**1. Introduction**

Neck pain is a common neuromusculoskeletal symptom with a prevalence ranging from 22% to 70%, increasing with age and a ffecting most often women around 50 years old [1]. It is the fourth leading cause of years lived with disability in 188 countries during the period 1990–2013 [2]. Therefore, the correct identification of the source of neck pain is paramount. However, probably due to imperfect diagnosis, the majority of patients with neck pain are still nowadays called "non-specific" [3].

According to the Bayesian inference, a medical diagnosis indicates that one disorder (e.g., muscular, discogenic, lack of sensorimotor control deficits) more than another is probably the cause of a patient's symptoms, and thus, investigations are needed to reinforce or refute the hypothetical diagnosis [4]. In accordance with the literature [2,5–7], diagnoses and therapeutic interventions for neck pain should be informed using quantitative (strength and range of motion) and qualitative (sensorimotor appraisal) assessment of neck rotation. Quantitative devices have been reported to be superior to visual estimation to assess the cervical range of motion [8], the most popular method used by clinicians being goniometry [9]. Although very easy to use, goniometry has a margin of error of about 5◦ [9]. Moreover, maybe more important than movement amplitude [5,10], the evaluation of sensorimotor function has demonstrated its importance in developing a better understanding of the pathophysiological mechanisms associated with cervical pain [6] both in cases of specific neck pain such as traumatic neck pain [11], as well as for idiopathic neck pain [12]. Therefore, in an attempt to better define the clinical picture of patients by focusing on head movement [5,13,14] especially in axial rotation [15], clinicians show increased interest in quantitative devices that can accurately monitor movement.

Various noninvasive three-dimensional motion capture systems are used in the field of cervical research in order to evaluate kinematic variables going beyond simple range of motion such as speed, acceleration and deceleration using electrogoniometers [16], ultrasound waves [17], optical-based systems [18,19] and inertial sensors [20] and so on. Nevertheless, their dimension, complexity, and cost make such systems often difficult to use in clinical practice. The need for compact, user-friendly and low-cost measurement devices that can bring relevant information in everyday clinical practice is therefore obvious and goes beyond neck exploration, although we chose to focus on that topic in the present study.

Inertial measurement units sensors (IMUs) began to be applied to human movement before 2000 [21]. IMUs consist of accelerometers and gyroscopes which are organized in orthogonal triads in order to obtain three-dimensional kinematics [22]. Most often, IMUs are now supplemented by magnetometers and thermometers and are called MARG sensors (magnetic angular rate and gravity sensors). This technology has the advantage of not requiring external equipment such as cameras to acquire the orientation and position of the human segments, and it does not limit the subject's movement to the volume covered by the cameras. IMUs or MARGs thus seem to be the appropriate basic tool to design a device which could be easily used in a clinical and ecological environment [23,24]. Note that this technology suffers from high measurement noise and drift [25] that can mostly be cured by a Kalman filter [25]. Nowadays, the large-scale production MARGs sensors make them affordable compared to gold-standard systems but nevertheless the prices are several thousand euros (e.g., Vicon®, XSENS®). Other MARGs may indeed be bought at typical prices less than 50 €, still without any packaging nor user-friendly interface.

It is in this context that our team designed a small-sized, light, and ultra-low-cost inertial sensor called DYSKIMOT. After first laboratory tests, our goal was to evaluate the accuracy of DYSKIMOT compared to a gold-standard optoelectronic system when performing a clinical sensorimotor test developed by Hage et al. (i.e., the DidRen-Laser Test) in small groups of asymptomatic and symptomatic neck pain participants [26]. We selected different dynamic outcomes to evaluate our DYSKIMOT [27,28]: range of motion, peak speed, average speed, peak acceleration, and peak deceleration.

#### **2. Materials and Methods**
