Data movement between separate processing and memory units in traditional von Neumann computing systems is costly in terms of time and energy. The problem is aggravated by the recent explosive growth in data intensive applications related to artificial intelligence. In-memory computing has been proposed as an alternative approach where computational tasks can be performed directly in memory without shuttling back and forth between the processing and memory units. Memory is at the heart of in-memory computing. Technology scaling of mainstream memory technologies, such as static random-access memory (SRAM) and Dynamic random-access memory (DRAM), is increasingly constrained by fundamental technology limits. The recent research progress of various emerging nonvolatile memory (eNVM) device technologies, such as resistive random-access memory (RRAM), phase-change memory (PCM), conductive bridging random-access memory (CBRAM), ferroelectric random-access memory (FeRAM) and spin-transfer torque magnetoresistive random-access memory (STT-MRAM), have drawn tremendous attentions owing to its high speed, low cost, excellent scalability, enhanced storage density. Moreover, an eNVM based crossbar array can perform in-memory matrix vector multiplications in analog manner with high energy efficiency and provide potential opportunities for accelerating computation in various fields such as deep learning, scientific computing and computer vision. This dissertation presents research work on demonstrating a wide range of emerging memory device technologies (CBRAM, RRAM and STT-MRAM) for implementing neuro-inspired in-memory computing in several real-world applications using software and hardware co-design approach. Chapter 1 presents low energy subquantum CBRAM devices and a network pruning technique to reduce network-level energy consumption by hundreds to thousands fold. We showed low energy (10×-100× less than conventional memory technologies) and gradual switching characteristics of CBRAM as synaptic devices. We developed a network pruning algorithm that can be employed during spiking neural network (SNN) training to further reduce the energy by 10×. Using a 512 Kbit subquantum CBRAM array, we experimentally demonstrated high recognition accuracy on the MNIST dataset for digital implementation of unsupervised learning. Chapter 2 presents the details of SNN pruning algorithm that used in Chapter1. The pruning algorithms exploits the features of network weights and prune weights during the training based on neurons’ spiking characteristics, leading significant energy saving when implemented in eNVM based in-memory computing hardware. Chapter 3 presents a benchmarking analysis for the potential use of STT-MRAM in in-memory computing against SRAM at deeply scaled technology nodes (14nm and 7nm). A C++ based benchmarking platform is developed and uses LeNet-5, a popular convolutional neural network model (CNN). The platform maps STT-MRAM based in-memory computing architectures to LeNet-5 and can estimate inference accuracy, energy, latency, and area accurately for proposed architectures at different technology nodes compared against SRAM. Chapter 4 presents an adaptive quantization technique that compensates the accuracy loss due to limited conductance levels of PCM based synaptic devices and enables high-accuracy SNN unsupervised learning with low-precision PCM devices. The proposed adaptive quantization technique uses software and hardware co-design approach by designing software algorithms with consideration of real synaptic device characteristics and hardware limitations. Chapter 5 presents a real-world neural engineering application using in-memory computing. It presents an interface between eNVM based crossbar with neural electrodes to implement a real-time and high-energy efficient in-memory spike sorting system. A real-time hardware demonstration is performed using CuOx based eNVM crossbar to sort spike data in different brain regions recorded from multi-electrode arrays in animal experiments, which further extend the eNVM memory technologies for neural engineering applications. Chapter 6 presents a real-world deep learning application using in-memory computing. We demonstrated a direct integration of Ag-based conductive bridge random access memory (Ag-CBRAM) crossbar arrays with Mott-ReLU activation neurons for scalable, energy and area efficient hardware implementation of DNNs. Chapter 7 is the conclusion of this dissertation. The future directions of in-memory computing system based on eNVM technologies are discussed

The brain is the most complex part of human body. It is the source of intelligence and human behavior. Understanding how the brain works is essential for treatment of neurological disorders and development of more powerful artificial intelligence algorithms. To investigate its function, novel neurotechnologies are needed to record large-scale neural activity at high spatiotemporal resolution. Besides that, computational models could also be built to investigate possible mechanisms for neural activities at different levels. This dissertation presents research work on novel neurotechnologies and computational frameworks based on machine learning and biophysical modeling for investigation of brain functions. Chapter 1 presents a flexible insertable transparent microelectrode array (Neuro-FITM) for simultaneous electrical recordings from hippocampus and wide-field calcium imaging of dorsal cortex. We successfully recorded simultaneous activity of large-scale cortex and hippocampus. Using two-stage tensor component analysis and decoding analysis with support vector machines, we demonstrated distinct associations between hippocampus and cortex during hippocampal sharp-wave ripples. Chapter 2 presents a compact closed-loop optogenetics system based on transparent graphene microelectrode array. We extensively investigated light-induced artifacts for graphene electrodes and gold electrodes. We also validated the system for different frequencies of interest for neural recordings. The developed system can be used for various applications involving optogenetic stimulation and electrophysiological recordings. Chapter 3 presents the decoding of cortex-wide brain activity using surface electrical recordings over the cortex. We performed simultaneous local electrical recording and wide-field calcium imaging in awake mice. Using recurrent neural network, we demonstrated successful decoding of pixel-level cortical activity using locally recorded surface potentials. These results show that locally recorded surface potentials indeed contain rich information of large-scale neural activities. Chapter 4 presents a biophysical model for coupled network of hippocampal CA1 and prefrontal cortex to study mechanisms of memory trace transfer and reactivation. We found that stored memory traces in cortex can be retrieved through two different mechanisms, namely cell-specific inputs and non-specific spontaneous activities. Our study presents mechanistic explanations for memory transfer and retrieval. Chapter 5 is the conclusion of this dissertation. The future directions of neurotechnologies and computational models for neuroscience research are discussed

Deep learning based on neural networks emerged as a robust solution to various complex problems such as speech recognition and visual recognition. Deep learning relies on a great amount of iterative computation on a huge dataset. As we need to transfer a large amount of data and program between the CPU and the memory unit, the data transfer rate through a bus becomes a limiting factor for computing speed, which is known as Von Neumann bottleneck. Moreover, the data transfer between memory and computation spends a large amount of energy and cause significant delay. To overcome the limitation of Von Neumann bottleneck, neuromorphic computing with emerging nonvolatile memory (eNVM) devices has been proposed to perform iterative calculations in memory without transferring data to a processor. This dissertation presents energy efficient hardware implementation of neuromorphic computing applications using phase change memory (PCM), subquantum conductive bridge random access memory (CBRAM), Ag-based CBRAM, and CuOx-based resistive random access memory (RRAM). Although substantial progress has been made towards in-memory computing with synaptic devices, compact nanodevices implementing non-linear activation functions for efficient full-hardware implementation of deep neural networks is still missing. Since DNNs need to have a very large number of activations to achieve high accuracy, it is critical to develop energy and area efficient implementations of activation functions, which can be integrated on the periphery of the synaptic arrays. In this dissertation, we demonstrate a Mott activation neuron that implements the rectified linear unit function in the analogue domain. The integration of Mott activation neurons with a CBRAM crossbar array is also demonstrated in this dissertation.

The advancement of neuroscience research often requires recording of complex neural activities at high spatiotemporal resolution. Electrophysiology, being the backbone of neuroscience for decades, has the advantage of high temporal resolution, yet lacks the high spatial resolution of fluorescent imaging at single cell level. On the other hand, fluorescent imaging suffers from low temporal resolution due to the slow kinetics of the indicators. Recently, optogenetics revolutionized the capacity to control selective neural populations and provides researchers with unprecedented opportunities to investigate the causal relationships among different brain circuits. However, the traditional neural electrode arrays based on silicon and noble metals are opaque and hence not suitable to integrate electrophysiology and optical modalities. This dissertation presents a novel transparent microelectrode array based on graphene that demonstrates crosstalk-free integration of electrophysiology, calcium imaging, and optogenetics in in vivo experiments on mice models.Chapter 1 reviews the recent progress in the field of graphene-based neurotechnology. Graphene is widely used for microelectrode, field effect transistor, chemical sensing, and cell culture owing to its flexibility, transparency, high conductivity, low noise, and biocompatibility. Chapter 2 presents a novel transparent graphene microelectrode array designed for multimodal neural interfaces. The fabrication process was designed to avoid crack formation and organic residue, which is essential to eliminate light-induced artifacts. In vivo experiments were conducted to demonstrate a crosstalk-free integration of electrophysiology, optical imaging, and optogenetics for the first time. Chapter 3 demonstrates that electrochemical impedance of graphene is fundamentally limited by the quantum capacitance. And to overcome such limit, we created an alternative conduction path with electrochemically deposited platinum nanoparticles and reduced the impedance by 100-fold while maintaining high transparency. Chapter 4 presents a flexible implantable transparent microelectrode array that enables simultaneous electrical recordings from hippocampus during optical imaging of neural activity across large areas. Our neural probe has three advantages, flexibility, transparency, and shuttle-free implantation. We demonstrated seamless integration of simultaneous wide-field fluorescence imaging of the cortex with electrical recordings from the hippocampus. Chapter 5 is the conclusion of this dissertation. The outlook and roadmap of graphene-based neurotechnology for both neuroscience research and medical applications are discussed.