A 16-ch Patient-Specific Seizure Onset and Termination Detection SoC with Machine-Learning and Voltage-Mode Transcranial Stimulation

📋 论文概要

提出了一款16通道非侵入性患者特异性癫痫发作起止检测SoC，集成通道复用模拟前端和脉冲电压经颅刺激，解决了现有癫痫检测SoC缺乏患者特异性、通道数少、无法检测发作终止且侵入性高的问题。

💡 主要创新点

• 实现16通道患者特异性癫痫发作起止检测，采用机器学习算法提高检测精度和适应性。
• 集成了电压模式经颅刺激功能，实现闭环治疗，且为非侵入式设计。
• 通过通道复用模拟前端和超低功耗设计，降低系统功耗和面积。

🏷 关键词

📄 原文摘要

Multichannel EEG seizure detection SoCs are widely used in medical practice and in research [1]-[3]. Due to huge variation in seizure patterns, patient-specific seizure detection is very crucial. [1], [2] presents 8-channel (ch) SoCs with moderate latency (~2s) but without seizure termination detection and stimulation. [3] implements a closed-loop SoC but is not patient-specific, and moreover, is invasive. This paper presents an ultra-low power 16-ch “noninvasive, patient-specific” seizure onset and termination detection SoC with channels multiplexing AFE and pulsating voltage transcranial electrical stimulation (PVTES). Figure 21.8.1 shows the implemented SoC. The 16-ch AFE is composed of 8 dual-channel charge recycled (DCCR) Chopper-Stabilized Capacitive-Coupled IA (CS-CCIA) in continuous time domain (40dB/1kHz), each channel followed by a discrete time-domain PGA (programmable gain 12/20/32/40dB and bandwidth

👥 作者与机构

Muhammad Awais Bin Altaf, Chen Zhang, Jerald Yoo

Masdar Institute of Science and Technology, Abu Dhabi, United Arab Emirates