RESNA 27th International Annual Confence

Technology & Disability: Research, Design, Practice & Policy

June 18 to June 22, 2004
Orlando, Florida

Measuring Target Acquisition Utilizing Madentec's Tracker head pointing system in Individuals with Cerebral Palsy (CP)

Albert M. Cook, PhD*, Sharon Warren, PhD, Bonnie Dobbs, PhD, Rebecca McKeever, BScPT

ABSTRACT

The purpose of this study was to determine the impact of repeated trials of the Tracker head-pointing system in a series of target acquisition tasks for persons with CP. Twelve persons participated in a single case study. Data were collected on time-to-target, time-to-select, and distance-moved-to target across 4 once weekly on hour sessions, one hour in length. Overall, nine of the 12 participants achieved a smaller target size at the end of the study, with the majority of subjects moving one level (e.g., from 2.5 to 2.0 inch targets). Comparison across sets for the same target size indicates that seven participants had decreased time-to-target on their last set compared to the initial comparator set. Only three participants showed a decrease in their time-to-select scores but nine showed a decrease in distance-moved-to target. The group as a whole decreased their target size without any increase in times or distance.

KEY WORDS:

computer access; cerebral palsy; target acquisition

BACKGROUND

Computer technology is becoming increasingly prevalent in many aspects of daily life. Not surprisingly, level of computer skills and training have been identified as significant predictors of employment outcomes in individuals with a variety of disabilities (1). Individuals with a variety of disabilities, including cerebral palsy (CP), may not possess sufficient upper extremity function to effectively operate a computer keyboard or mouse utilizing traditional methods. In response to this, the assistive technology (AT) industry has developed alternative input methods and devices, with the intent to allow performance on par with non-disabled individuals (2).

The Tracker One head tracking system (Madentec) plugs directly into a computer's USB port and mounts directly on top of the computer monitor. It moves the mouse cursor by tracking the position of a small reflective dot, which the user wears on his or her forehead or glasses. The Tracker is compatible with USB, PS/2, and Serial, systems, and performs well in the direct sunlight. Although there is evidence to support its effectiveness in clients with spinal cord injury, little is known about its potential for individuals with CP (3).

The purpose of this study was to determine the impact of repeated trials of the Tracker One system in a series of target acquisition tasks for persons with CP: specifically the impact of repeated trials on time-to-target and time-to-select, and on distance-moved-to targets of increasingly smaller size between sessions.

METHODOLOGY

Participants

A convenience sample of 12 participants was involved in the study. Potential volunteers were recruited through various treatment facilities and service organizations for persons with CP. The inclusion and exclusion criteria for the participants are outlined in Table 1. Potential participants were screened to ensure that they possessed physical and sensory abilities adequate to operate the Tracker One system.

Procedure

A single case experimental design was used. Participants and their parents/guardians were engaged in weekly sessions of one full hour each to a total of four sessions. Speaking Dynamically Pro (Mayer-Johnson) was used to create grids of square targets with pictures that are selected by mouse movement and click (or dwell). There is speech or sound effect feedback when a target is selected. Total time was separated into time required to move to the target (time-to-target) and time to select (time-to-select). Mouse Offroad (A!K Research Labs) (a mouse odometer) was used to measure distance of mouse travel from the starting point to the point of target acquisition. A central target on each grid was designated as the ‘home' target. Participants were asked to move the cursor from the home target to a peripheral target and hold the cursor until the feedback was elicited. Once the home target was selected, another target appeared in one of four peripheral locations arranged in the 12, 3, 6 and 9 o'clock positions. The location in which peripheral targets appeared was randomly varied. Following a selection, the participant was requested to move the cursor to the home target. A successfully completed session consisted of selecting a minimum of two targets in each of the four peripheral locations (a minimum of eight selections) for each size of target: 2.5, 2.0, 1.5, and 1.0 inches on a side. All participants began with the largest target. If a successful set was completed at that size within the hour session, the next smaller size target was tried. However, most participants could not complete more than one target size per session and they began with that same size at the next session, until successful and then proceeded to the next smallest target.

Time (in seconds) was measured, using a stopwatch to separately record the time required to move to the target; and the total time to acquire the target (including the one second minimum dwell time required by Speaking Dynamically Pro). Time-to-select was calculated as the difference between the total time and the time to move to the target. The time-to-target measure began when the cursor left the central home position and ended when the cursor was held on the peripheral target location for the required dwell time. The mouse odometer recorded cursor travel to move to and select the target as ‘distance' in metres. The difference between the start odometer reading and the end odometer reading in a data set was recorded as the travel distance for that set.

Data analysis

Because the participants were variable in their skill level and because of small sample size time and distance, results were graphed for each participant rather than generating group means and standard deviations to compare statistically over time. Times-to-target, times-to-select, and distances-moved-to target were averaged within a session to give a representation of participant performance. Visual inspection was first used to assess change.

RESULTS

Of 12 participants, 17% (N=2) had athetosis and 83% (N=10) had spastic quadriplegia. Fifty-eight per cent (N=7) were male and 42% (N=5) were female. The average age of participants was 30.5 years with a range from nine to 49. Thirty-three per cent (N=4) of the participants were living with family, 33% (N=4) in a group home, 17% (N=2) alone, 8% (N=1) in assisted living, and 8% (N=1) in other arrangements. Forty-two per cent (N=5) were currently attending school and another 42% (N=5) were attending some other type program, such as a work program. Tables 2 and 3 show selected participant characteristics with respect to mobility assistance and computer use.

Tables 4 through 6 illustrate the raw data. Few participants were able to complete all target sizes in one session. Last session scores are typically based on smaller target sizes. The typical pattern was for a participant to gradually decrease scores (faster time and shorter distance traveled) on one target size, then increase initially when moving to a smaller target size, and decrease again at that target size. The shaded areas for each participant allow for comparisons across sets using the same sized target (e.g., 2.5 vs. 2.5 inch targets, 2.0 vs. 2.0 inch targets). Nine of the 12 participants were able to achieve a smaller target at the end of the sessions compared to the initial target size of 2.5 inches. All completed at least two sets at the same size target: 10 at 2.5, and two at 2.0. Of these 12, nine had decreased times-to-target on their last equivalent size target set from their first set and three had increased times (Table 4). Only three showed a decrease in time-to-select and one remained the same (Table 5). Ten participants showed a decrease in distance and two an increase. (Table 6).

DISCUSSION

Participants' skill in using the Tracker One head pointing system generally improved over time. The majority could gradually hit increasingly smaller targets, decrease their time-to-target; and decrease their distance-to target. The exception was in time-to–select. This result may be due to the inclusion of a fixed acceptance time of one second in all results. The pattern of results suggests that more than four one-hour sessions may be required for participants to achieve maximum proficiency, since it appeared that two to three sessions at any one target size were needed before most could progress to a smaller target size.

REFERENCES

1. Anson, D. (1991). Using the Headmaster with Macintosh Apple II, and MS-DOS computers. American Journal of Occupational Therapy , 45(10): 889-897.
2. Pell, S.D., Gills, R.M. & Carss, M. (1997). Relationship between use of technology /and employment rates for people with physical disabilities in Australia: Implications for education and training programs. Disability Rehabilitation , 29(8): 332-338.
3. http://www.madentec.com

AKNOWLEDGEMENT

This work was done under the 'Medical Device Development Program' funded through the Canada/Alberta Western Economic Partnership Agreement (WEPA). Drs. Johanna Darrah and Ron Willis offered many constructive comments during the course of this work. Madentec, Ltd loaned the Tracker One for this study.

Albert M. Cook, Dean,
Faculty of Rehabilitation Medicine,
Room 348 Corbett Hall,
University of Alberta,
Edmonton, Alberta, Canada T6G 2G4.
Phone 1-780-492-5991