研究目的
To design a real-time computing system for processing electrophysiology data to identify relationships among living cells from their activity signals, with low latency and high flexibility.
研究成果
The system successfully processes electrophysiology data in real-time with low latency, identifying cell relationships through spike detection and correlation. It demonstrates the maturity of HLS methodologies for generating flexible processing systems, though scalability is constrained by FPGA memory resources. Future work should focus on optimizing memory usage and extending to more complex cell experiments with higher channel counts.
研究不足
The system's memory usage is a bottleneck, limiting the number of channels to 512 on a Nexys-4 FPGA board; exceeding this requires a mid-range FPGA device. Performance depends on the quality of SystemC models and user-defined annotations.
1：Experimental Design and Method Selection:
The system uses a digital signal processing chain with spike detection and inter-channel correlation modules, implemented using High-Level Synthesis (HLS) from SystemC models for flexibility and genericity.
2：Sample Selection and Data Sources:
Data is acquired from electrophysiology cultures using Micro-Electrode Arrays (MEA) with 64 analog signals sampled at 10kHz, but can be adapted to other numbers of electrodes.
3：List of Experimental Equipment and Materials:
Includes FPGA (ARTIX-7 XC7A100T-1CSG324C), analog-to-digital converters, MEA setups, and amplification systems.
4：Experimental Procedures and Operational Workflow:
Signals are filtered and processed in real-time using high-pass filters, stationary wavelet transform (SWT) with Haar wavelet for spike detection, and spike timing-dependent plasticity (STDP) for correlation analysis. Data transfer between modules is handled by FIFOs.
5：Data Analysis Methods:
Functional validation on FPGA with real or synthetic data, resource usage analysis (LUTs, FFs, DSPs, BRAMs), and performance evaluation based on number of channels and event rates.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
