Improving Computer Interaction for People with Parkinson's Disease

Miriam Eileen Nes Begnum

MediaLT, Jerikoveien 22, 1067 Oslo, Norway


Abstract: As the situation and main challenges for PC users with Parkinson's disease (PD) has been investigated, it has become evident that suitable adaptations are needed. Using standard computer mouse solutions are particularly problematic. This article describes the results from real end-user tests of potential suitable adaptations for people with PD, in order to determine their effect upon inertia, muscle stiffness, tremor, pain, strain and coordination. We show that shelf-ware could significantly improve mouse pointer control for many users. Our tests also illustrate how projected interactive PC-management may improve computer interaction - without risking strain injuries.

1. Introduction

The results of a Norwegian survey (Parkinson’s IT challenges – PIKT) on the computer use of people with Parkinson's disease (PD) (Begnum 2010) showed that nearly 80 % of PC-users with PD have significant and severe difficulties using a computer due to their illness. According to the survey respondents, alternative keyboard and mice solutions impact ease of computer use, but the user population lack experience with and knowledge of such alternatives.

A person with PD is her/himself responsible for acquiring and paying for relevant alternative PC equipment. No governmental or organizational recommendations are available at this point, making the process of optimizing the computer interaction a tiresome and possible expensive process. Some PC-users with PD recommend certain facilitating devices to their peers (with similar PD symptoms). Besides this, individuals with PD have to find relevant adjustments themselves. The aim of this article is to ease this process, by investigating possible computer adaptations through real user tests.

2. Background

Parkinson's disease (PD) is caused by a disorder of the central nervous system, afflicting one in 500 people (Parkinson’s New Zealand). It is usually diagnosed from the age of 55-60 years and upwards. The disease is projected to double in the next 25 years (Dorsey 2006). PD leads to a number of symptoms, such as tremor, muscle stiffness and pain, reduced and slowed motor skills, reduced balance, difficulties coordinate and resulting clumsiness, dementia-related symptoms, weakened voice, visual disturbances and lack of energy. It is progressive, degenerative and chronic. PC interaction challenges are mainly linked to the motor symptoms, and using the computer mouse and keyboard peripherals are main issues (Begnum 2010). About half of PC-users with PD struggle with significant, severe or highly severe muscle stiffness when using a computer, and equally many with inertia. Tremor is also a common challenge in relation to PC use, seriously affecting about 30%. A smaller sub-group of PD-users are only significantly challenged by tremor, reducing the arm/hand control.

In spite of general PD ailments, people with PD want to continue living as normally as possible, being able to use a computer efficiently and effectively without unnecessary fatigue, pain and strain. Knowledge of beneficial adaptations is thus important, however little has been done to investigate this. One other study focusing on device usability for PC users with PD has been found (De Wet 2005); however testing only three possible adaptations, none of which improved computer interaction for their seven participants, and focusing solely on tremor as an interaction issue.

Several software and hardware adaptations exist to address tremor related troubles, such as the Mousecage software (Softpedia), the IBM Assistive Mouse Adapter (Levine 2005) and various keyguards. Keyboards exist with fewer keys, larger keys and enhanced key label visibility. Keyboards with less recoil are available - even some made of soft rubber. A wide range of computer mice are also obtainable - providing alternative sensitivities, weights, pointer control techniques and buttons. Many of these are optimized towards computer gaming.

Although few standard peripherals are designed to lessen the impact of such symptoms as akinesia, muscle stiffness and pain, a wide variety of ergonomically designed equipment, keyboards and computer mice are available. These are typically designed to avoid strain injuries for non-disabled computer users. Equipment designed to reduce ergonomic stress, such as strain and fatigue, seem quite well known in the population. To the degree that ergonomic solutions can provide help for the user group, they are considered useful.

The Microsoft Accessibility tutorials can be used to systematically try out OS built-in MS accessibility features. In addition comes specialized aid technologies - e.g. the foot operated mouse.

3. The Quest for Suitable Adaptations

A test-group of eight persons with PD was selected for our user tests: five men and three women. The age span was 48 to 65 years. Two of the testers work full-time; two are disabled and four work part-time. All testers have general computer skills, and use computers on a regular basis; all but one person use computers daily. Participants described interaction challenges, and their seriousness, during a pre-test interview. In addition, each tester filled out a questionnaire (developed and validated trough the PIKT survey) to provide further individual information and insight into the detailed challenges, experiences and computer use of each participant.

Based on this input, and verified against post-test data from test observations, seriousness ratings on each individual’s interaction challenges were determined. The members of the test-group were then categorized according to computer interaction challenges and their seriousness. Comparing the answers from the test-group with the responses from the PIKT survey, we see a similar distribution – indicating the test group is fairly representative.

Table 1 summarizes what computer interaction areas each tester’s challenges are related to, and their seriousness. Challenges described as insignificant or trivial are not included. As the table shows, all testers experience difficulties using a computer mouse, keyboard devices are problematic, and ergonomics are also a problem area.

Table 1. The Main Computer Interaction Challenges of the Test Users
Par. 1 Par. 2 Par. 3 Par. 4 Par. 5 Par. 6 Par. 7 Par. 8
Ergonimics Serious Serious - Very serious - - Serious Serious
Keyboard Quite serious - - Quite serious Quite serious Quite serious Serious Serious
Computer mouse Quite serious Serious Quite serious Quite serious Serious Quite serious Serious Very serious
Screen - - - Quite serious - - - -

The testers were grouped into four categories, described in Table 2. The even distribution is coincidental, as the testers were not paired when categorized. All participants other than testers 3 and 5 have complex challenges relating to hand and arm control. In addition to mouse related difficulties, they also have varying degrees of ergonomic challenges and keyboard issues.

Table 2. Test User Categories
Category Participants Short descriptions
A 3 & 5 Mouse related problems are the only, or by far the most prominent, area of difficulty. Tremor is the only, or by far the most prominent, challenge related to computer mouse control.
B 1 & 4 Ergonomic challenges are the most significant, with related difficulties regarding using keyboard and computer mouse. Tremor is a central problem area. Muscle stiffness and inertia are less problematic, but present.
C 7 & 8 The testers in this category have serious issues related to both computer mouse and keyboard, as well as ergonomics. The most prominent ergonomic challenges are pain and fatigue. Tremor and arm/hand control is also a challenge for this category.
D 2 & 6 Muscle stiffness and inertia are the main challenges. Fatigue and pain may be problem areas. Tremor is an insignificant trouble. The testers focus on device related issues.

3.1 Implementation

User tests were held over a three day period. A selection of hardware and software alternatives for testing were made based on assumed usefulness and potential, while securing a broad range of alternatives and including new and innovative solutions. 45 different input devices and adaptations were offered the participants, of which 34 were chosen to be tested by the users. The tested peripherals are listed in Table 3.

Each peripheral tested had a description attached, clarifying to the testers the issue(s) assumed to be improved. Since needs and type of computer use varied, the decision was made to leave it up to the testers to choose what peripherals they should test and also not to require a specific scenario to be completed for each test. A second questionnaire form was however designed, and attached to each peripheral’s description. This form aimed at assessing the ease of use and benefit of a peripheral in a specific trial; how a particular user assessed a certain adaptation. Items made the testers describe in detail how the device/adaptation worked, the aim of the trial and whether expectations were met. This form was compulsory to fill out for all tested peripherals. Each tester thus filled out one form for every specific trial he/she completed.

Each tester was assigned a personal observer, who followed them during the tests. The observer helped guide the tester, suggested potential adaptations to try out and assisted in correctly filling out test forms. The observers were responsible for making sure participants took the time and effort to thoroughly fill out test forms. Thus, the form-filling was usually conducted in the form of a mini-interview, where observations and user feedback were discussed, clarified and documented.

Table 3. The Peripherals and Adaptations Tested
Mouse peripherals: Tablets with a pen stylus, Penmouse, Perific handheld mouse, Ergonomical left-hand mouse, AirOrbic vertical right-hand mouse, Whalemouse, Touchpad, Logitech marble trackball mouse, Kensington trackball orbit expert, Kensington trackball orbit optical, Mouse with adjustable weight, Joystick-mouse with changeable top, Anir joystick-mouse, Headmouse (with reflectors), IBM Assistive Mouse Adapter, Norlink New Concept Mouse, Mouse stick, Projected Interactive PC-control prototype solution, Foot mouse NoHands Mouse.
Keyboard peripherals: Ergonomic split keyboard, Rubber keyboard, "Easier life" keyboard, Logitech ergonomic keyboard, Miniature keyboard, Keyguard, Virtual laser keyboard, X-keys, On-screen keyboard, DigiScribble.
Ergonomics: Arm rest, Ergonomic chair, Adjustable height office table, Laptop pillow.
Other adaptations: Adaptations in Vista - Changing color/size of mouse pointer, High contrast screen setting, Mouse keys, Repeat keys, Slow keys, Shortcuts, Adjusting mouse button clicking rate, Automatic mouse pointer movement. Larger computer monitor.

To more accurately and consistently determine levels of mouse pointer control, a simple drawing test was used: An ellipse was drawn in Microsoft Paint as a template, and the testers were asked to follow the elliptic line with the mouse in drawing mode. The mouse control challenges and the expediency of the adaptation were demonstrated by repeating the drawing test with different computer mice, tremor filter settings on the IBM Assistive Mouse Adapter etc.

4. Test Results

The different peripherals’ suitabilities were determined based on their documented effect upon inertia, muscle stiffness, tremor, pain, strain and coordination. During data analysis, participants as well as observers were asked to clarify and elaborate on the described results when necessary to ensure a coherent report. Before completing the analysis, summarized findings on individual test results were discussed with, and verified by, each participant.

4.1 A Troublesome Computer Mouse

Even though a large individual difference within the user group was verified through interviews and tests, using a computer mouse was confirmed as the main problem for the user group. Reflecting the main user problem, all testers chose to put their main focus on alternative mouse devices. Difficulties using computer mice were prominent for all testers, even though the reasons for this were slightly different.

Using a computer mouse generate two common troubles for the user group: controlling the mouse pointer and clicking the mouse buttons. Our tests show that the solution to one of these issues unfortunately often further complicates the other.

In order to optimize clicking for the user group, our tests show that two device attributes are needed: 1) Buttons that are easy to reach (- avoiding fatigue and pain when clicking), and 2) a design facilitating holding the mouse in a stable position while clicking (- even with reduced arm/hand control).

For optimized mouse pointer control, the tests showed that the pointer must be easy to move (e.g. slide well along a surface) whilst being insensitive to small movements as well as facilitating increased movement control (e.g. being able to use multiple fingers on a trackball).

4.2 Clicking and PD Challenges

No general solution was found to solve issues related to mouse button clicking. The trackball mouse solutions did not aid users challenged with troubles related to click-control. On the contrary, testers complained that steadying the mouse pointer while clicking was difficult using these solutions – particularly for testers in category D, where mouse clicking were problematic and tremor less of a problem.

Several testers preferred a joystick mouse for mouse clicking, having a different placement of mice buttons compared to a standard mouse and also far more stable. The joystick mouse seem suited for users with larger ergonomic issues such as fatigue, pain, muscle stiffness and inertia, especially when combined with other ergonomic support devices, such as arm supporters. On the other hand, the joystick solution also received complaints from testers experiencing pain and fatigue, and was judged as being heavy and hard to maneuver, providing poor overall mouse pointer control.

Suitable solutions for clicking when suffering from reduced arm/hand strength and/or control seem particularly lacking.

4.3 Mouse Pointer Control and PD Challenges

Generally, touchpad and mouse stick are not at all beneficial for solving serious mouse control issues for PC users with PD. For left-handed and ergonomic mice the feedback varies, though some found these beneficial. Head mouse and foot mouse solutions received an overall negative rating, being alien control concepts to the users. Touch technology, IBM Assistive Mouse Adapter and trackball mice received however positive feedback.

When it comes to improving the control of the mouse pointer, trackball mice seem appropriate for people with PD - especially where trembling is the main problem. All tester categories were helped by trackball mice, which where denoted as the best mouse solution by most of the testers. Our tests show that this type of mouse provides significantly increased mouse control. The trackballs were reported to be easy to use, not moving too fast, "cleaner" and more accurate than ordinary mice, intuitive, steady, good grip, ergonomically fitting with the opportunity to use multiple fingers as well as both hands, comfortable, good ball control, nice to scroll with and providing great pointer and mouse control. Trackballs of different types and sizes were tested, and the most important feature for optimized control seems to be whether the mouse is ergonomically fitting the users hand size - smaller mice fitted users with smaller hands, while users with large hands preferred larger mice. In addition, some minor differences in mouse design related to individual testers particular additional challenges influenced the usability.

The IBM Assistive Mouse Adapter filter for tremor was useful for some to increase drawing and precision control, and generally received a lukewarm positive rating. Often, medication is used to remove the main shaking of the hands/arms, and this is a possible cause of why the filter does not play a more important role in increasing mouse control. It seems finding the appropriate computer mouse has the greatest effect on solving the challenges related to trembling, and not the filter.

4.4 Mouse Issues and Interactive Touch Technology

Our tests surprisingly revealed that touch based computer interaction could prove a solution to both challenges related to mouse usage. Interactive touch technology detects pressure and positioning on a screen area, and is a promising technology for users with mobility impairments since it provides possibilities for movement of navigational peripherals on larger surfaces - reducing the need for detailed coordination skills. Two such solutions were tested: tablet with a pen stylus and Projected Interactive PC-control (PIP) prototype solution.

The PIP solution consisted of a glove with a point and clicking device attached to one of the fingers, and the screen projected on a table in front of them. A sensor above the table registered the position of the glove and when it touched the surface of the table, subsequently communicating a “click” to the computer. The computer picked up movements and clicks through infrared light in the pointing device - the glove. Steering the computer by directly manipulating the perceived computer screen was an easy concept to grasp for the testers, and the control solution seemed familiar. It seemed the direct touch with natural tactile feedback increased mouse pointer control, as well as minimizing pain and fatigue. The testers did not have to care about the whereabouts of the fingers and limbs not equipped with the pointing/clicking device - since no light were projected from these. This meant the testers could lean in and support their movements if they wanted, reducing fatigue and strain, stabilizing the pointer when steering - further increasing the mouse control and steadying the hand and finger while clicking - providing increased click precision. The solution spurred no unfavorable reactions, receiving remarkably positive ratings for being such an early prototype.

The tablet was mostly positively rated, especially from category C in which the testers mainly struggled with pain and fatigue, though some users’ initial reactions were that the interaction felt strange and difficult – finding the concept of steering what happened at the computer screen with the tablet hard to grasp. Others did not appreciate the ergonomic design solution and had difficulties holding the pen, resulting in the tablet not picking up movements with the necessary level of detail. In addition, clicking on the tablet was difficult. The tablet might have proven more useful if extended user training had been provided, and the particular peripheral were tested over a longer period in time - providing the opportunity to get accustomed to the new control concept before judging it.

4.5 Keyboard Preferences

Our tests show that there was a large spread in the assessment of keyboards. For some users, smaller keyboards worked well – reducing the need for finger movement and thus reducing pain and fatigue. For others, lack of precision made small keyboards hard to use, and larger alternatives were necessary despite the strain effects. The individuals’ main challenges and also the combination of these resulted in different keyboard preferences – for example preferred one tester a keyboard with enlarged key labeling and improved contrast/colors ("easier life") due to reduced eyesight.

In general, different ergonomic keyboards were beneficial compared to standard solutions. Ergonomic and/or split keyboards seem appropriate to optimize for the touch method, and to reduce ergonomic stress, but do not necessarily resolve any specific challenges related to the use of keyboards. Ergonomic and split keyboards seem especially appropriate for users suffering from slowness, rigidity and muscle stiffness, too help reduce clicking error rate and increase the speed of keyboard writing. Rubber keyboards received a slightly negative rating in the test. Other than this, there are no findings suggesting a specific type of keyboard more generally suitable than another. Keyguards were not considered helpful.

4.6 Other Adaptations and Peripherals

Different types of arm rests were described by testers as relaxing and supporting the arm, and considered beneficial to combine with appropriate keyboards and mice solutions. Arm supporters seem to work particularly well as an aid for those struggling with trembling.

Adjustable chairs and tables that can be raised and lowered are favored over non-adjustable alternatives, used to facilitate ergonomic and variable working positions, and assessed as suitable for minimizing troubles such as strain, pains and fatigue. So are ergonomic chairs.

Less than half of the testers chose to try the suggested adjustments in Windows Vista. All who did found these to be beneficial. Improved visibility of the mouse pointer as well as high contrast settings reportedly improved mouse pointer control. Setting for automatic movement of the mouse pointer reduced the users need for moving the pointer. Also, the ability to adjust mouse pointer speed was felt to be beneficial. Further, Vista has several keyboard key settings that may be adjusted. The testers particularly liked the filter/bounced keys and slow keys. Shortcut possibilities in Vista were also considered useful.

Only one of the testers had serious visual disturbances and reduced eyesight, which could be solved by a larger computer screen.

5. Discussion

The individual symptoms are quite different from person to person with PD. This hampers efforts to identify general recommendations. The need for individual adaptation and device selection is apparent. No single solution could be identified for the user group as a whole. For most problem areas, the study indicates potential solutions. However, it should be noted that the trials were all of relatively short time span, and included only typical tasks. The results may therefore need adjustments related to issues of use over time. Users may also be more negative when quickly testing unfamiliar ways to control a computer, such as the foot mouse, where prolonged acquaintance may be required for a fair assessment.

A major computer interaction issue is controlling the computer mouse. The difficulties related to mouse control is two-fold: pointer control and mouse button clicking. Tests show improved button placements often collide with the need for a design giving optimal mouse pointer control. The challenge of finding an optimal mouse peripheral solution both in relation to pointer control and button placement is ongoing. It is worth noting that the optimal solution is not necessarily to find one single mouse. Switching between different mice solutions - used for different purposes and/or in different situations- could be beneficial. The combined use of several computer mice input devices is an area worth investigating.

For the sub-group of users mostly troubled by tremor and subsequently lacking mouse control, an ergonomically fitting trackball mouse seems beneficial. This was an interesting test results, as there were no specific expectations towards the trackball mice alternatives prior to the tests. These mice are shelfware, both cheap and easy to get hold of, and can easily be distributed to end-users. It is therefore regrettable that only 2% of the user population (Begnum 2010) has ever tried such a computer mouse alternative, something that points to the need for providing more knowledge of suitable adaptations to users. Knowledge of Vista customizations should also be distributed.

Clicking in general, and mouse button clicking in particular, proved to be a far greater challenge than expected. Finding beneficial solutions for clicking, both with regards to finding general solutions for the user group in question as well as for a particular individual, is problematic. This is a problem area that needs further research activities. Unfortunately the trackball mice tested were all poorly designed in relation to ease of button clicking. Buttons on tested trackball mice were on the socket, and so further tests with a clickable trackball would be interesting. Touch technology proved an interesting platform for a combined pointer and click precision and control solution, and should be further looked into as a mouse alternative for this user group.

6. Conclusion

Results from this study indicate which solutions could be beneficial for computer users with Parkinson's disease, depending on a person’s main challenges. Results from this study refine and significantly extend the deWet 2005 findings. Generally inappropriate solutions are identified. Individual needs are clarified and the need for individual adaptations verified. Possible solutions for most device-related challenges seem to exist. Challenges related to clicking proved to be a far more significant problem area than previously considered. Touch technology emerged as an interesting PC-control alternative for PC-users with complex interaction challenges.

7. Acknowledgements

The author would like to thank to the volunteer participants, the Norwegian Parkinson Foundation and Oslo University College for invaluable contributions.

8. References