Enabling Model-Based Design for Real-Time Spike Detection.

Area of Science:

• Neuroengineering
• Computational Neuroscience
• Embedded Systems

Background:

• Developing real-time neural signal processing systems is complex due to intricate low-level coding challenges.
• Hardware advancements in neurotechnology often outpace the ease of software development, documentation, and maintenance.

Purpose of the Study:

• To simplify the creation of neuroengineering systems for real-time neural signal processing.
• To reduce the technical barriers in developing and maintaining neurotechnology software.
• To accelerate the prototyping of diverse neuroengineering system architectures.

Main Methods:

• Utilized Model-Based Design (MBD) to streamline the workflow from modeling to implementation.
• Implemented a spike detection algorithm on a Field-Programmable Gate Array (FPGA) system.
• Performed all data handling and processing within the Simulink environment, generating hardware description language (HDL) code for the FPGA.

Main Results:

• Achieved real-time processing with latency <= 100.37 µs, matching offline spike detection performance.
• Validated the Model-Based Design approach through in vivo experiments in six animals.
• Demonstrated the efficiency and accuracy of the MBD-based spike detection system.

Conclusions:

• The Model-Based Design methodology significantly impacts neuroengineering by accelerating system prototyping.
• This approach simplifies the development of complex neuroengineering systems.
• All project code files are open-source, promoting collaborative development in the field.