Scalable Nanofabricated Neural Interfaces for Retinal Signal Decoding and Electrochemical Sensing

Abstract

Array electrophysiology is a useful approach for studying neural systems at scale, enabling simultaneous recording from large populations of neurons with high spatial and temporal resolution. Realizing this capability requires electrode platforms that are both scalable and reliable, while maintaining high-quality signal transduction at the electrode–electrolyte interface. This dissertation focuses on the development of such nanoelectrode array systems and their application to neural recording and sensing. Chapter 1 provides an overview of microelectrode array technologies and outlines the motivation for developing high-density neural interfaces. It introduces the challenges associated with large-scale neural recording, including electrode design, signal fidelity, and the need for advanced computational methods to interpret complex neural data. The chapter also frames retinal decoding and multimodal sensing as key application domains. Chapter 2 presents the system-level design and implementation of the nanoelectrode array platform. This includes the architecture of the recording system, integration with CMOS electronics, and considerations for achieving large-scale parallel recording with stable performance. Chapter 3 focuses on the nanofabrication processes used to realize the electrode arrays. Techniques such as photolithography, electron beam lithography, thin-film deposition, and etching are discussed in detail, along with surface modification methods to optimize electrode impedance and interface properties. Special attention is given to achieving reliable fabrication of three-dimensional electrode structures and ensuring compatibility with biological environments. Chapter 4 addresses the computational modeling of neural data, with an emphasis on decoding visual stimuli from retinal recordings. Machine learning models, including convolutional and recurrent neural networks, are developed to reconstruct stimulus information from population spike activity. The results highlight the importance of capturing nonlinear and temporal dynamics in neural responses, particularly in the mouse retina. Chapter 5 introduces a dual-mode sensing technique that combines electrochemical and electrophysiological measurements on the same chip. By integrating neurotransmitter sensing with electrical recording, the platform enables simultaneous measurement of dopamine concentration and neural activity. This approach provides new insights into the relationship between chemical signaling and electrophysiology, offering a pathway toward studying disease models and complex neural dynamics in a more comprehensive manner. Together, this work establishes a scalable nanoelectrode platform that bridges advanced nanofabrication, large-scale neural recording, and multimodal sensing, contributing to both engineering development and neuroscience applications.