Abstracts

Rationale: Some patients with pharmacoresistant partial epilepsy are not candidate for surgery. There has been growing interest in neuro-responsive intracerebral local treatment of seizures such as focal drug delivery, focal cooling, or electrical stimulation. The latter requires an effective seizure-detection system and a brain stimulator. We present a low-power implantable integrated device for responsive electrical stimulation. Methods: The proposed implantable closed-loop neuro-stimulator (CLNS) combines an epileptic seizure detection (EPSD) with simultaneous electrical stimulation feedback. The EPSD provides continuous long-term monitoring of intracerebral EEG (iEEG). The sensitivity of EPSD is enhanced and several decision boundaries are introduced to reduce the number of false alarms for the patient-specific seizure pattern. The seizure-onset information is extracted through early modulation and proper rectification of the intracerebral signal. The EPSD determines the high-frequency patterns and the progressive amplitude increase of the seizure signal. The iEEG is analyzed over a certain time frame and a higher number of the detection indicates an upcoming seizure event. Results: The EPSD algorithm was validated through behavioral simulations in MATLAB. The EPSD was implemented and fabricated in a 0.18- m CMOS technology. The EPSD circuits fit in 2 mm X 1 mm chip area. Mixed-mode (analog/digital) tuning circuitry was used to enable adjusting the amplitude threshold and the time frame of the seizure-onset detection. Thus, this EPSD chip can be adapted to the patient specific seizure onset pattern. Furthermore, it can be tuned to be non-responsive to high-frequency brief electrical seizures if needed. This EPSD chip demonstrated accurate detection of seizure onsets, based on iEEG recordings from 8 patients with epilepsy. Furthermore, the influence of low-frequency noise was found to be negligible. Moreover, the total power dissipation was less than 6.80 W. The electrical stimulator has been highly miniaturized. The external controller provides energy and transmits data to the implanted stimulator by means of inductive coupling of spiral antennas. The control unit of the implanted stimulator is based on a commercially available Field Programmable Gate Array (FPGA) that present advantageous low-power and small-scale features. The CLNS is assembled on two circular printed circuit boards (PCB) of 2 cm diameter each, which are connected together with a flexible bus connector. The power consumption of the CLNS has showed that the system could run on a button cell battery for more than 8 years. Conclusions: The experimental results demonstrated the detection accuracy and the low-power dissipation of this implantable CLNS.