Development of the circumduction metric for identification of cervical motion impairment

Abstract

Introduction: Chronic neck pain results in considerable personal, clinical, and societal burden. It consistently ranks among the top three pain-related reasons for seeking healthcare. Despite its prevalence, neck pain is difficult to both assess and treat. Quantitative approaches are required since diagnostic imaging techniques rarely provide information on movement-related neck pain, and most common clinical assessment tools are limited to single plane motion measurement.

Methods: In this study, the ability of an inertial measurement unit to document the cervical motion characteristics of 28 people with chronic neck pain and 23 healthy controls was assessed. A total of six circumduction metrics and one neck circumduction trajectory model were proposed as identification metrics.

Results: Five metrics demonstrated significant differences between the two groups. The neck circumduction trajectory model successfully distinguished between the two groups.

Discussion: The evaluation of the proposed metrics provides proof of concept that novel metrics can be captured with relative ease in the clinical setting using an inexpensive wearable sensor headband. The derivation of the proposed model may open new lines of inquiry into the clinical utility of assessing the multiplanar movement of cervical circumduction. The results obtained from this study also provide additional insight for the development of a sensitive, quantifiable and real-world neck evaluation strategies.

Keywords: Chronic neck pain; cervical circumduction evaluation; cervical circumduction metric; cervical motion impairment; cervical position; neck circumduction trajectory model.

Figure 1.

Elastic headband with one Shimmer IMU. The axis directions of the IMU are shown in the Left. The reference system for the head was defined as z+ as anterior, y+ as left and x- as superior. IMU: inertial measurement unit.

Figure 2.

Diagram of the neck circumduction movement and the neck circumduction model. 1 and 5 are neutral positions, 2 and 4 are flexion extremity point, 3 is extension extremity point. The neck circumduction movements started in the neutral position (1), followed by neck flexion (2), followed by an arc-shaped motion through lateral bending, extension (3), and flexion (4). Lastly, the head was returned to the neutral position (5).

Figure 3.

FFT of both healthy control group and pain neck group. The red, green, and yellow curves represent the power attenuation contour at 10, 5, and 3% of the peak power. The color bar indicates the signal’s power value.

Figure 4.

Diagram of the relative circumduction vectors magnitude during the full circumduction motion. The purple curve shows the shape of rCVM during a full circumduction motion, and the black arrows indicate the sequence of a circumduction motion. The green points represent the flexion extremity point, the blue point represents the extension extremity point. The beginning and end of the circumduction motion are denoted t₀ and t_n, respectively.

Figure 5.

Two exemplar neck circumduction movements from the neck pain and control groups. The starting points of both neck circumduction movements and the axes were aligned to assist with visual comparison. The green and purple dashed lines represent the major axis and minor axis of the neck circumduction movements for the control participant. The top left figure presents the neck circumduction movements in 3D configuration, the top right, bottom left, and bottom right figures present the neck circumduction trajectories projected onto 2D planes. The value of each data point is unitless and has meaning in relative terms only.

Figure 6.

Boxplot of the ROMs of both groups. Mann-Whitney U test showed significant difference in each ROM comparison. All differences between groups were significant at the p < 0.05 level.

Figure 7.

Metric 1: Cycle time. The top figures (Left: control; Right: Neck pain) show the gravitational accelerations in the three axes during a sample neck circumduction trajectory. The starting time is labeled as t₀ and the finishing time is labeled as t_n. The bottom figures show the distribution of the cycle time of the two groups, p < 0.05.

Figure 8.

Left: MCV difference between flexion and extension. Right: MCV difference between two lateral bending extremity points, significant differences were obtained for both metrics, p < 0.0001. MCV: magnitude of circumduction vector.

Figure 9.

Metric 3: the rMCV difference between vectors at the opposite extremity points. Top figures show two rMCVs from both groups. The red triangles represent the extension (middle) and flexion (left and right) extremity points. Bottom figures show the boxplots of the rMCV_ef, p < 0.0001 and rMCV_lr, p = 0.067. rMCV: relative magnitude of circumduction vector.

Figure 10.

Metric 4: NJP. Top figures show the jerk peaks, which are labeled as yellow triangles, within a neck circumduction movement. Bottom figures show the comparison between the NJP and NJI. For the same data sets, NJI failed to show a significant difference (p = 0.078). rMCV: relative magnitude of circumduction vector.

Figure 11.

Histogram of the NJI and square root of the NJI.

Figure 12.

Diagram of the neck circumduction trajectories from simulation and the participants. One healthy neck circumduction trajectory is shown in the top figure, one neck pain circumduction trajectory is on the bottom. The blue curves represent the data recorded from the trial, the red curves represent the outputs of the neck circumduction trajectory model with the best fit to the data sets. The unit is degree.

Figure 13.

Fit error magnitude, significant differences were obtained for both metrics, p = 0.007.

Figure 14.

OC curve of all diagnostic metrics. MCV: magnitude of circumduction vector; NJP: number of jerk peaks; rMCV: relative magnitude of circumduction vector.