Carpal Tunnel Syndrome (CTS) is a crippling disorder of the wrist that afflicts millions of workers around the world, and ultimately it can result in the loss of use of the hands. Since the late 1980s, the incidence of CTS and other related cumulative trauma disorders (CTDs) of the musculoskeletal system has skyrocketed, especially among those performing computer intensive work. Although the etiology of CTS is complex and not thoroughly understood, repeated forceful movements made by the hands when these are in deviated postures are known to dramatically increase the risks of developing this syndrome. In particular, repeated extremes of vertical wrist extension can put excessive pressure on the median nerve as it passes through the carpal tunnel of the wrist, and it is thought that eventually this impairs nerve function and ultimately results in injury.
In the past 5 years, ergonomics research at Cornell has studied how the hand and wrist moves during different types of work, and while using different types of tools, such as different computer keyboards. Although data have been collected using advanced biomechanical recording methods, the conventional statistical analyses of the wrist motion data have only yielded an incomplete picture of effects.
Consequently, researchers turned to the visualization group at the Cornell Theory Center to explore using scientific visualization tools to provide a more complete understanding of hand and wrist motion when using a conventional keyboard and to evaluate different designs and ergonomic arrangements to create low-risk postures.
A team including the author, graduate student Simonetta Rodriguez, and CTC visualization project leader Bruce Land chose a study recently conducted with Dr. Daniel McCrobie, corporate ergonomist at Honeywell, Inc., in which 38 professional office workers were monitored on site while using different keyboard arrangements. In one condition 23 workers were tested using various combinations of conventional keyboards, e.g., on the desk or on adjustable keyboard trays, with or without wrist rests. These same workers were tested again three weeks later using a Preset Tilt-down Keyboard System (PTKS). Fifteen other workers served as a control group.
In this study, vertical and lateral angular deviation data for right hand wrist positions was dynamically recorded at 5 Hz using an electronic exoskeleton measuring device. Conventional parametric statistical analyses of the data yielded several significant differences favoring use of the PTKS, but exactly what was happening to hand movements was not easy to grasp from these analyses. The challenge posed was to develop a scientific visualization of the data that would allow the researchers to understand what was happening to wrist movements in different situations. Also, the team was charged with developing visualizations that could be readily understood by non-technical people who either might be suffering from CTS or its early discomfort signs, or be involved in purchasing of ergonomic equipment to lower CTS risks.
Early on in the project, the team decided to develop all visualization tools using the IBM DataExplorer (DX) software. Team members also realized that in previous hand motion research vertical (extension/flexion) and lateral (ulnar/radial deviation) movements were always analyzed independently as two time varying angles, yet during typing unique combinations of movements occur over time in a three- dimensional space.
Results
Early visualizations that allowed both vertical and lateral angular deviation data to be seen over time showed considerable promise, and after exploring different options, the team decided to employ a colored tube to represent the trajectory of the wrist through time. Previewing this material with a number of non-technical subjects showed that people found it difficult to grasp whether the movement patterns constituted a risk or not, since the visualization didn't contain a reference point. In the next generation of images, researchers used a reference line at 0 deg. on both the vertical and lateral axes, but problems of interpretation emerged because all hand movements occur beyond this point. Consequently, the researchers turned to published empirical studies of changes in intracarpal pressure within the wrist as a function of vertical or lateral hand position, and used these data to model a neutral zone of movement. Hand movements inside of this zone produce minimal changes in intracarpal pressure. Now, actual typing movements could be visualized with respect to this neutral zone. Color variations along an individual data tube trajectory were used to emphasize proximity to the neutral zone. Figure 1 shows an example of this visualization for data generated by a person typing on a computer keyboard on a conventional keyboard tray. The visualization clearly shows that the pattern of movements is occurring outside of a neutral range of movement. Figure 2 shows the same worker typing the same text on the same keyboard, but using a PTKS. The movements now clearly are within the neutral zone. Figure 3 shows an end view of the visualization both before and after the addition of the PTKS. The effect of using the PTKS is obvious. What is also remarkable from these images is that the data were recorded some 2 months apart, yet the pattern of typing movements remains remarkably similar. This suggest that, like handwriting, typing at a keyboard may generate movement patterns idiosyncratic to particular individuals. Further work is planned to explore this idea and its potential applications.
Figure 1: Wrist tube trajectory for a subject typing on a keyboard placed on a
conventional keyboard tray
Figure 2: Wrist tube trajectory for the same subject typing on the same keyboard placed on
a Preset Tilt down Keyboard System
Figure 3: End view of before-and-after wrist tube trajectories for the same subject typing
on a keyboard placed either on a conventional keyboard tray (red) or on a Preset Tilt down
Keyboard System (blue)
Looking at individual data proved intriguing, but insufficient to determine whether or not the use of the PTKS produced an overall benefit for all of the workers who used it. The team next developed surface plot visualizations of the frequency distributions of combined vertical and lateral movements before and after use of the PTKS. These visualizations showed concurrent wrist vertical/lateral angular deviation measurement data for all subjects in a group collapsed over time, and collected into a two-dimensional matrix of "bins." The count of data points in each bin was extended into the third dimension to form a three-dimensional surface. The height of the colored surfaces represents the comparative number of data points for each combination of vertical/lateral angular deviation position. The longitudinal axis represents vertical wrist angles from 60deg. extension to 20deg. flexion. The transverse axis represents lateral wrist angles from 30deg. ulnar to 20deg.. The yellow surfaces represent the Survey 1 data collected prior to any treatment. The blue surface represents Survey 2 data collected from the same subjects 3 weeks after using the PTKS.
Additional visual reference information was added to these visualizations to aid interpretation. The curved elliptical low wall indicates the neutral zone, i.e., that region of the vertical/lateral planes assumed to represent little or no injury risk due to inferred low pressure in the carpal tunnel. The red box indicates that portion of the vertical/lateral planes assumed to represent significant risk of injury due to inferred high pressure in the carpal tunnel. Movements inside this danger zone will produce undesirable increases in intracarpal pressure.
Left-side and right-side views of visualizations of the distributions of movements before and after use of the PTKS are shown in figures 4 and 5. These images clearly show that when typing with the PTKS, the hands moved within a neutral zone 67 percent of the time, compared with 42 percent of the time using the other keyboard arrangements that were tested. In addition, the predicted pressure in the carpal tunnel remained below the critical threshold 82 percent of the time using PTKS, compared to only 48 percent of the time with other keyboard arrangements.
Hidden surfaces, however, could not be viewed in these static images The whole image was therefore animated to rotate to reveal otherwise hidden surfaces, and the before and after data were separately grown to emphasize the transition that occurred as a result of intervening with the PTKS. The figures shown here and additional images, along with details of this work, are available on a Web page that has been placed in the Nation al Science Foundation MetaCenter Science Highlights Repository (http://www.tc. cornell.edu/Research/Articles/CIE/IRI/Hedge/). A copy of the report and/or a video summary of this work can be obtained from Alan Hedge. Additional studies of the benefits of scientific visualization for improving understanding of ergonomic data are underway.
Copyright 1995 by Cornell University. All Rights Reserved.
