The Bernstein Center Freiburg congratulates Sebastian Spreizer on his distinguished doctorate on “Dynamics of spatial network models”.
Computational role of space in the brain
The Bernstein Center Freiburg congratulates Sebastian Spreizer on his distinguished doctorate on “Dynamics of spatial network models”.
Neurons in the brain are distributed across space and their connectivity is constrained by their physical size and location in the brain. This aspect of the brain’s connectivity is often ignored. Therefore, we have a rather limited understanding of how the structure of the spatial connectivity shapes brain activity dynamics and, ultimately, brain function.
With his doctoral research, Sebastian Spreizer has provided two crucial results that advance our understanding of the brain's networks. The starting point of his doctoral research was two related experimental observations. First, it was observed that neurons located in a small neighborhood tend to activate together. These spatial neuronal assemblies encode animal behavior in the striatum, a network with only inhibitory interactions.
Sebastian built a network model of the striatum – a brain region composed of only inhibitory neurons. He wanted to wire the neurons in a distance-dependent manner. But experimental data was not available. So he studied two different, but biologically plausible, connectivity schemes. He demonstrated that in an inhibitory network like the striatum, if spatial neighbors become co-active, they must not connect to each other. That is, connectivity should be donut shaped – and striatal neurons should not inhibit their neighbors [see more in When in Striatum, Avoid Thy Neighbour, Link].
The second experimental observation that Sebastian explained in his research was that in many brain regions, irrespective of the task, neurons are activated in a temporal order, as the animal behavior unfolds. There are several proposals to explain the emergence of such temporal patterns, but all these models require some sort of learning. Sebastian showed that with simple network connectivity rules we can endow a network with an ability to generate spatial and temporal sequences. This connectivity rule requires that neurons project preferentially in a specific direction – which is plausible given their non-spherical shapes, and that neighboring neurons have similar preferential direction to connect. Such a network with spatially correlated anisotropic connectivity can generate a large repertoire of spatial and temporal sequences and provides a simple and biologically plausible explanation of a wide range of experimental activity. Moreover, these networks have properties that can be exploited for brain-inspired robotics applications. [The space-time fabric of brain networks, Link].
With these two studies Sebastian brings the connectivity structure of brain networks in the focus to understand brain dynamics and function.
Sebastian, during his doctoral studies realized that it is not easy for a young student to learn to model biological neuronal networks. So he devised the NEST Desktop which provides a web-based graphical user interface to build a variety of biological neuronal networks using the simulator NEST. This makes an excellent teaching aid for young students to explore the amazing world of brain networks [Nest-Desktop, Link].
Thus, in his doctoral research Sebastian has not only advanced our scientific understanding of brain networks, but has also improved the teaching aids for teaching computational neuroscience.
Contact
KTH Royal Institute of Technology
Prof. Dr. Arvind Kumar
Dept. of Computational Science and Technology
Lindstedtsvagen 5
Stockholm, Sweden
Tel: +46 (8) 790 62 24
E-mail: arvkumar@kth.se
Further reading
The space-time fabric of brain networks [Link]
Spreizer Sebastian, Aertsen Ad, Kumar Arvind (2019) From space to time: Spatial inhomogeneities lead to the emergence of spatiotemporal sequences in spiking neuronal networks. PLOS Computational Biology, 10(15):15(10):e1007432
When in Striatum, Avoid Thy Neighbour [Link]
Original publication:
Spreizer Sebastian, Angelhuber Martin, Bahuguna Jyotika, Aertsen Ad, Kumar Arvind (2017) Activity Dynamics, Signal Representation in a Striatal Network Model with Distance-Dependent Connectivity. eNeuro 4 (4) ENEURO.0348-16.2017; DOI: https://doi.org/10.1523/ENEURO.0348-16.2017
Nest-Desktop [Link]
Original publication:
Spreizer S, Senk J, Rotter S, Diesmann M, Weyers B. NEST Desktop-An educational application for neuroscience. eNeuro, 8 (6) ENEURO.0274-21.2021; DOI: https://doi.org/10.1523/ENEURO.0274-21.2021