SpICENet - Spinal Input Connectome: Extracting the Functional Network
The purpose of the partnering project is to evaluate through simulation in the NRP the functionality of a new spinal cord model developed by DLR and Lund University in another project. To tis end, SpICENet will:
- implement this model in the NRP
- link the model of the spinal cord to models of sensory neurons as implemented by Courtine's group
- test if/how the model can produce meaningful movement;
- iteratively adjust the model according to the functional insights
An association with the HBP will grant us privileged access to platforms for the validation of our network models. This includes both the NRP and the myrorobotics systems in combination with SpiNNaker boards.
In addition, we would like to spread our results within the neuroscientific community by offering our Python-model as optional transfer function that interfaces the brain and robotics simulations in the NRP, as discussed previously with HBP members from FZI SP10.
Collaboration with HBP
The SC model from DLR will be integrated as a new component into the neurorobotics platform. While a spinal cord model has already been implemented in the platform by SPlO, the new one will provide users with an alternative that will enable comparative studies. Concretely, the ·model will be implemented as transfer function in Python code that interfaces the brain simulation and sensory input from and motor signals to the neurorobotic platforms. This code will be made available as a model on the NRP.
Within the HBP, two groups are predominantly responsible for the SC models, lead by Auke ljspeert and Gregoire Courtine. Possible future interactions (already discussed with these groups) may consist in:
- Having the DLR model receive input from primary sensory neurons as modelled by Courtine's group, to probe interoperability of the models.
- Having the DLR model act as a complete model of the SC or as modular input stage for functional CPG and reflex models as developed by ljspeert.
- Seeking indications about where and how supraspinal signals converge onto the spinal circuitry, thereby informing future projects both in DLR and the HBP.
While preliminary SC models are available within the HBP, the SC model from the previous collaboration between DLR and Lund University would be guaranteed by HBP to be fully integrated inside the NRP as an extremely bio-realistic SC model. It will thereby strongly support the research planned by SPlO members such as ljspeert and Courtine, as well as others in HBP.
The integration of the SC model into the NRP has already been discussed with the other groups responsible for SC models, as well as with both Florian Rohrbein and Fabrice Morin, the past and current SPlO managers, respectively. The SC models for the NRP are currently being developed and core members of our PP participate in the regular skype meetings and group discussions at the HBP events.
We will therefore continuously make reasonable efforts that the interfaces between sensory models implemented in the HBP and our SC model are compatible. In addition, we will adjust our Python code to suit the software requirements set by the models for motor output and requirements of the brain models as they are implemented in the NRP.
In addition, Philipp Stratmann, who participates in this PP as "external employee" of DLR, is supervising SPlO work in SGA2 (unrelated to SplCENet) carried out in Prof. Albu-Schaffer's group at TUM. He regularly takes part in formal events of the NRP, such as group meetings, seminars, or summer schools. This guarantees a strong linkage (both formal and informal) between the DLR and the HBP.
Time frame: 2019 to 2021
Origin: Spontaneous Application
Funding: Lund university, Deutsches Zentrum