Enabling Real-time Signal Processing Techniques for Intracortical Brain Stimulation

Neurological disorders remain a leading cause of disability worldwide, significantly impacting motor and cognitive function. Conventional rehabilitation strategies, such as physical therapy and pharmacological interventions, often provide only partial recovery, underscoring the urgent need for innovative neurotechnological approaches. Neuromodulation techniques, including deep brain stimulation, transcranial magnetic stimulation, and intracortical microstimulation, have demonstrated the ability to modulate neural circuits and promote plasticity. However, these methods are typically applied in an open-loop fashion, lacking the capacity to adapt in real-time to the dynamic nature of neural activity. The emergence of closed-loop neurostimulation, where stimulation is adjusted based on real-time neural signals processing, offers a promising avenue for more effective and personalized treatment strategies. This project underscores the importance of developing advanced neuroprosthetic systems that can effectively exploit both high-frequency and low-frequency brain signals to improve closed-loop system architecture design. The brain operates across a range of frequencies, with each band offering unique insights into neural states and functions. High-frequency content (300-3000 Hz range) provides precise, timely information on the excitability of single neurons, while low-frequency content (<100 Hz) captures population-level neuronal activity, reflecting shifts in cortical excitability over time. By integrating both types of signals, a neuroprosthetic system can deliver stimulation that aligns with the brain’s natural rhythms, enabling a more adaptive, state-dependent framework for neuromodulation and enhancing the effectiveness of motor rehabilitation strategies. This thesis explores novel approaches for advancing real-time neuromodulation techniques, with a specific focus on intracortical brain stimulation. The research is structured around three primary aims. 1) the design and implementation of a hardware-based real-time spike detection system utilizing model-based design approach to accelerate system development, 2) design of a real-time cortical state detection algorithm to deliver intracortical stimulation based on the ongoing cortical oscillations of the brain area of interest, and 3) the exploration of novel approaches for electroceutical by means of a hardware implementation of a spiking neural network in an open-loop fashion. Aim 1: Simulink®, Fixed-Point Designer, and HDL Coder were employed to streamline the transition from algorithm development to hardware implementation. This approach minimized design errors and accelerated real-time neural signal processing system design. The in vivo experimental validation of the developed system demonstrated high accuracy and efficiency. The spike detection algorithm was successfully translated into HDL code and then implemented in an FPGA-based system. Aim 2: A novel cortical state detection algorithm was also developed, tested, and validated in software before being translated into HDL and implemented on FPGA. The cortical state detection algorithm achieved over 91% accuracy in identifying brain states (specifically depolarized state) compared to state-of-the-art methods, without relying on invasive intracellular recordings. Aim 3: an open-loop neuromodulation system based on a spiking neural network, successfully increased post-simulation mean firing rate. These findings are consistent with previous studies that highlight the effectiveness of closed-loop stimulation in increasing neural firing rates. However, the observed effect was achieved through an open-loop fashion. Advancements in neuroprosthetic systems through the adoption of the model-base design approach and the design of a cortical state detection system pave the way for both accelerating the development of neuroengineering systems and enabling novel applications in motor rehabilitation. These contributions are instrumental in shaping the future of personalized neuromodulation, enhancing neuroprosthetic technology, and laying the groundwork for next-generation electroceutical therapies aimed at improving functional recovery in patients with neurological impairments.