Recent advances in machine learning demonstrated that mouse behavior appears to be composed of stereotyped, sub-second modules. Mouse body language is built from identifiable components and is organized in a predictable fashion (Wiltschko 2015). While high-resolution imaging is sufficient to reveal motifs of full-body motion, it may miss the tiny movements of the mouse head, which plays a crucial role in navigation and social interaction. To track head movements we have developed "head IMU" system to monitor linear and angular accelerations. In combination with the Kalman filter, it allows reconstructing mouse head movement with great precision.