ElectroCorticoGraphy (ECoG) acquisition exploiting signal characteristics for reduced power

Abstract

We present two analog front-end circuits for acquiring ElectroCorticoGraphic (ECoG) signals from the brain for Brain-Computer Interfaces (BCIs). They exploit a priori knowledge of ECoG characteristics to reduce resolution requirements for digitization and transmission and thus power. The first saves 2 bits with a power overhead of 127 nW; the second theoretically saves up to 8.3 bits with overhead of 0.85 µW. Both are in 130 nm CMOS using a 1.2 V supply.

Exploiting Electrocorticographic Spectral Characteristics for Optimized Signal Chain Design: A 1.08 W Analog Front End With Reduced ADC Resolution Requirements<formula formulatype="inline"> <tex Notation="TeX"/> </formula>

“…These bands reflect anatomical and functional characteristics of cortical and subcortical networks and their interactions, and they are correlated with a variety of cognitive processes and behaviors. Starting with the lowest frequency, researchers have distinguished the following bands: (1-4 Hz), (4)(5)(6)(7)(8), (8)(9)(10)(11)(12)(13), (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), , and high-( 60 Hz). These rhythms have been tied to cortical activity during language production and understanding [17], [18], vision [19], movement [20], and movement planning [21], among others.…”

“…The idea of preemphasizing parts of the ECoG spectrum is not new. In [30] the authors create a pre-emphasis at highsignal frequencies. This technique has a particular corner frequency that must be precisely calibrated for proper reconstruction.…”