Event-driven data and power management in high-density neural recording microsystems

We present design approaches and circuit techniques to perform data and power management in high-density neural recording implants. The proposed strategies employs low-power automatic detectors and dedicated neural signal processing circuits to take advantage on the discontinuous nature of neural signals presenting low duty cycles. We show that the suggested approaches enable for significant data and power reductions compared with standard design practices. Also, design methods to optimize low-power analog circuits are discussed.