Myself and a colleague are keen to explore the impact altering band pass filters (high and low) has on ubiquitous accelerometer output. Whilst other steps in the (pre)processing of accelerometer data, such as epoch and recording frequency, are relatively well explored, the impact of band pass filters applied is less well known. The application of a "high" recording frequency and therefore observance Nyquist-Shannon sampling Theorem (2fmax) is often cited as sufficient to capture all relevant human movement. Yet, the application of band pass filters is predominantly not considered, where logically, if a high band pass filter Is too low, irrespective of the recording frequency, biologically relevant movement will be missed and perhaps misinterpreted. Given the interest in even modest activity level improvement, and the pervasive nature of accelerometry; we therefore intend to explore this assertion. For those that have read this far, we would welcome correspondence regarding collaboration ([email protected] or [email protected]). Cain