2022 Theses Doctoral
Sensory Processing and Associative Learning in Connectome-Based Neural Circuits
There has been a significant increase in the amount of connectomics data available at the level of single neurons and single synapses in the last few years. This increase enabled investigations into the structure and function of neural circuits in much greater detail than ever before. Thus, the next step in our quest to understand the brain's functional logic is the development of tools and methods to enable us to extract data from and model these new connectomics datasets, and their use to start to examine the brain computationally. Specifically, for Drosophila melanogaster, the fruit fly, a large amount of data on the connectome have become available in the last few years. In this dissertation, we start by introducing the tools we have built to extract information from the Drosophila connectome and to create spiking models of neuropils using this information to model sensory processing and associative learning circuits at single-synapse scale. We then use the toolkit we have introduced to explore sensory processing and associative learning in the brain.
First, we introduce FlyBrainLab, an interactive computing environment designed to accelerate the discovery of functional logic of the Drosophila brain. Then, we propose a programmable ontology that expands the scope of the current Drosophila brain anatomy ontologies to encompass the functional logic of the fly brain, providing a language not only for modeling circuit motifs but also for programmatically exploring their functional logic; we introduce the FeedbackCircuits library for exploring the functional logic of the massive number of feedback loops (motifs) in the fruit fly brain, and NeuroNLP++, an application that supports free-form English queries for constructing functional brain circuits fully anchored on the available connectome/synaptome datasets. Thirdly, following up on the second, we explore the construction of antennal lobe circuits using models of glomeruli. We explore the composability of the connectivity of glomeruli with local neuron feedback loops, and quantitatively characterize the I/O of the AL as a function of feedback loop motifs in the one-glomerulus, two-glomerulus and 23-glomerulus scenarios. Lastly, in the final chapter, we consider the modeling of the mushroom body, a second order olfactory neuropil and a center of associative learning, to demonstrate how the architecture of the circuit interacts with the circuit mechanisms by which sensory inputs are represented and memories are updated.
Thus, in this dissertation we introduce an approach for the analysis and modeling of neural circuits based on connectomics data, and apply this approach to neural circuits spanning multiple neuropils to extract and analyze the principles of computation in the brain. The methodology described here is designed to be applied to different sensory systems and organisms to infer the functional logic of connectome-based neural circuits.
- Turkcan_columbia_0054D_17528.pdf application/pdf 1.31 MB Download File
More About This Work
- Academic Units
- Electrical Engineering
- Thesis Advisors
- Lazar, Aurel A.
- Ph.D., Columbia University
- Published Here
- October 12, 2022