ElectroCorticoGraphy (ECoG) acquisition exploiting signal characteristics for reduced power

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.