模拟前端
电子工程
模式(计算机接口)
闭环
前端和后端
计算机科学
控制理论(社会学)
电气工程
工程类
控制工程
人工智能
CMOS芯片
操作系统
控制(管理)
作者
Xiaolong Li,Senlin Ren,Xiaowei Li,Tianyu Zhao,Xiaoqiao Deng,Wei Zheng
标识
DOI:10.1109/tbcas.2023.3333369
摘要
This article presents a local field potential (LFP)/action potential (AP) mode reconfigurable analog front-end (AFE) dedicated for the closed-loop vagus nerve stimulator (VNS). It combines an inverse electrical model of the intracranial electroencephalogram (iEEG) conducting in the brain tissues and been recorded at scalp as the extended electroencephalogram (EEEG). The AFE contains a LFP/AP mode reconfigurable EEEG preamplifier, a tunable integrator to compensate the effect of either the recording electrodes or head tissues, and an adder. The LFP/AP mode reconfigurable EEEG preamplifier consists of a tunable chopper-stabilized amplifier (CSA) and a 2nd-order tunable low pass filter (LPF). For better separation of LFP and AP signals, a high-order DC servo loop (DSL) characterized as a 2nd-order DSL in parallel with a 1st-order DSL is exploited in the tunable CSA to achieve a tunable high-pass frequency with a stopband attenuation slope (SAS) of +40 dB/dec. In addition, the tunable LPF can obtain a tunable low-pass frequency with a SAS of −40 dB/dec and provide additional 20 dB gain for AP signals. Fabricated in a SMIC 180 nm CMOS technology, and in the LFP band (0.5 Hz–200 Hz) and AP band (300 Hz–5 kHz), the measured mid-band gains of the LFP/AP mode reconfigurable EEEG preamplifier are 39.6 dB and 59.5 dB, the input-referred noises (IRNs) are 2.2 μVrms and 6.3 μVrms, the DC/in-band input impedances are 1.27/1.26 GΩ and 0.3/0.22 GΩ, respectively. The power consumption per channel AFE is 6.3 μW, and the die area is 1.4 mm $ \times$ 0.25 mm.
科研通智能强力驱动
Strongly Powered by AbleSci AI