Abstract
The computational capacity of the brain (and artificial neural networks, which draw inspiration from brain function) is derived from the nonlinear transformations carried out in neural circuits. Such non-linearity is instantiated in the brain, not only by nonlinear transformation of neural “codes” across brain regions (which could be taken to correspond to layers in an artificial neural network), but also by similar dynamic (i.e. changing) transformations within single brain regions over time. Such temporal transformation of encoding in brain regions is allowed by recurrency of neural connections (circuits providing inputs to themselves) and generates dynamic activity patterns during neural processing for many tasks. Furthermore, such dynamics allow the same neurons to essentially behave in different ways over time, and hence perform different functions over time. Consequently, this means that neural circuits interact with each other in different ways over time. However, brain function is further influenced by (long-lasting) body states (e.g. sickness), context (e.g. new vs. familiar environments), and behavioral modes as attention vs. disengagement or awake vs. sleep. Therefore, “intrinsic” dynamics in the brain are occurring with a backdrop of changing external input, hence, to advance our understanding of neural processing we need to account for the intra- and inter-circuit interactions and dynamics that are embedded in a hierarchy of timescales.
Although the dynamics of neural activity is a rapidly evolving topic of research, the contribution of this aspect of neural processing is still being understood. The work in this thesis investigates how neural processing and communication between neural populations change with brain/body states, and over shorter periods of time during stimulus-evoked responses in the context of taste processing. Starting with a general introduction into neural dynamics, and inter-region communication, I go on to discuss how the same neural population responds differently depending on body state (sickness). From there, I move on to discuss the dynamics of multi-region communication during taste processing, ending with an exploration of changing neural “encoding” over time, and discussing how any instance of neural responses can be considered as being generated from hierarchical processes at multiple timescales.