An integrated sub-scalp EEG sensor for diagnosis in epilepsy

There is a growing demand for low-noise and low-power wireless neurosensors for emerging clinical indications such as epilepsy, where long-term, neurological health must be monitored to detect seizures. In this paper, we present an integrated recorder chip that is part of a wireless system to record and process sub-scalp electroencephalography (EEG) data for seizure detection and treatment. The recorder is designed based on a novel frequency-shaping architecture that offers competitive circuit specifications on power, noise, precision, and chronic stability. The design is fabricated in a 0.13μm CMOS process. When measured at a 2.5kHz clock and 1.0V supply, the recorder can be switched between two bandwidths, where the lower bandwidth (50Hz) is for recording regular EEG data and the higher bandwidth (500Hz) is for recording high frequency oscillations [1]. The recorder is designed with a wide data dynamic range in order to avoid saturation from 50mV ambulatory motion artifacts when the gain of the active filter is set to 2V/V. The total power consumption is under 2.5μW/ch and the input-referred noise is 0.8μV from 150Hz and 1.4μV from 1-500Hz. A proof-of-concept prototype has been also developed to demonstrate that the recorder is insensitive to ambient interferences and movement artifacts in experiments.

[1]  C. Camfield,et al.  Stigma of Epilepsy , 2008, Canadian Journal of Neurological Sciences / Journal Canadien des Sciences Neurologiques.

[2]  Long Yan,et al.  A two-electrode 2.88nJ/conversion biopotential acquisition system for portable healthcare device , 2008, 2008 IEEE Asian Solid-State Circuits Conference.

[3]  Fernando Lopes da Silva,et al.  High-frequency oscillations – Where we are and where we need to go , 2012, Progress in Neurobiology.

[4]  Chen Zhang,et al.  21.8 A 16-ch patient-specific seizure onset and termination detection SoC with machine-learning and voltage-mode transcranial stimulation , 2015, 2015 IEEE International Solid-State Circuits Conference - (ISSCC) Digest of Technical Papers.

[5]  Jian Xu,et al.  A 50 μW/Ch artifacts-insensitive neural recorder using frequency-shaping technique , 2013, Proceedings of the IEEE 2013 Custom Integrated Circuits Conference.

[6]  Jian Xu,et al.  A New System Architecture for Future Long-Term High-Density Neural Recording , 2013, IEEE Transactions on Circuits and Systems II: Express Briefs.

[7]  Naveen Verma,et al.  A Micro-Power EEG Acquisition SoC With Integrated Feature Extraction Processor for a Chronic Seizure Detection System , 2010, IEEE Journal of Solid-State Circuits.

[8]  Refet Firat Yazicioglu,et al.  A 30 $\mu$ W Analog Signal Processor ASIC for Portable Biopotential Signal Monitoring , 2011, IEEE Journal of Solid-State Circuits.