Event
Ph.D. Research Proposal Exam: Sahar Khosravi
Friday, August 2, 2024
2:00 p.m.
AVW 1146 (ISR)
Maria Hoo
301 405 3681
mch@umd.edu
ANNOUNCEMENT: Ph.D. Research Proposal Exam
Name: Sahar Khosravi
Committee:
Professor Behtash Babadi (Chair)
Professor Shihab Shamma
Professor Jonathan Z. Simon
Date/Time: Friday, August 2, 2024 at 2:00 PM
Location: AVW 1146(ISR)
Title: Reliable Inference of Functional Connectivity Measures from Two-Photon Calcium Imaging Data
Abstract:
Recent advances in neural recording technologies, such as high-density electrodes and two-photon calcium imaging, enable the simultaneous acquisition of neural activity from large populations of neurons. This thesis investigates two major aspects of neural coding and connectivity in neuronal populations: modeling functional interactions in neuronal assemblies and estimating functional connectivity using advanced statistical methods.
In the first part of this dissertation, we focus on developing a framework to infer Granger causal (GC) links from two-photon imaging observations in neuronal populations. By explicitly modeling latent endogenous processes that govern spontaneous activity and exogenous effects of external stimuli, we develop a direct inference framework for identifying GC connectivity in neuronal ensembles. Our methodology integrates dynamic Bayesian network modeling, point process models, and multivariate autoregressive modeling. Through simulated and real data from two-photon calcium imaging, we demonstrate that our approach effectively identifies GC links, providing insights into the functional interactions underlying stimulus discrimination tasks.
The second part pertains to the analysis of functional connectivity in the primary auditory cortex (A1) during two-tone presentations. We utilize Gaussian processes with Zernike means to project correlation maps to a 2D space, capturing the functional connectivity patterns of neurons. This study examines the facilitative and suppressive interactions in L2/3 neurons, highlighting the spatial structure of signal and noise correlations during harmonic and non-harmonic two-tone stimuli. Our findings suggest that these spatially precise interactions are crucial for understanding the complex processing in A1.
Finally, future work will extend these methodologies to further explore the functional connectivity of neuronal networks in conjunction with optogenetic interventions. This includes applying the developed GC framework to new datasets, such as those involving more complex behavioral paradigms or different sensory modalities. We aim to investigate how temporal dynamics influence GC links and how these interactions change under various conditions. Furthermore, we will extend our methods to be applied to two-photon data recorded under optogenetic stimulation, with the goal of revealing behaviorally-relevant functional connections. By broadening the scope of our research to encompass a wider range of experimental conditions and neuronal populations, these advancements will contribute to a deeper understanding of how complex neural interactions underpin cognitive processes and sensory perception. Ultimately, this research has the potential to provide new insights into the fundamental principles of brain function and to inform the development of therapeutic strategies for neurological disorders.
