**1. Introduction**

Individuals with mild traumatic brain injury (mTBI) suffering from impaired vestibular and ocular-motor impairments may be prescribed vestibular rehabilitation consisting of head and trunk movements during physical therapy. Vestibular rehabilitation typically includes gradual increases in the range of motion (ROM) and velocity of head movements and has shown promising improvements in the reduction of symptoms and greater overall function in mTBI patients [1–3]. While a physical therapist trains the patient in these exercises, a home exercise program is a vital part of vestibular rehabilitation therapy. Head and trunk movements can be compromised in this population, as the

perceived head position relative to the trunk is generally impaired in individuals with vestibular pathologies [4]. Consequently, avoidance behavior and maladaptive strategies, such as limiting the head ROM and rotational velocity, may be used in an effort to minimize symptoms [4]. These subtle impaired movements, such as head and trunk velocity, are often not detected visually [5], and have the potential to interfere with successful rehabilitation. Thus, the ability to quantify and track these movements both within the clinic and during a home exercise program may be highly beneficial.

Optical motion capture is commonly considered the gold-standard for tracking human movement. However, these motion capture systems are costly, can be limited by optical occlusion, and require a dedicated motion laboratory with specialists to collect, process and interpret the data. These factors make motion capture a suboptimal choice for clinical assessment in practice. Conversely, inertial measurement units (IMUs) provide a promising alternative due to the lower associated cost, simplified experimental set-up, and the ability to acquire data during everyday life, such as home monitoring [6].

IMUs have been shown to be a viable tool for measuring human motion across a wide range of activities and environments (for review across multiple domains see References [7–11]). While Duc and colleagues [12] have reported very good agreement with motion capture, and good retest reliability when measuring cervical ROM [12], it is unclear if IMUs can accurately estimate ROM and angular velocity during more complex movements. Additionally, the accuracy of IMU output can be subject to sensor placement [13], the developed algorithm, and how well it compensates for factors such as magnetic distortion and gyroscopic drift [14,15]. Further, it is important to validate an IMU algorithm across the range of measures of interest.

Our lab is currently evaluating the use of IMUs within a standard vestibular focused home-exercise program involving head and trunk movements (ROM and peak rotational velocity) during standing and walking in individuals with mTBI. Prior to conducting the larger intervention, we wanted to determine the validity of the sensors that will be used. Therefore, the aim of this study was to determine the validity of IMUs to estimate head and trunk ROM and peak rotational velocity during head movements made when standing and walking, in comparison with three-dimensional motion capture data.
