Congenital muscular torticollis (CMT) is a frequently observed musculoskeletal condition occurring in infancy and has been reported to be the third most common orthopedic diagnosis in the pediatric population.1,2 The reported incidence of CMT in children ranges from 4% to 16%.3,4 Congenital muscular torticollis is the result of unilateral shortening of the sternocleidomastoid muscle, which causes limited cervical range of motion (ROM) in lateral flexion toward the uninvolved side and in rotation toward the involved side.5–8 As a result of these ROM limitations, children with CMT typically demonstrate a preferred head position that is laterally tilted toward the involved side and rotated toward the uninvolved side.5,6 Asymmetric head posture caused by these limitations is the most prominent clinical feature.8 Most children with CMT are successfully treated with conservative physical therapy, including a combination of manual stretching, active range of motion (AROM) exercises, and handling and positioning strategies.1,8–12
Cervical ROM in rotation and lateral flexion are key measures used to assess severity at initial diagnosis, predict outcome, and evaluate response to treatment interventions.1,8,10,12,13 Active cervical motion allows assessment of functional motion and muscle strength, especially for assessment of neck lateral flexor muscle function during the head-righting response.6,7 In a recent survey conducted by Fradette et al,8 77% of pediatric physical therapists (PTs) identified cervical ROM as one of the critical factors influencing their clinical decision making regarding assessment of severity of torticollis and intervention needs. Despite the identified importance of this clinical parameter, this study reported that the use of psychometrically robust outcome measures for cervical ROM was uncommon.8 A survey of PTs in New Zealand found that 95% to 100% of therapists assessed cervical ROM in the treatment of torticollis and 86% reported using visual estimation (VE) for this assessment.14 Respondents were aware of the low accuracy of VE, indicating that the lack of a suitable alternative measurement tool was the reason for their continued use of VE, especially when working with distressed infants. Youdas et al15 found VE to have overall poor reliability in the measurement of cervical AROM with an interrater reliability intraclass correlation coefficient (ICC) of 0.69 to 0.82 for rotation, and an ICC of 0.63 to 0.70 for lateral flexion. In a recent systematic review of studies measuring active and passive cervical ROM poor VE reliability, both for rotation (κ = 0.16-0.66) and lateral flexion (κ = 0.05-0.60) was reported.16 The authors also noted the absence of validity studies for VE, reinforcing the conclusion that VE is not a reliable enough measure for continued use as the primary assessment of cervical ROM.16
Although VE is unreliable, the selection of a suitable measurement tool for AROM in the infant population is challenging. Prior studies using goniometers and arthrodial protractors have only demonstrated reliability and validity for passive cervical motion in infants.7,17 Radiographic measurement is considered the reference standard for assessment of cervical ROM because it provides accurate identification of bony landmarks and joint angles crucial to assessment of muscle length.16,18 However, the risks associated with x-ray exposure make repeated use of radiographs impractical and unsafe as an assessment tool in children. Among other options, 3-dimensional (3D) video or electromagnetic motion analysis systems are complex and often costly, making them typically feasible only for the research setting.16 The identification of bony landmarks can be very difficult, particularly on a restless infant. Complicating potential use of goniometric measurement with a typical cervical range of motion (CROM) device is the concern that the device does not fit an infant's head size, a pediatric model is not available, and fit may be difficult because of associated plagiocephaly.
Pediatric PTs indicate that the qualities required in a desirable assessment tool include good reliability and validity, easy and rapid administration, simple equipment, and easy interpretation.8 A 2-dimensional (2D) video analysis (VA) system may present a feasible measurement tool for CROM in the pediatric clinical setting. Two-dimensional video motion analysis systems use computer software to measure the position of relevant body segments on the basis of reference markers attached to the skin of the subject. This noninvasive ROM measurement strategy may be particularly useful with infants and may reduce the risk of agitation during assessment. Relative to past 2D models and software, the continued development of software systems has improved ease of use, cost for acquisition (multiple options that are low cost or free), and more efficient analysis time, thus making 2D an attractive and reasonable option. High concurrent validity (Pearson r ≥ 0.95) as well as interrater and intrarater reliability (ICC ≥ 0.91) has been demonstrated with 2D systems used to measure lower extremity kinematics in the sagittal plane.19 Few studies, however, were found describing 2D analysis for CROM. Arbogast et al20 demonstrated criterion validity of 2D VA against a CROM device measuring active CROM in children aged 3 to 12 years and reported accuracy within 2° to 4°, but did not report reliability. Rahlin and Sarmiento6 evaluated still photography for measuring cervical lateral flexion in infants with CMT and reported good intrarater reliability (ICC = 0.79-0.84) and moderate to good interrater reliability (ICC = 0.72-0.99). Data from a pilot study performed at the University of Wisconsin–La Crosse measuring infant cervical motion using a 2D analysis system revealed an interrater reliability ICC of 0.97 (0.80-0.99) for measurement of angles marked by a pediatric PT, and an ICC of 0.89 (0.39-0.99) for PT students independently marking and measuring angles.21 To our knowledge, no literature has been published evaluating 2D VA for measurement of CROM in infants.
Ohman et al22 described the validity and reliability of the muscle function scale (MFS) for measuring muscle function in the lateral flexors of the neck in infants with CMT. However, this method is not commonly used clinically as demonstrated by the survey conducted by Luxford and colleagues who found that 86% of therapists used VE for assessment of cervical motion in CMT.14 The MSF assessment technique also relies on VE of the head position relative to a horizontal reference to assess muscle function/strength and assign a score on an ordinal scale of 0 to 5.8,22 The intent of the current study was to evaluate the infant's ability to produce an effective neck-righting response to a graded lateral trunk tilt from the vertical orientation. Therefore, the MSF was not deemed appropriate for use in this study. The purpose of this case series was to evaluate the clinical feasibility of using a 2D VA system compared with VE for measurement of active cervical rotation and lateral flexion in infants with CMT.
A small case series of 12 subjects was enrolled in this pilot study. All subjects were infants between 2 and 14 months of age referred to Gundersen Health System Pediatric Physical Therapy for evaluation and treatment of CMT. Ethical approval for this research project was obtained from the Gundersen Health System and University of Wisconsin-La Crosse Institutional Review Boards. Voluntary parental consent to participate in the study was obtained by 1 of the investigators at the time of the therapy visit, using an informed consent form approved by the local Institutional Review Board.
Two experienced pediatric PTs and a PT intern performed all data collection for this study. The PT intern set up video equipment and performed 2D VA of cervical motion for all subjects. Marker placement, infant positioning during data collection, and VE of cervical motion for all subjects were performed by 1 of the 2 experienced PTs. For each subject, video of cervical motion was captured simultaneously while VE of cervical motion was assessed by 1 of the 2 experienced PTs. All PTs remained blinded to corresponding VA results or VE for each subject until completion of the study.
Two digital video cameras (Sony Handycam model #HDR-CX160 and JVC Everio model GZ-MG730) were used to capture active cervical motion. Video in the frontal and transverse planes was recorded either simultaneously using both cameras, or sequentially using a single camera. An axis indicating the location of the frontal plane was marked on the floor of the gym and across the mat used for testing. A mirror was positioned parallel to the frontal plane axis at a distance of 10 ft for VE of lateral flexion, per the typical clinical technique used by the pediatric PTs. A hole 2 inches in diameter was cut in the mirror at a height of 23 inches from the floor, to allow video capture of lateral flexion perpendicular to the frontal plane at the height of a subject's neck (the fulcrum of cervical lateral flexion) when held by the PT who was sitting on the mat holding the subject (Figures 1A and B). A second camera to capture cervical rotation was positioned on a tripod directly over the frontal plane axis at a height of 3 ft above the subject's head when held by the therapist (Figure 1B). A laser pointer held along the longitudinal axis of each camera was used to locate a single position approximating the center of the video fields for both cameras. A reference mark indicating this position was made on the frontal plane axis for positioning of the therapist and subject during testing, and the positions of the mirror and tripod were marked on the floor of the gym.
All data for VA and VE for each subject were collected during a single therapy session. The locations of the anterior fontanelle, apex of the skull along the sagittal suture, tip of the nose, sternal angle, xiphisternal junction, and acromioclavicular (AC) joints were identified by palpation for placement of reflective markers. For locations of the anterior fontanelle and apex of the skull, the markers used were reflective balls approximately 1 cm in diameter attached with adhesive to elastic headbands. All remaining markers were flat, reflective stickers approximately 1 cm in diameter attached with adhesive to the skin of the subject, except for the marker on the nose that was approximately 0.5 cm in diameter. The time needed for palpation and marker placement was recorded for each subject.
Active CROM was measured in rotation and lateral flexion. For all measurements the PT holding the infant was in a seated position on a mat, at a point in alignment with horizontal and vertically oriented video cameras. To isolate the movement of the cervical spine from the thoracic spine, the supporting therapist placed both hands under the infant's arms with thumbs providing trunk support posteriorly and fingers providing support anteriorly (Figure 1B). This served to stabilize the thoracic spine and limit shoulder elevation.
The video cameras captured the motion at 30 frames per second in the frontal and horizontal planes during the therapy session. One camera was positioned behind the hole in the mirror 10 ft from the subject to record the motion in the frontal plane and the second camera was directly above the infant at a distance of 3 ft from the top of the infant's head. A frontal plane reference axis was marked on the floor of the examination area with colored tape. An additional mark on this axis indicated the center reference point with respect to the horizontally and vertically oriented cameras. The therapist supporting the infant used these references to keep the infant aligned with the cameras during testing.
For cervical rotation, the infant was supported in sitting or standing with shoulders level and in the frontal plane and presented with a visual stimulus of a toy to facilitate cervical rotation to the left or right. The PT supporting the infant performed VE of maximum rotation through direct observation of the infant.
For lateral flexion, the infant was supported in sitting, standing, or held in free space in the frontal plane while being tilted left or right to facilitate active lateral flexion (head righting) against gravity. The infant was presented with a visual stimulus of his or her reflection in a mirror directly in front of him or her at a distance of 10 ft or the PT performing the video capture held a toy directly above the top of the mirror to prevent rotation during active lateral flexion. The therapist supporting the infant used the reflection in the mirror to assess maximum lateral flexion by VE.
The PT supporting the infant recorded visual estimates of cervical motion on data sheets. These results were kept in a separate file and not shared with the PT intern performing the video analyses until after all video analyses were complete.
Two-dimensional video analyses were performed using the Dartfish Connect 6.0 software (Fribourg, Switzerland) to measure cervical rotation and lateral flexion for each infant. Video analyses were performed by a single therapist and included 3 steps: (1) importing video files from the camera into the Dartfish software, (2) visual determination of apparent maximum excursion points in the recorded video, and (3) use of the angle measuring capabilities of the software to measure the degrees of maximum cervical lateral flexion and rotation (Figure 2). These results were not shared with the other therapists until the completion of the study. Analysis time for each infant's data set was recorded, including the time to import video files into the software.
Outcome Measures and Statistical Analyses
A paired t test (α = 0.05) was performed using the SPSS software to assess the difference between VE and 2D VA for measures of right rotation, left rotation, right lateral flexion, and left lateral flexion. A t test was performed separately for infants assessed by each therapist performing VE, as well as a t test comparing results across all subjects.
In addition, other outcome measures used to evaluate clinical feasibility included time for marker placement, time for image analysis, and cost of hardware and software. Marker placement included the time spent to remove the child's clothes, to identify bony landmarks by palpation, and to place/position adhesive markers and elastic headbands. Image analysis included the time to import video files into the image analysis software, visually determine maximum excursion points, and perform angle measurement to determine maximum lateral flexion and rotation. Trends in time efficiency with experience were evaluated, and the mean time for marker placement and image analysis was calculated. Software costs and estimated hardware costs were reported to provide an additional important parameter for clinical practicality.
Table 1 shows the results of the paired t test comparing VE of cervical motion by observer 1, observer 2, and combined results from both observers, with the results of the VA performed by a third observer. For the combined results from both VE observers, a significant difference (2-tailed P < .05) between VE and VA was identified for left and right lateral flexion and for right rotation. A significant difference for right rotation and left lateral flexion was also identified independently for VE observer 1. A significant difference for right lateral flexion was also found for VE observer 2.
The average time required to set up equipment for data collection was 5.51 minutes (3.67-12.75 minutes). Setup time was initially 12.75 minutes, then quickly decreased on subsequent setups. On 2 occasions, equipment remained in place to collect data from more than 1 subject, so the setup values refer only to the number of times equipment was set up, and do not reflect the number of subjects.
The average marker placement time across all 12 subjects was 2.75 minutes (1.92-4.75 minutes). Marker placement time was 4.75 minutes for subject 1, then generally decreased across the remaining subjects. The marker placement time included the removal of clothing from the child as necessary for placement of adhesive markers on bony landmarks.
Greater variability was observed in the VA time. The average VA time was 23.96 minutes (18.32-32.52 minutes), but no clear trend in VA time was apparent.
Cervical ROM in both active and passive rotation and lateral flexion have been identified as key measures used to assess severity of CMT, predict outcome, and evaluate response to treatment.1,7–9,12,13,23 Pediatric PTs have identified CROM as 1 of the critical factors influencing their clinical decision making in determining intervention needs and plan of care, yet use of a robust outcome measure is uncommon.8,23 Fradette et al8 reported that a more accurate assessment of the severity of CMT may lead to the development of more appropriate intervention strategies and, ultimately, to better outcomes. Additional research to identify more reliable, valid, and clinically feasible methods of measurement of cervical AROM was recommended as an outcome of the American Physical Therapy Association's Section on Pediatrics Clinical Practice Guidelines for CMT.23 Our pilot study is a response to the need for a more psychometrically robust method for measuring active CROM in infants with torticollis.
In this study, the paired mean differences in active CROM measures between VE and VA shows that measurement of cervical motion by VE consistently resulted in higher values than VA (Table 1). Because the direction of the measurement differences were not reported by other authors comparing VE with other measurement techniques for CROM, we do not know whether this directional tendency is consistent with other reported results.15,16 On the basis of the documented poor reliability of VE15,16 and high validity and reliability of 2D VA systems for measurement of adult lower extremity motion,19 and good preliminary reliability for 2D analysis of infant cervical motion,21 we presumed that this difference may reflect greater accuracy of the VA assessment. Results of the paired t test across all 12 subjects (combined for observers 1 and 2) show a difference (P < .05) between VE and VA for left lateral flexion, right lateral flexion, and right rotation. The difference detected in right rotation and not left rotation may possibly be related to the position of the tripod for the overhead camera. The tripod was positioned to the right of the therapist supporting the infant, and may have created visual interference that affected VE of right rotation. Similar t test results are seen for observer 1, with a difference between VE and VA found for left lateral flexion and right rotation, and a difference approaching significance (P = .052) for right lateral flexion. Only a difference in left lateral flexion was significant between VE and VA for observer 2, but this may be due to the smaller number of subjects (n = 3) evaluated.
The time required to set up the equipment for video capture quickly decreased after the initial setup, demonstrating a rapid learning trend in efficiency with this process. With increasing numbers of repetitions, it would be reasonable to expect the average time for setup to gradually decrease and approach our minimum time of 3.67 minutes. If this type of VA was implemented in a clinical setting, further reduction in time for setup could be achieved with the use of a permanent ceiling-mounted bracket for the camera capturing cervical rotation. This would eliminate the time for setup and positioning of a tripod.
A similar trend in improvement of speed and efficiency is demonstrated in the time needed for marker placement. After the first subject, the time required for this task quickly decreased across most of the remaining subjects. The efficiency of this task is highly dependent on the tolerance of the child, as demonstrated by subject 9, who did not tolerate the headbands well and required several attempts to position these correctly. Considering this variability in response of the subject, it is uncertain whether the trend across repetitions would increase or decrease from the average time of 2.75 minutes recorded in this study. In fact, the headbands were the least tolerated markers for all subjects and a skull cap designed for infants is suggested as a possible way to improve tolerance and time efficiency in future trials.
Performance of the VA was, by far, the most time-intensive component of this 2D motion analysis system. The average time for VA was 23.96 minutes, with no apparent trend indicating a decrease in time across repetitions. Time for performance of separate analysis tasks was not recorded, but the greatest proportion of time was spent in measurement of angles with the software. The Dartfish software, while easy to use, required drawing axis lines and measuring the angle on each video frame representing a possible maximum excursion point. This process was typically repeated several times on each video, extending the analysis time. Variability in analysis time may be due, in part, to varying length of video files. During assessment of cervical rotation, markers on AC joints were sometimes obscured by the subject's head, requiring additional time scanning forward/backward in the video to identify approximate marker position. Initial use of 2 different camera types (Sony and JVC) required 2 different video import procedures, which added time to analysis for subject 1. Problems uploading the video files from the JVC camera into the Dartfish Connect 6.0 software were encountered, resulting in early loss of data for 3 subjects (omitted from study). Data for all subsequent subjects was collected using only the Sony camera and sequentially capturing rotation and lateral flexion motion. Clearly, VA time would need to be reduced markedly for this type of system to be feasible in the clinical setting. Other 2D motion analysis software is available that potentially supports the ability to automatically track assigned marker positions through the entire video and generate minimum or maximum angles at selected segments (MaxTRAQ 2D, Innovision Systems, Inc). Future evaluation of software with this function may dramatically reduce analysis time and improve clinical feasibility.
The primary expenses to consider for clinical implementation of a 2D VA system are the video cameras and the software. The Dartfish Connect 6.0 software (used on a trial basis for completion of this study) is available for purchase for $1000, or by annual subscription for $380. The Sony Handycam (model #HDR-CX160) used for this study belonged to one of the therapists, and is available for purchase for approximately $349 ($698 if 2 cameras are to be used simultaneously). In comparison, other expenses such as the tripod, mirror-with-hole, and a flash drive were nominal with a total cost of approximately $200. With a 2-camera system and purchased software, the total cost for clinical implementation would be approximately $1900. Using other software options might result in further cost savings. For example, the MaxTRAQ Lite software from Innovision Systems, Inc, essentially offers the same capabilities for this application as the Dartfish Connect 6.0 software, but at a lower cost of $99. However, the MaxTRAQ Lite software is only compatible with an AVI video file format and this specificity may require file conversion, thus requiring more time and negating some of the cost-benefit.
The variability of lateral flexion results in this study suggests a need for clarification in measurement technique regarding bony reference points. Because the sternocleidomastoid attaches to the sternum and clavicle, head position relative to both of these bony landmarks is crucial to accurately determine muscle length. Some of the variability in VE may have resulted from differences in attention to specific bony reference points used by each therapist.
Although the intent of this study was to determine a clinically feasible alternative to VE of infant cervical motion, it remains unclear whether 2D analysis is truly a more accurate measurement tool in this population as no comparison to a reference standard was conducted. Two other limitations were noted regarding marker position during the course of this study. Marker shift was observed in heavier infants with more mobile skin during positioning for active lateral flexion, especially with the marker at the xiphisternal junction (Figure 3A, identified by an arrow). The markers at the sternal angle and xiphisternal junction are positioned closely enough that a relatively small shift in position of the xiphisternal junction marker might reduce the measured angle by VA. This could potentially increase the difference between VE and VA. Visual estimate of left lateral flexion for the subject 1 was 50° (Figure 3A), and VA of left lateral flexion for subject 1 was 25.7° (Figure 3B).
The second marker limitation was observed in cervical rotation. Markers on the AC joints were sometimes obscured by the head of the infant at the point of maximum excursion. In Figure 4A, the right AC joint marker is just visible, but the left AC joint marker is obscured. This required frame-by-frame scanning backward and forward to identify the marker position on either side of the maximum excursion point to estimate the left AC joint marker position at the position of maximum excursion. This not only required additional analysis time, but possibly decreased the accuracy of the VA measurement.
On the basis of this pilot study of 12 subjects, a statistically significant difference between VE and 2D VA was identified for 3 of the 4 active cervical motions: left lateral flexion, right lateral flexion, and right rotation. Given the reported low reliability of VE when compared with VA, it is presumed that this difference reflects the greater accuracy of VA. Gains in measurement accuracy could support improved clinical assessment and outcomes. On the basis of this pilot study, the time required for VA was found to be too long to make the manual marking system clinically feasible. However, this time factor might be addressed through further exploration of the use of more automatic software analysis options. This study suggests that the use of VA may improve clinical measurement of active CROM; therefore, further studies are indicated to explore the utility of other software for this application and to examine the concurrent validity of 2D VA compared to 3D motion analysis.
Clinicians regularly assess both passive and active CROM in infants with CMT. In this study, only the assessment of active CROM was addressed and there is no intent to suggest that the use of passive ROM measures be eliminated as part of the VA assessment of active ROM in infants with CMT.
The authors thank Charlene Deters, PT, for her assistance with data collection.
1. Tessmer A, Mooney P, Pelland L. A developmental perspective on congenital muscular torticollis: a critical appraisal of the evidence. Pediatr Phys Ther. 2010;22(4):378–383.
2. Do TT. Congenital muscular torticollis: current concepts and review of treatment. Curr Opin Pediatr. 2006;18(1):26–29.
3. Peitsch WK, Keefer CH, LaBrie RA, Mulliken JB. Incidence of cranial asymmetry in healthy newborns. Pediatrics. 2002;110(6):e72.
4. Stellwagen L, Hubbard E, Chambers C, Jones KL. Torticollis, facial asymmetry and plagiocephaly in normal newborns. Arch Dis Child. 2008;93(10):827–831.
5. Rahlin M. TAMO therapy as a major component of physical therapy intervention for an infant
with congenital muscular torticollis: a case report. Pediatr Phys Ther. 2005;17(3):209–218.
6. Rahlin M, Sarmiento B. Reliability
of still photography measuring habitual head deviation from midline in infants with congenital muscular torticollis. Pediatr Phys Ther. 2010;22(4):399–406.
7. Ohman AM, Beckung ER. Reference values for range of motion
and muscle function of the neck in infants. Pediatr Phys Ther. 2008;20(1):53–58.
8. Fradette J, Gagnon I, Kennedy E, Snider L, Majnemer A. Clinical decision making regarding intervention needs of infants with torticollis. Pediatr Phys Ther. 2011;23(3):249–256.
9. Cheng JC, Wong MW, Tang SP, Chen TM, Shum SL, Wong EM. Clinical determinants of the outcome of manual stretching in the treatment of congenital muscular torticollis in infants. A prospective study of eight hundred and twenty-one cases. J Bone Joint Surg Am. 2001;83-A(5):679–687.
10. Ohman A, Mårdbrink EL, Stensby J, Beckung E. Evaluation of treatment strategies for muscle function in infants with congenital muscular torticollis. Physiother Theory Pract. 2011;27(7):463–470.
11. Demirbilek S, Atayurt HF. Congenital muscular torticollis and sternomastoid tumor: results of nonoperative treatment. J Pediatr Surg. 1999;34(4):549–551.
12. Cheng JC, Tang SP, Chen TM, Wong MW, Wong EM. The clinical presentation and outcome of treatment of congenital muscular torticollis in infants—a study of 1,086 cases. J Pediatr Surg. 2000;35(7):1091–1096.
13. Luther BL. Congenital muscular torticollis. Orthopaedic Nursing. 2002;21(3):21–28.
14. Luxford BK, Hale L, Piggot J. The physiotherapy management of infants with congenital muscular torticollis: a survey of current practice in New Zealand. N Z J Physiother. 2009;37(3):127–135.
15. Youdas JW, Carey JR, Garrett TR. Reliability
of measurements of cervical spine range of motion
—comparison of three methods. Phys Ther. 1991;71(2):98–104.
16. Williams MA, McCarthy CJ, Chorti A, Cooke MW, Gates S. A systematic review of reliability
and validity studies of methods for measuring active and passive cervical range of motion
. J Manipulative Physiol Ther. 2010;33(2):138–155.
17. Klackenberg EP, Elfving B, Haglund-Akerlind Y, Carlberg EB. Intra-rater reliability
in measuring range of motion
in infants with congenital muscular torticollis. Adv Physiother. 2005;7:84–91.
18. Wu SK, Lan HH, Kuo LC, Tsai SW, Chen CL, Su FC. The feasibility of a video-based motion analysis system in measuring the segmental movements between upper and lower cervical spine. Gait Posture. 2007;26(1):161–166.
19. Norris BS, Olson SL. Concurrent validity and reliability
of two-dimensional video analysis of hip and knee joint motion during mechanical lifting. Physiother Theory Pract. 2011;27(7):521–530.
20. Arbogast KB, Gholve PA, Friedman JE, Maltese MR, Tomasello MF, Dormans JP. Normal cervical spine range of motion
in children 3-12 years old. Spine. 2007;32(10):E309–E315.
21. Barske EM, Draeger D, Keck K, et al. Interrater Reliability
of 2-Dimensional (2D) Motion Analysis to Measure Infant
Cervical Range of Motion
. Wisconsin Dells, WI: Poster presented at: Wisconsin Physical Therapy Association; 2014.
22. Ohman AM, Nilsson S, Beckung ERE. Validity and reliability
of the muscle function scale, aimed to assess the lateral flexors of the neck in infants. Physiother Theory Pract. 2009;25(2):129–137.
23. Kaplan SL, Coulter C, Fetters L. Physical Therapy management of congenital muscular torticollis: an evidence-based clinical practice guideline. Pediatr Phys Ther. 2013;25:348–394.