This paper discusses designing technology for brain-machine interfaces (BMI). Advances in BMI technology might potentially help improve devices treating neurological conditions. Designing a BMI system is a complex engineering problem drawing on such diverse fields as applied physics, circuit design, algorithms and biology. Fundamental to device design is the neurophysiology of the 'brain circuits' and how a circuit designer chooses to interface with these networks. To put these concepts in context, we describe recent work in prototyping bi-directional BMIs that seek to provide adaptive neuromodulation of a neural circuit based on detected electrical biomarkers in the brain. This application highlights how advancements in merging silicon and biological systems, reflected in BMI design, seeks to help advance the treatment of a variety of neurological disorders.