Where's your shoulder?

A study of shoulder proprioception and an experiment in scientific communication.

Background

(Explained using only the ten hundred most used words.)

Proprioception has been described as a sixth sense, formed by the integration of sensory information about joint position, muscle tension, and the orientation of the body in space. The three distinct elements of proprioception are: joint position sense, awareness of force application and kinaesthesia; these are neural impulses which originate in mechanoreceptors within joint capsules, ligaments, muscles, tendons and skin. Proprioceptive pathways combine in the central nervous system, undergoing integration with information from other branches of the nervous system. Proprioception is vital for co-ordinated movement patterns and functional stability throughout the human body.

Proprioception refers to the body's ability to sense joint position and movement. The three aspects of proprioception are senses of: position, force and movement; these are electrical messages sent from within joints, ligaments, muscles, tendons and skin. Messages come together in the central nervous system, combining with information from other parts of the body. Proprioception is vital for co-ordinated movement patterns and stability.

Body position knowing is the sense that people have of where their body parts are. The three parts of this are senses of: position, force and moving; these senses come from within points of the body that move, tight inside body parts, forcing body parts, and the skin. They come together in the brain, joining with knowing from other parts of the body. Body position knowing is used for moving around and not falling over.

Abnormalities in proprioceptive feedback loops have been implicated in shoulder pathology and its management. Despite much academic interest in the field, clinical assessment of shoulder proprioception remains rare. One reason for this is practicability; in the laboratory setting, electromagnetic tracking and isokinetic dynamometry have provided elegant, yet complex solutions to proprioceptive assessment, unfortunately these methods are poorly suited to the time- and resource-limited environment of a busy musculoskeletal clinic.

Problems with proprioception have been linked to shoulder complaints and their management. Despite many academic studies, assessment of shoulder proprioception within clinics remains rare. In laboratories complex methods have been used for the assessment of proprioception; unfortunately, these methods are not well suited to use in a busy clinical environment.

Problems with body position knowing are often seen in people with shoulder problems. Although looked at in many studies, checks of shoulder position knowing are not often done within hospital sick areas. In studies of shoulder knowing the checks are not simply done and would take too long or not be practical to use in a busy hospital type place.

Balke et al, reported a clinically-feasible alternative joint position sense assessment method – the laser pointer assisted angle-reproduction test (LP-ART). This uses a wrist-mounted laser and a target, and has been found practical and effective for use in clinic.

Recent research has suggested an alternative method of assessing a patients sense of joint position, termed the laser pointer assisted angle-reproduction test (LP-ART). This uses a wrist-mounted laser and a target, and has been found practical and effective for use in a clinical setting.

New studies have suggested other ways of checking a persons sense of shoulder position, called the "light gun corner making check". This uses a light gun fixed to the persons arm and a mark on the wall to point at. It has been found practical for use in a hospital setting.

This study proposes a revised LP-ART, and presents initial trial results using novel interactive data graphics. These aim to be easily interpretable and encourage a full investigation of the data set, making them suitable for use by researchers, clinicians, patients and the general public.

This study proposes a revised LP-ART, and presents initial trial results using interactive data graphics. These aim to be straightforward to interpret and to encourage a full investigation of the data set, making them suitable for use by researchers, clinical staff, patients and the general public.

This study suggests a slightly changed "light gun corner making check", and presents check numbers using changing pictures. These are made to be easy to understand and to help you play with the numbers, making them good for use by both study people and normal people.

Method

Results

Subject:
Abduction/Flexion:
Gender:
Handed:

Age / years:

Height / cm:

Weight / kg:

Sleep / hours:

Show within-subject mean variance:
Show between-subject mean variance:

Analysis

Is elevation joint position sense acuity affected by target angle?

Show working
Start with two identical sets of results which include both genders and both arms in .
Consider only the Elevation errors.
Combine all targets in the first model.
Fit mixed effects models with subject as a random effect, and target angle as fixed effect only in the second model.
Check model assumptions of linearity and homoscedasticity by inspecting the conditional residual plots.
Check model assumption of normality by inpecting histograms of the conditional residuals.
Normalise each model by its residual standard deviation and then take the residual sum-of-sqaures to form log-likelihood measures.
Form a likelihood-ratio between the two models.
Assess the probability of achieving that likelihoood-ratio result by chance - assuming the chi-square distribution is a good approximation to the null distribution.

There is a statistically significant difference in error and variance between target angles for both flexion and abduction.

Is there a difference in elevation joint position sense acuity between dominant and nondominant arms?

Show working
Start with results for both genders in flexion.
Sort into dominant and non-dominant arm results.
Look only at the elevation errors.
Fit a mixed effects model with target angle and arm dominance as fixed effects and subject as a random effect.
Perform an F-test between the model and a similar model but with only target angle as a fixed effect.
Assess the probability of achieving that F-test result by chance.
Check model assumptions of linearity and homoskedasticity by inspecting the residual plots.
Check model assumption of normality by inpecting a histogram of the residuals.

In abduction there is evidence of a reduction in elevation error in subjects dominant arms.

There is no statistically significant difference in elevation error between arms during flexion.

There is no significant difference in elevation error variance in either abduction or flexion.

Is there any gender difference in elevation joint position sense?

Show working
Start with two identical sets of results which include both genders and both arms in flexion.
Consider only the Elevation errors.
Combine all targets in the first model.
Fit mixed effects models with subject as a random effect, and target angle as fixed effect only in the second model.
Check model assumptions of linearity and homoscedasticity by inspecting the conditional residual plots.
Check model assumption of normality by inpecting histograms of the conditional residuals.
Perform a likelihood-ratio test between the two models.
Assess the probability of achieving that likelihoood-ratio result by chance - assuming the chi-square distribution is a good approximation to the null distribution.

There is no statistically significant difference in elevation error between male and female subjects in either abduction or flexion.

There is some difference in elevation error variance between male and female subjects, but not to statistically significant levels.

Does elevation joint position sense acuity differ between flexion and abduction?

Show working
Start with two identical sets of results which include both genders and both arms in flexion.
Consider only the Elevation errors.
Combine all targets in the first model.
Fit mixed effects models with subject as a random effect, and target angle as fixed effect only in the second model.
Check model assumptions of linearity and homoscedasticity by inspecting the conditional residual plots.
Check model assumption of normality by inpecting histograms of the conditional residuals.
Perform a likelihood-ratio test between the two models.
Assess the probability of achieving that likelihoood-ratio result by chance - assuming the chi-square distribution is a good approximation to the null distribution.

There is a significant difference in error between flexion and abduction.

References

Boyar A, Salci Y, Kocak S, Korkusuz F. Shoulder proprioception in male adolescent tennis players and controls: The effect of shoulder position and dominance. Isokinetics and exercise science. 2007 Jan 1;15(2):111-6. Witherspoon JW, Smirnova IV, McIff TE. Neuroanatomical distribution of mechanoreceptors in the human cadaveric shoulder capsule and labrum. Journal of anatomy. 2014 Sep 1;225(3):337-45. Hatterman D, Kernozek T, Palmer-McLean K, Davies GJ. Proprioception and its application to shoulder dysfunction. Critical Reviews™ in Physical and Rehabilitation Medicine. 2003;15(1). Fyhr C, Gustavsson L, Wassinger C, Sole G. The effects of shoulder injury on kinaesthesia: A systematic review and meta-analysis. Manual therapy. 2015 Feb 28;20(1):28-37. Balke M, Liem D, Dedy N, Thorwesten L, Balke M, Poetzl W, Marquardt B. The laser-pointer assisted angle reproduction test for evaluation of proprioceptive shoulder function in patients with instability. Archives of orthopaedic and trauma surgery. 2011 Aug 1;131(8):1077-84. Haik MN, Camargo PR, Zanca GG, Alburquerque-Sendín F, Salvini TF, Mattiello-Rosa SM. Joint position sense is not altered during shoulder medial and lateral rotations in female assembly line workers with shoulder impingement syndrome. Physiotherapy theory and practice. 2013 Jan 1;29(1):41-50. Inspired by xkcd Tested here Details of mixed effect model...