ESR 5 : Theory of memory storage for sequences
Sequence activation, thought to be key for memory consolidation and future planning, and for linking fragmentary maps of extended environments, is likely to be generated in recurrent networks in the hippocampus that support attractor dynamics. However, understanding the encoding of temporally continuous memories and the ensuing neuronal dynamics requires a new theoretical framework. Statistical physics and computational neuroscience may provide the tools to develop methods to explore the ‘free-energy landscape’ of neural network models for hippocampal subfield CA3, storing both spatial and non-spatial information; this may offer new ways to study both the equilibrium states of the network and a number of dynamical properties. The model will be informed by relevant experimental data gathered by experimental partners in M-GATE.
This project is in collaboration with RU, INMED, UCL, NTNU.
Oleksandra Soldatkina was recruited as M-GATE fellow for this project.