IPAB Workshop-16/07/2020
Title: Spatio-temporal Memory for Navigation in a Mushroom Body Model
Abstract: Insects, despite relatively small brains, can perform complexnavigation tasks such as memorising a visual route. The exact formatof visual memory encoded by neural systems during route learning andfollowing is still unclear. Here we propose that interconnections betweenKenyon cells in the Mushroom Body (MB) could encode spatio-temporalmemory of visual motion experienced when moving along a route. In ourimplementation, visual motion is sensed using an event-based cameramounted on a robot, and learned by a biologically constrained spikingneural network model, based on simplified MB architecture and usingmodified leaky integrate-andfire neurons. In contrast to previous image-matching models where all memories are stored in parallel, the continu-ous visual flow is inherently sequential. Our results show that the modelcan distinguish learned from unlearned route segments, with some toler-ance to internal and external noise, including small displacements. Theneural response can also explain observed behaviour taken to supportsequential memory in ant experiments. However, obtaining comparablerobustness to insect navigation might require the addition of biomimeticpre-processing of the input stream, and determination of the appropriatemotor strategy to exploit the memory output.
Jul 16 2020
-
IPAB Workshop-16/07/2020
Le Zhu
https://eu.bbcollab.com/guest/097fcdf5f281425c8e6d3679a9d021bc
" />
Blackboard Collaborate https://eu.bbcollab.com/guest/097fcdf5f281425c8e6d3679a9d021bc