The Woodruff School

SUMMARY

Understanding how neurons of the brain communicate, connect, and respond to stimuli is a fundamental goal of neuroscience. Whole-cell patch clamp recording in vitro represents the gold standard method for measuring electrophysiology of single neurons because of its high spatiotemporal resolution. However, the manual and time-consuming nature of patch clamping experiments has limited the throughput and number of cells that can be sampled per day. To overcome these limitations, this dissertation aimed to (1) integrate automated patch clamp with discovery experiments for cellular indicators and effectors, (2) develop a machine learning algorithm for real-time neuron detection of neurons in brain slices for in vitro patch clamping, and (3) create a coordinated, multi-pipette patch clamp algorithm for enabling high throughput synaptic connectivity studies. Towards these aims, this thesis demonstrated the first robotic system to perform ligand-gated ionotropic receptor protocols autonomously leading up to a 10-fold reduction in research effort over the duration of the experiment. In addition, a fully automated patch clamp robot was deployed to discover a brighter and more sensitive chemigenetic voltage indicator, Voltron2, over its predecessor exhibiting 3-fold higher sensitivity in response to sub-threshold membrane potential changes. Towards the second aim, a novel, deep learning-based method was developed to accomplish automated, real-time neuron detection in brain slice with high accuracy, achieving an F1 score of 80%. To facilitate efficient probing of local synaptic connections between neurons, the first ever forward-thinking multipatching robot demonstrated automatic, sequential recordings in a brain slice using a coordinated route plan. With these technologies combined, this thesis enabled the first robot that can automatically search for connected neurons in brain tissue and also outperforms manual patch clamping-based screening assays to significantly advance the field of neuroscience and reveal new insights into brain function.

SUBJECT:	Ph.D. Dissertation Defense

BY:	Mighten Yip

TIME:	Friday, April 14, 2023, 9:00 a.m.

PLACE:	IBB Building, 1128

TITLE:	Towards automation of multimodal cellular electrophysiology

COMMITTEE:	Dr. Craig Forest, Chair (ME) Dr. Ming-fai Fong (BME) Dr. Brandon Dixon (ME) Dr. Christopher Valenta (GTRI) Dr. Matt Rowan (Emory) Stephen Traynelis (Emory)

SUMMARY Understanding how neurons of the brain communicate, connect, and respond to stimuli is a fundamental goal of neuroscience. Whole-cell patch clamp recording in vitro represents the gold standard method for measuring electrophysiology of single neurons because of its high spatiotemporal resolution. However, the manual and time-consuming nature of patch clamping experiments has limited the throughput and number of cells that can be sampled per day. To overcome these limitations, this dissertation aimed to (1) integrate automated patch clamp with discovery experiments for cellular indicators and effectors, (2) develop a machine learning algorithm for real-time neuron detection of neurons in brain slices for in vitro patch clamping, and (3) create a coordinated, multi-pipette patch clamp algorithm for enabling high throughput synaptic connectivity studies. Towards these aims, this thesis demonstrated the first robotic system to perform ligand-gated ionotropic receptor protocols autonomously leading up to a 10-fold reduction in research effort over the duration of the experiment. In addition, a fully automated patch clamp robot was deployed to discover a brighter and more sensitive chemigenetic voltage indicator, Voltron2, over its predecessor exhibiting 3-fold higher sensitivity in response to sub-threshold membrane potential changes. Towards the second aim, a novel, deep learning-based method was developed to accomplish automated, real-time neuron detection in brain slice with high accuracy, achieving an F1 score of 80%. To facilitate efficient probing of local synaptic connections between neurons, the first ever forward-thinking multipatching robot demonstrated automatic, sequential recordings in a brain slice using a coordinated route plan. With these technologies combined, this thesis enabled the first robot that can automatically search for connected neurons in brain tissue and also outperforms manual patch clamping-based screening assays to significantly advance the field of neuroscience and reveal new insights into brain function.