Dataset of Methods for Lowering the Power Consumption of OS-Based Adaptive Deep Brain Stimulation Controllers

Published: 29-03-2021| Version 1 | DOI: 10.17632/wp2kcrymgg.1
Contributors:
Roberto Rodriguez-Zurrunero,
Alvaro Araujo,
Madeleine M. Lowery

Description

Power Consumption measurements and emulated LFP data when using an adaptive dual threshold deep brain stimulation algorithm on top of a YetiOS based controller. The stimulation amplitude varies depending on the beta band amplitude of the LFP signal. The power consumption was measured for an operating system implementing this controller for different sample rates, duty cycling, tick-less idle mode and read mode (single sample read, buffered read). The base power consumption of the OS based controller ("Single_Sample_1000Samples_s_Soft_Tickless.csv") is drastically reduced when using tick-less idle without main clocks, buffered read, reduced sample rate and duty cycling modes.

Files

Steps to reproduce

The power consumption data were measured on a YetiMote with an STM32L476RE microcontroller. The LFP data was generated by a mean field model based real time emulator (aDBS-EMU) of a parkinsonian basal ganglia. A dual threshold adaptive deep brain stimulation algorithm was running on top of the YetiOS operating system. Power consumption data were acquired with Keysight B2901A SMU with 2.5 ms of aperture and interval times, 10 mA measurement range and 3.3V source. Each test acquired 250000 samples (250 seconds).