The workshop is organized by the CNS group at the Department of Biophysics of the KFKI Research Institute for
Particle and Nuclear Physics.
Abstract: I will discuss an approach to causal inference that is based on the concept of time-lagged conditional mutual information, I(xp(t-s),yf(t)|zp1(t1)...zpn(tn)), i.e, the mutual information of the "past-oriented state" of process x at time t-s and the "future-oriented state" of process y at time t, taking into account a list of "non-causal confounds" z. This will be illustrated using EEG-derived time series associated with macroscopic brain regions of interest.
Abstract: An ongoing modelling study that uses the biologically realistic, detailed, compartmental technique will be presented. The purpose of this research is to explain the generation of the so called theta oscillation, a low frequency (5-12 Hz) population oscillation in the CA1 region of the hippocampus of the rat. The model consits of three neuron populations: pyramidal cells, olveus-lacunosum moleculare interneurons, basket interneurons, described by Hodgkin-Huxley-like currents. Besides generating theta-frequency population oscillation, the model shows similar firing-phase-preference distributions of neuron types to in vivo measured ones. It has been shown that a hyperpolarization activated cation (H) current and NMDA receptors are required for robust oscillation generation. The model suggests a possible way of low-frequency oscillation generation in a tri-synaptic, essentially feed-forward neural circle and explains the stabilizing role of the H current. It also explains why preferred firing phase distributions have the experimentally measured forms.
Abstract: Cross-modal integration, similarly to other functions of the cerebral cortex, is based on parallel distributed processing across the cortical areas. However, the elements of such a cross-modal network and their relationship have not been exactly defined. We analyzed the connectional pattern and topographical arrangement of visual and somatosensory cortices of primates to identify areas involved in visual-tactile integration and to characterize the architecture of the network of these areas using graph theoretical methods. Visuo-tactile network is characterized by a high incidence of nonreciprocal connections (82.4%). The network is densely connected providing that all areas are separated by short distances. Areas 7b/MT/7a/LIP/46/FEF occupy a central position in the network. This was confirmed by multidimensional scaling in the whole set of visual- and somatomotor areas. MT was the most frequently occurring area in cliques defined by mutual connections between the areas. Virtual lesions further supported the central role of MT in the formation of cliques. The shortest paths from early somatosensory cortices to early visual areas V1/V2 were through MT from areas 3a/1/2. These results suggest a central role for MT in visuo-tactile integration and in cross modal compensational plasticity.
Abstract: Validity of the monopole and other point source approximations were tested on neocortical spikes. Spikes were measured by extracellular linear multi microelectrode, in vivo feline primary auditory cortex. Instead of point source models, a new counter current model of the current source density distribution on a spiking cell was set up. Circumstances of its validity were determined. Using the counter current model, a new method for cell-electrode distance estimation was presented. Precision of the distance estimation was examined on simulated data. By applying the new method, two dimensional localization of the spiking cells arround the electrode was performed.