator effects revealed that the distance between Ca2+ channels and Ca2+ sensors of exocytosis was 10 – 20 nm. Furthermore, modeling of the relation between transmitter release and amplitude of the presynaptic Ca2+ transient revealed that two or three open Ca2+ channels trigger transmitter release at BC output synapses.
In conclusion, fast-spiking, parvalbumin-expressing BCs are specialized at both the input and the output level. These specializations promote the generation of rapid feedforward and feedback inhibitory signals, which may be important for temporal encoding of information in principal neurons in neuronal networks of the hippocampus.
|
| 14:20 - 14:40 |
Talk
H104 |
Clemens Boucsein
Number, reliability and precision of long-distance projections onto neocortical layer 5 pyramidal neurons |
| 14:40 - 15:10 |
Coffee Break |
| |
Session 5b |
Chair: Benjamin Blankertz |
| 15:10 - 15:30 |
Talk
H104 |
Armin Biess
Calcium spread in crowded dendrites |
| 15:30 - 15:50 |
Talk
H104 |
Jonathan Caplan
Homeostatic Regulation of Neuronal Activity with Temperature Variations |
| 15:50 - 16:20 |
Coffee Break |
| 16:20 - 19:00 |
Poster Session
Lichthof |
|
Thursday, September 30, 2010
|
| |
Session 6a |
Chair: Andreas Herz |
| 09:00 - 09:45 |
Keynote
H104 |
Misha Tsodyks
Attractor neural network models of space representations
Attractor networks were proposed as a possible neuronal underpinning of place representation in Hippocampal formation. I will show how empowering them with short-term synaptic plasticity can explain a surprisingly wide repertoire of phenomena in a unified manner. In particular, the network can exhibit fast transitions between theta-modulated activity and sharp waves, and generate recently observed replay of place-specific activity during immobility. When several correlated environments are stored in the same network, there can be a strong interaction between them even when one environment is activated by the external cue.
|
| 09:45 - 10:05 |
Talk
H104 |
Dimitrije Markovic
Intrinsic plasticity in autonomous recurrent neural networks |
| 10:05 - 10:35 |
Coffee Break |
| |
Session 6b |
Chair: Christian Leibold |
| 10:35 - 10:55 |
Talk
H104 |
Daniel Krieg
An objective function for STDP: increasing the separability in self-organized recurrent neural networks |
| 10:55 - 11:15 |
Talk
H104 |
Felipe Gerhard
Estimating small-world topology of neural networks from multi-electrode recordings |
| 11:15 - 12:00 |
Final Keynote
H104 |
Gerwin Schalk (tentative)
Perception and Cognition in Human Electrocorticographic Signals
Recent developments have sparked substantial interest in recordings from the surface of the brain (electrocorticography (ECoG)) to investigate the basis of normal brain function related to motor control, language, or memory, as well as of abnormal function such as epileptic seizures. For the past several years, my laboratory has utilized human ECoG recordings to study neural correlates of motor, language, and cognitive function. In this talk, I will describe the types of signals that can be detected in ECoG and the emerging understanding of how they relate to each other. I will then demonstrate that ECoG encodes detailed aspects of auditory perception, auditory and visual spatial attention, and language function at high spatial and temporal resolution. |
| 12:00 - 12:15 |
Goodbye |
| 12:15 - 14:00 |
Lunch Break |
| |
| |
Special Event |
GRK Opening Ceremony |
| 14:00 - 14:30 |
Greetings
H104 |
|
| 14:30 - 14:45 |
Music
H104 |
|
| 14:45 - 15:30 |
Keynote
H104 |
Alain Destexhe, Centre National de la Recherche Scientifique (CNRS)
Combining experiments and theory to investigate stochastic brain states |
| 15:30 - 16:00 |
Coffee Break |
| 16:00 - 16:45 |
Keynote
H104 |
Maneesh Sahani, University College London
Watching time go by: the statistics of the sensory environment contributes to estimates of temporal intervals
Recent experiments have shown that the nature of a stimulus can have a pronounced effect on observers' estimates of its duration. Such observations are not easy to reconcile with standard models of temporal judgments in animals, which focus on mechanisms that exploit internally-driven dynamics---oscillations within specialised timing circuitry, predicatably sequenced dynamic behaviours, or the evolution of general nonlinear recurrent neural networks---to generate the timing signal.
I will suggest that the stimulus-interaction results are best understood in a framework where observers combine internal processes with temporal information available from the environment. In particular, I will propose a new frameowrk for stimulus-based timing, in which statistical expectations about the change in sensory input are combined with sensory observations to yield probabilistic estimates of the elapsed time. Results of two novel behavioural experiments provide support for this view, showing that stochastic stimuli that evolve differently from observers' expectations bias their temporal judgments, while stimuli consistent with expectations actually improve the accuracy of timing.
A stumbling block for many timing models has been the need to provide a natural account for the scalar property of the variance in timing judgements---the temporal Weber law. The new change-based framework (by contrast to schemes based on counting) robustly predicts that stimulus-derived judgments will be Weberian, a prediction that is borne out by the results of a third new experiment. This scalar behaviour of the model framework suggests that a similar approach to internal processes may also provide a naturally Weberian account of stimulus-independent timing. As such the change-based framework may, in the end, offer a comprehensive model for the perception of temporal intervals.
|
| 16:45 - 17:30 |
Keynote
H104 |
Jan Koendrink, Delft University of Technology
Vision: The Optical User Interface
An often quoted example of a fixed-action pattern is the egg-rolling behavior of the graylag goose: the bird will even attempt to roll a brick back to its nest. It apparently “takes a brick for an egg”, despite its excellent visual acuity. Evolution optimizes utility, not veridicality. Perversely, textbooks take it for a fact that human vision has evolved so as to approach veridical perception. But do humans indeed escape the laws of evolution? No, human vision is an idiosyncratic user interface. I draw examples from pictorial perception. Gleaning information from still images is still an important human ability and is likely to remain so for the foreseeable future. I discuss a number of instances of extreme non-veridicality and huge inter-observer variability. Despite their importance in applications (just think of information dissemination, personnel selection, …) such huge effects have remained largely undocumented. They can be traced to artistic conventions though. The reason is that conventional psychophysics fails to address the qualitative aspect by design. But these qualitative aspects are exactly the meaningful aspects of visual awareness, which is why they are the target of the visual arts. I consider possible extensions of classical psychophysics.
|
| 17:30 - 17:45 |
Music
H104 |
|
| 17:45 - Open End |
Reception
Lichthof |
Music/Food/Fingerfood
|
Friday, October 1, 2010
|
| |
Special Event |
PhD Student Symposium |
| 09:00 - 09:45 |
Talk
H104 |
Klaas Enno Stephan
Model-based inference on synaptic mechanisms of (mal)adaptive behaviour
Synaptic plasticity and its regulation by modulatory transmitters, such as dopamine or acetylcholine, are essential for learning and decision-making and thus represent a neurophysiological cornerstone of adaptive behavior. They have also been identified as key mechanisms in the pathophysiology of various psychiatric diseases. However, non-invasive techniques for measuring these processes in the human brain remain to be established. This presentation outlines a research program that aims to fill this gap by developing models for inference on synaptic plasticity and neuromodulation. This approach rests on three key advances:
(i) developing physiologically interpretable models of neuronal population dynamics, that can be fitted to non-invasively measured brain activity data (such as fMRI or EEG) from individual subjects and provide probabilistic estimates of underlying synaptic processes,
(ii) embedding computational models of learning into neurophysiological models to prescribe the expected dynamics of plasticity,
(iii) validating models using neuropharmacology and invasive recordings in human and animal studies.
Given successful validation studies, model-based assays of synaptic plasticity and neuromodulation would enable a more mechanistic understanding of individual (mal)adaptive behavior and may lead to pathophysiologically interpretable diagnostic classification schemes and individualized treatment strategies in psychiatry. |
| 09:50 - 10:35 |
Talk
H104 |
Guido Nolte
The Neuroimage debate on Granger Causality and Dynamical Causal Modeling. A critical review of critical reviews.
Recently, a debate emerged on the relative merits of two methods, Dynamic Causal Modeling (DCM) and Granger Causality Modeling (GCM), to infer causal relationships between different brain regions from fMRI data. DCM is a hypothesis driven approach which estimates parameters of a biophysically inspired causal model and compares different models according to model evidence. GCM, in contrast, makes less restrictive assumptions about the biophysical dynamics and bases the causality estimates on information theory exploiting the principle "the cause precedes the effect". In this tutorial I will shortly introduce both methods and discuss objections raised by participants of this debate. I will finally discuss limitations of the models due to low time resolution, model assumptions, and measurement noise. |
| 10:45 - 11:30 |
Talk
H104 |
Pål Westermark
The Neurons of the suprachiasmatic nuclei, circadian rhythms, and how to measure them
Most organisms have an endogenous circadian rhythm: if placed in an controlled environment with completely constant conditions, the organism will still operate with a ~24 hour daily rhythm. In mammals, the suprachiasmatic nuclei (SCN) of the hypothalamus function as a master pacemaker for the entire organism. Here, we will give an overview of the circadian clock of the SCN neurons, how to measure the pace of the core clock accurately, and how to quantitatively analyze the corresponding data. |
| 11:30 - 11:45 |
Coffee Break |
| 11:45 - 13:00 |
Discussions
H104 |
|
