The upper extremity has an important role in several daily activities such as eating, drinking, clothing, grooming, writing, as well as in different sports and leisure activities. These activities require coordination of multiple joints and involve both the musculoskeletal and neural systems [1]. Impairment of upper extremity is one of the most common sequels following CNS lesions [2, 3] and is also frequent in patients with musculoskeletal impairments involving the upper extremity. A dysfunction in the upper extremity can significantly limit a person's level of activity and participation in their social and physical environment [4].

Upper extremity function is in neurological settings generally assessed with observer-initiated standardized measures based on ordinal scales, e.g. Fugl-Meyer Assesment, Frenchay Arm Test, Motor Assessment Scale, Action Research Arm Test, or as timed tests, e.g. Box and Block Test, Nine Hole Peg Test [5–7]. These outcome measures are reliable and sensitive for measuring gross changes in functional performance but have less sensitivity to smaller and more specific changes. Furthermore, despite the extensive experience in using these observer-initiated measures by clinicians, the subjectivity of these tests cannot be denied.

A better understanding of human movement requires more objective testing and accurate analysis of motion to describe the arm movements more precisely and specifically during functional testing. Kinematic analysis is one such method. Kinematics describes movements of the body through space and time, including linear and angular displacements, velocities and accelerations, but without reference to the forces involved [8, 9]. Three-dimensional imaging measurement techniques, including optoelectronic systems, have turned out to be a powerful tool for a quantitative assessment of movement in all degrees of freedom. The models for lower extremity movements and gait analysis have been well established in biomechanical and clinical research and are now applied to the detailed diagnosis and treatment planning of patients. However, the variety, complexity and range of upper-extremity movements is a challenge to assessment and interpretation of data and the clinical routines for three-dimensional analysis in upper extremities are not fully established [10].

Movement analysis of reaching can provide precise quantitative and qualitative data of arm movement in space including movement velocities and accelerations. Joint angles and interjoint coordination can be calculated. In addition to the assessment of performance, kinematic measures can be useful for elucidating the motor strategies in goal-oriented tasks, as well as for evaluating upper-extremity therapies [7]. Most studies with kinematic analysis of upper extremities involve reaching movement carried out in the horizontal plane with the arm supported and in highly constrained conditions [7, 11]. There are, however, an increasing number of studies with more "natural" reaching or pointing movements were in some cases even grasping an object is included [12–15].

Only a few studies have analyzed a functional task for upper extremity with three-dimensional kinematic analysis. Two of these studies have had the intention to attain data for upper limb kinematics in order to support the development of upper limb joint replacements [16, 17]. Murgia et al investigated the motor control of wrist movements in two activities of daily living (jar opening and carton pouring) in four healthy persons [18]. One pilot study investigated the two-dimensional forearm movement during transport phase of drinking activity, with focus on effects of concentric and eccentric exercise training in elderly healthy women [19].

Kinematics studies have shown that reaching and grasping movements vary according to the goal and constraints of the task. For example, a pointing movement has different kinematics than a movement combined with grasping an object, in the same way that reaching movement kinematics are different depending on if the real-life object is present or not [9, 20]. Studies of natural and goal-oriented movements are of particular relevance to clinical practice since they provide essential information of person's real capabilities [7, 10].

There are no studies, in our knowledge, which have analyzed the whole drinking movement with kinematic characteristics. The starting-point for this study was to investigate and analyze this daily activity with kinematic analysis without physical restraints on the normal movement of drinking. We were also interested to explore the potential of this method for use in clinical practice. To get answers for these considerations it is necessary to develop a method of three-dimensional analysis of the drinking activity and gather the reference data for control persons. The aims of the present study were:

1. To develop a protocol and test the consistency of that protocol for the three-dimensional motion analysis of an daily activity "drinking from a glass",

2. To obtain descriptive group data for this drinking task in control subjects.

Statistical analyses were performed with SPSS (Statistical Packages for Social Sciences, 11.0). Descriptive statistics including mean, standard deviation and 95% confidence intervals (CI) were calculated for the study group of 20 subjects and for test-retest data. The mean of the three recordings was used in statistical calculations.

The difference between test and retest was analyzed with a paired t-test with alpha level at 0.05 and with hypothesis testing based on confidence intervals of the test-retest data. The agreement between test and retest was evaluated with 95% limits of agreement (LOA) method [22, 23]. The 95% LOA were calculated as the mean of difference ± 1.96 standard deviations of difference. This method calculates the limits within which expected differences between two measurements will lie with 95% probability. To check the assumptions of the limits of agreement the differences were plotted against the average of the two measurements for every variable.

The present study provides a detailed three-dimensional kinematic analysis of the drinking task in control subjects. A standardized test protocol for the drinking task was developed, including the set-up of cameras, measurement volume, location of markers and position of the subjects. The test protocol demonstrated a good consistency in test-retest and provided clear and accurate results. The phase analysis which divided the drinking task into five sequential phases was unique for the present study. Kinematic data are presented for the movement times, marker positions, velocities and joint angles for the control subjects.

The drinking activity is a complex task in terms of kinematics. It contains several different movements as reaching, grasping, transporting the glass and drinking. In the present study, five sequential phases were identified: reaching, forward transport, drinking, back transport and returning. This phase analysis gives a logical and easy observable structure for the drinking task and provides the possibility to investigate different variables separately in each phase. There are no studies, to our knowledge, which have presented the phase definitions for the whole drinking movement; hence the phase analysis applied for the drinking task in this study is unique.

The concept of three-dimensional joint movements and joint angles is a quite unusual for clinicians. It is common to think of two-dimensional movement in different planes. To make this study more obtainable for clinicians and more comparable with other clinical studies, we divided the shoulder elevation into abduction and flexion and recalculated the values for elbow angle into an anatomical angle rather the technical/mathematical angle values. One limitation of this study, for example, is that we have not analyzed the rotations in shoulder and elbow joints or joint angles in the wrist joint. Wrist joint motion and forearm rotations could provide additional information of different drinking strategies. Although the system has the potential to provide this information, a much more complex set-up and analysis would be required. The goal of this study was not to measure the joint angles in all joints and in all degrees of freedom. Our intention was to collect informational and clinically useful data for the drinking task with the existing camera system and according to the current stage of software development.

We have used surface markers and computed the joint angles as the angles between the corresponding vectors joining the adjacent markers or vertical vector. It must be understood that those joint angles do not pass through the centers of rotation of the joints, thus they are not the true joint angles. However, placing the markers on the well-defined superficial bony prominences increases the reproducibility of data on different occasions. This was confirmed by the good consistency in test-retest in this study as well in other studies using surface markers [21].

Earlier studies of reaching movements have reported that trunk movement is acting both as stabilizer and as an integral component in positioning the hand close to the target [24]. Several studies have also shown that when reaching within arm's length, healthy subjects use minimal trunk displacement in contrast to hemiparetic subjects who use a compensatory strategy involving trunk recruitment [24–26]. In the present study the object was placed within an arm's length and the subject could reach the glass while sitting against the chair back during the whole task. Any unintentional trunk displacement was also measured using a marker on the thorax. The intention was to keep the drinking movement as natural as possible but allow some trunk displacement during the task.

Interjoint coordination for the elbow and shoulder joint excursions during a reaching movement is shown to be highly coupled in normal subjects, but has a significant decrease in stroke subjects with hemiparetic arm [11, 12]. Levin et al suggested that a measure of interjoint coordination can give us clinically beneficial information about the subject's motor function. In the present study we found a high correlation between the shoulder and elbow joint excursions in reaching phase indicating a good interjoint coordination. Several studies have suggested using angle/angle graphs for qualitative analysis of interjoint coordination [11–13]. We have constructed the angle/angle graphs for the reaching movement and found a smooth and almost linear curve between elbow and shoulder joint excursions.

All measurements systems including the kinematic systems, suffer from measurement error. A typical average value for measurement error is estimated to be 2–3 mm in all dimensions in gait analysis [8]. Turner-Stokes et al found the absolute mean difference (differences between two measurements in either direction are treated as positive) in joint range to be 2.3° for shoulder and 2.7° for the elbow in repeated measures in the bowing arm of string-playing musicians when the whole measurement system was dismantled and re-set-up [21]. Replacing the markers produced slightly greater differences than repositioning and recalibration of the whole system [21]. In this study we tested the consistency regarding the replacing the markers and the repositioning the subject. The absolute mean difference in test-retest for shoulder abduction was 2.7°, for shoulder flexion 1.1°, for elbow extension 2.0° and for elbow flexion 1.2°. These values are comparable with results in study of Turner-Stokes et al.

The analysis of upper extremity tasks requires a measurement set-up and camera system which can track all the necessary markers throughout the whole movement. This has been a challenge in the present study. From total 112 recordings we had to exclude 8 recordings (7%), because of the high segmentation and gaps in data. This data lost is acceptable considering the biomechanical complexity in the upper extremity movement analysis [10]. The problem with segmentation and gaps in data have also been reported in other studies, especially when few (two to four) cameras are used and when the movement is complex in degrees of freedom [17, 20, 21]. As reported in other studies, the set-up of camera systems and data analysis in special software programs required a good collaboration between clinicians and engineers in the present study [10, 21]. Results of this study corroborate that that the use of the existing motion capture system, as used in the presented protocol, still has some limitations and requires further refinement to be feasible for clinical use with persons with neurological disorders. The main clinical benefit of this study is the establishment of the phases of an important activity of daily living (ADL). The establishment of the phases can simplify the camera measurement system and allow for increased clinical use. In addition, the phases provide a descriptive methodology to classify strategies of drinking task of healthy persons.

Kinematic analysis has great possibilities to be used as an outcome measure in clinical research especially when optoelectronic camera systems become more readily available in clinical settings. Our approach to investigate and analyze a goal-oriented daily task has a great clinical potential. However a certain degree of standardization is indispensable in order to obtain repeatable and reliable results. Further development of objective and reliable evaluation methods for upper extremity tasks is required, particularly for natural and goal-oriented movements. Consequently the next step is to use and test this protocol on persons with impairments and disabilities from upper extremities e.g. persons with neurological and musculoskeletal diseases.