German - Japanese Collaborations
With the transnational funding initiative "German - Japanese Collaborations in Computational Neuroscience", the German Federal Ministry of Education and Research (BMBF), the German Research Foundation (DFG) and the Japan Science and Technology Agency (JST) support German-Japanese collaborative projects in Computational Neuroscience.
Within the scope of these collaborations, we were heading two reserach projects:
"Impact of top-down influence on visual processing during free viewing: multi-scale analysis of multi-area massively parallel recording of the visual pathway" led by Sonja Grün, Research Center Jülich, Tamura, Osaka University, Osaka, Japan and S. Shinomoto, Kyoto University, Kyoto, Japan
Primates perform frequent sequences of saccadic eye movements (duration about 30 msec) followed by fixations (duration about 200 msec) to sample visual information from their environment. During such a sequence, incoming visual information is processed, the next eye movement is prepared, and behavior is initiated. The information flow through the neuronal system causing these events is, however, unknown. The short time duration of the sequences suggests that incoming, sensory information and the preparation of the next eye movements are largely processed in parallel and tightly coordinated.This joint Japan-German project aims at getting a better understanding of the complex neuronal processing during free viewing, with a special emphasis on the interaction within and across multiple visual areas. The project requires on the one hand highly demanding experiments, which combine behavioral tasks and simultaneous electrophysiological recordings of massively parallel activity of multiple single neurons of the visual pathway. On the other hand it requires the analysis of higher-order correlations for identification of neuronal interactions within these complex multi-area multi-channel data.
"Neuronal network mechanisms of reinforcement learning", led by Abigail Morrison, Research Center Jülich, Kenji Morita, University of Tokyo, Japan and Mieko Morishima, National Institute for Physiological Sciences, Okazaki, Japan
The goal of this project is to understand the processes of reward learning in the cortico - basal ganglia loops on the single-cell and network level. To this end, neuronal network models will be developed based on known anatomic and physiological characteristics of the cortical-striatal circuit. In addition, model parameters will be specifically restricted by an accompanying program of experimental investigations. As the main result, we aim to describe the underlying dynamics of the development of dopaminergic activity response during stimulus / reward association tasks. It is expected that the results of this project will significantly contribute to the investigation of learning and decision-making processes in mammals. In addition, a broader understanding of the mechanisms of human learning is an advancement with great clinical relevance, because it is needed to better understand and treat dopamine-related diseases such as Parkinson's disease and Tourette's syndrome.In the first half of the funding period, local circuit models will be developed that realize sub-functions of the entire system. In addition, anatomical and physiological experiments will be carried out to specify the models as accurately as possible. In the second half, the local models will be integrated into a global model of the cortico - basal ganglia loop.
For more information, visit the German - Japanese collaborations website