Cueing and Logical Problem Solving in Brain Trauma ...binde1.verio.com/wb_fluency.org/Publications/MerbitzMillerHansen... · 33 individual treatment sessions with an experienced speech-language

31EUROPEAN JOURNAL OF BEHAVIOR ANALYSIS NUMBER 1 & 2 (2003)2003, 44444, 31 - 43

31

Following severe traumatic brain injury (TBI),persons undergoing rehabilitation are often char-acterized as having problems with impulse con-trol, judgement, attention and concentration, er-ror monitoring and utilization, and consistency(Hanlon, 1994; Willer, Allen, Liss, & Zicht, 1991;Dixon, & Layton, 1999). Behavioral excesses, suchas irritability and agitation, and behavioral defi-cits, such as poor initiation and sluggishness, arecommonly described (Prigatano, 1992; Lezak,1978). These difficulties can interfere with apatient�s participation in treatment and ability tobenefit from instruction. Service delivery is also

hampered by the paucity of controlled researchon treatment efficacy after TBI, which in turn ispartially due to the diversity of this population.Heterogeneity in impairment, intervention cur-ricula, and implementation strategies in TBI re-habilitation substantially reduce the utility of tra-ditional between-groups inferential statistics. Withthese concerns in mind for TBI as well as otherdisabling conditions, single-subject designs havebeen increasingly recommended for measuringrehabilitation processes and outcomes (Gonnella,1989; Sohlberg & Mateer, 1989; Backman, Har-ris, Chisholm, & Monette, 1997; Michaud, 1995;Ottenbacher, 1990; Mozzoni & Bailey, 1996;

Cueing and Logical Problem Solving in Brain

Trauma Rehabilitation: Frequency Patterns in

Clinician and Patient Behaviors

Charles T. Merbitz1*, Trudy K. Miller2 and Nancy K. Hansen1

1Institute of Psychoiogy, Illinois Institute of Technology, Chicago, Illinois, USA2Rehabilitation institute of Chicago, Department of Communicative Disorders, Chicago, Illinois, USA

Frequencies (count per minute) of patient and therapist behaviors during rehabilitation sessions aftertraumatic brain injury were tracked in order to evaluate an intervention curriculum and the effects of cueing.Frequencies of correct solutions to logical problems and other verbal tasks during speech-language treatmentswere measured for an adult male with memory impairments and impulsivity who underwent inpatientrehabilitation 14 months after severe traumatic brain injury, Daily frequency of cues by the clinician during thepatient�s logic exercises also was measured. These behaviors were recorded each treatment session for 14 weeks.Simple plots of the behavior frequencies were maintained by the clinician on standardized semi-log graphpaper. The patient�s performance in solving logical problems improved measurably but gradually. Day-to-daypredictability of patient performance was seen, as was predictability in cueing by the clinician. Celerations(trends measured as changes in count per minute per week) in the clinician�s cueing were inversely related tocelerations in the patient�s logical problem solving, a pattern evident through multiple reversals. Data fromother verbal performance tasks showed no general improvement, nor any pattern of variability paralleling thedata on logical problem solving. The data suggest that, for clients with brain trauma, routine continuousmeasurement of frequencies of behavior may facilitate clinical application of experimental analysis and inter-vention techniques to improve performance. Measuring the frequency of clinician behavior can help identifythe events and conditions that control aspects of patient behavior. Frequency measurement has also beenused in physical therapy and occupational therapy, and it is suggested that this method be examined as acommon data language for use across rehabilitation disciplines. Copyright © 2000 John Wiley & Sons, Ltd.

Originally Published in Behavioral Interventions, 15, 169-187(2000). Republished by Permission.

32

Merbitz, 1996; Hansen, 1999; Wilson, Emslie,Quirk, & Evans, 1999).

Clinicians who use an empirical approach (Ber-nard, 1957) implicitly use single-subject designsto manage the day-to-day flow of treatmentevents for each patient. A difficulty in using thesedesigns is that any inconsistency of performanceinterferes with assessing the effects of changes intreatment, because the logic of these designs restson predicting what the subject�s behavior wouldhave been if no changes were made in the pro-gram; thus baseline data usually are gathered untila stable level of performance can be inferred.�Precision Teaching� (Lindsley, 1992; Pennypacker,Koenig, & Lindsley, 1972; Carr & Williams,1982), a standardized behavioral measurement-and-display system for experimental analysis, fa-cilitates the detection of change in rate of change,and thus is well suited for circumstances wheresome change may be expected. This of course isthe situation found in both education and reha-bilitation, in which developmental processes,spontaneous recovery, or progressive decline cre-ate �moving baselines�.

The Precision Teaching system provides rapiddata plotting of daily behavioral frequencies andthe standardization permits quick visualization oftrends, variability, and effects of interventions.The time-line on the x-axis of the standardizedgraphs is �real time�, not �sessions�. Accurate track-ing and prediction of behavior can be used toempirically adjust ongoing treatment, thus maxi-mizing the effectiveness of treatment by titratinginterventions to produce movement toward de-sired outcomes. It also can provide a robustmethod for measuring functional change with-out resorting to ordinal instruments (Merbitz,Morris, & Grip, 1989; Merbitz, 1996).

Better reasoning, memory, and/or attentionskills are common goals for rehabilitation ofpersons with brain injury (Sohlberg & Mateer,1989; Hanlon, 1994; Williamson, Scott, & Adams,1996). Verbal exercises that require the use oflogic to find the solution to a problem are a com-mon vehicle for working with these deficits (vonCramon & Matthes-von Cramon, 1990, cited inHanlon, 1994; Wood, 1994; Williamson et al.,1996). Further, it is common for clinicians to useverbal and non-verbal cues or prompts to assist

patients in solving these problems (Wood, 1994),and a reduction in the amount of cueing as treat-ment progresses is desired. The present study is areport of findings applying the Precision Teach-ing method to interventions for one patient un-dergoing rehabilitation after severe TBI. We ap-plied this data gathering and display system toelements of speech-language treatment, whichincluded verbally mediated problem-solving ex-ercises and cueing provided by the clinician, inorder to track the interaction of patient and cli-nician behaviors throughout treatment so that thetreatment could be adjusted as needed to main-tain progress.

Method

ParticipantPatient AB81 was a 20-year-old white male

who sustained a severe TBI (closed head injury)in a motorcycle accident approximately 14months prior to the collection of these data. Hisinjuries also included fractures of the right hu-merus, left clavicle, and left wrist. He presentedwith a subdural hematoma and a right frontalcraniotomy was performed. Computerized to-mography revealed a small chronic right subdu-ral hematoma and bilateral cerebral atrophy. Atadmission to the Brain Trauma Program 12months post-trauma, his problems included lefthemiparesis, a left lower extremity contracture,aggressive behavior, a short attention span, andpoor impulse control. Neuropsychological evalu-ation suggested diffuse cortical and subcorticalimpairment. Deficits were noted throughout thebattery of cognitive functions and the prognosisfor marked improvement was guarded. Speech-language evaluation revealed moderate cognitiveand communicative deficits associated with braininjury and a mild spastic dysarthria. His medica-tion regimen included 200 mg of Tegretol TID,later increased to 300 mg.

Procedures: Speech-Language Treatment Program, and ProcessMeasurement

After initial testing, a speech-language treat-ment program was established and data werecollected on deductive reasoning, as well as cat-egorical naming and reading comprehension, in

Charles T. Merbitz, Trudy K. Miller and Nancy K. Hansen

33

individual treatment sessions with an experiencedspeech-language pathologist (MA, CCC-SP). AMasters level speech pathology student acted asthe observer for the first 40 days of data collec-tion and the clinician both recorded data andconducted the sessions thereafter.

‘Mind Benders’ Deductive Logic ExercisesThe curriculum for the deductive reasoning

therapy was based on books A1 and A2 of the�Mind Benders� series (Harnadek, 1978). Theseexercises each give several statements of facts andrelationships from which the patient is to deduceother unstated relationships by matching theknown elements from lists provided until theunknown relationships are revealed. Figure 1 il-lustrates a sample �Mind Bender�. A set of 22�Mind Benders� was selected which appeared rela-tively homogeneous in difficulty, based on theexperimenters� clinical judgement, and containedonly three or four elements from which relation-ships were to be deduced. Two exercises per daywere presented, so that the problems were re-cycled just over three times during the 95 calen-dar days reported here.

CuesWhen the patient paused or gave an incorrect

response or a partial correct response, the clini-cian provided cues or hints that were intended tohelp the patient learn how to arrive at the correct

solution. These cues could be statements thatcontradicted an incorrect inference, affirmed or�reinforced� a correct inference, identified a spe-cific fact, or modeled correct behavior. The cli-nician was sensitive to the issue of overcueingand made efforts to keep cues to a minimum allthrough the treatment. Cues were tallied as theyoccurred. When the clinician (rather than the stu-dent assistant) began counting cues after day 49,an effort was made to limit cues to a maximumof one cue per �Mind Bender� for the balance ofthe study.

Precision Teaching Behavior MeasurementDirect, objective measurement of cueing and

correct logical inference was accomplished byapplying the Precision Teaching methodology.The patient�s work on each problem was timedwith a stopwatch, and correct and incorrect state-ments of inference about the problem werecounted. Cues were also counted for each prob-lem. The count of each behavior (correct infer-ences, errors, and cues) divided by the time inminutes provided the average frequency of oc-currence for each. Average frequency (count ofbehavior divided by unit time) reflects fundamen-tal dimensional quantities of behavior (Merbitz,1996; Johnston & Pennypacker, 1980) and pro-vides a continuous, direct measure of behavior.

Average daily frequencies of all behaviorswere plotted by the clinician on Standard

Mindbenders - A1 DEDUCTIVE Thinking Skills

6.

The favorite sports of Brian, Horace, Ian, and Tammy are baseball, hiking, ice-skating, and tennis.

1. No person's name begins with the same letter as his or her favorite sport.

2. Tammy and Horace don't like team sports.

3. Horace and Brian don't like cold weather sports.

What is each person's favorite sport?

Figure 1. Sample from Harnadek (1978), adapted with permission.

Cueing and Logical Problem Solving in Brain Trauma Rehabilitation

34

Celeration Charts (SC charts), which are 8 ½ � x11� semilog graphs standardized for display ofbehavioral data (Pennypacker et al., 1972; Lindsley,1992, 1994). In Precision Teaching, behavior fre-quencies are plotted by calendar day (i.e. the �realtime� axis), and trends in frequency as the patientlearns are described by a line of best fit throughthe data points representing multiple days. Theterm �celeration� (with the prefix ac- or de- de-pending on the direction of change, defined ascount per unit time per unit time) is used to des-ignate these lines. Celerations can be fitted by aleast-squares formula or by a graphical �quarter-intercept� method with generally equal accuracy(Koenig, 1972). Celerations are easily seen on theSC chart, and the dispersion of data points arounda celeration line is a measure of variability.Clinicially, the celeration line drawn through pre-viously collected data points summarizes progressto date, while the projection of the celeration lineinto the future (across days on the SC chart) pre-dicts what progress is most likely to be seen ifno causal factors change. Then the treatment pro-gram can be maintained if predicted progress isgood, or it can be changed in an effort to im-prove upon the predicted progress. After achanged treatment is imposed and data are col-lected for several days, a clinician may draw anew celeration line covering the days under thenew treatment and compare the obtained line withthe earlier prediction.

Note that three types of celeration impliedby this logic are used, and they may be distin-guished by the span of days covered and by thetype of clinical control that is exerted. The firsttype, �overall� celerations, are fitted over the en-tire course of treatment to estimate the total ben-efit or change seen. The second type, �event-fol-lowing� celerations, begin after a known event,such as a new phase of treatment, and continueas long as that new phase is in effect, Drawingthese celerations allows the clinician to track andassess the behavior frequencies associated witheach planned change within an extended periodof treatment, for example after an adjustment inthe delivery of an intervention, or after the pa-tient moves to another environment. Finally,�trend-following� celeration, are drawn todemarcate observed changes in celeration that are

not associated with known events; this analysiscan facilitate the search for sources of influencethat the clinician may want to remove, or to rep-licate in the case of positive unexpected changesin the patient�s behavior. Thus, celeration (countsper minute per week) was the general expressionof dynamic change in behavior used in this study,and was graphically represented as a line of bestfit drawn through a given collection of frequencydata points.

All SC charts in the present study were �syn-chronized� to the same starting day, so that daynumbers correspond exactly across SC charts.Data plots were reviewed at least weekly overthe course of treatment at a routine meeting ofthe first two authors.

Results

Figure 2 shows the patient�s frequency of cor-rect inferences per minute for each of the 22problems in the set, during the first and third timeshe cycled through the set (over the course of 95calendar days). Cycle 2 fell between the othersand was omitted for clarity, and cycle 4 was notcomplete at discharge. Note that the �overall�celeration seen in cycle 1 was substantially repeatedin cycle 3, with an acceleration in frequency ofcorrect inferences from the beginning to end ofthe set in each cycle. Thus, acceleration occurredprior to the repetition of any of the Mindbenderproblems, and continued at approximately thesame rate during recycling of the set. Also notethat the patient had similar patterns of correctinferences for the same problems in cycle 1 wheneach was repeated in cycle 3, i.e. some problemsappear to have been consistently easier or harderfor him compared to others. There was also agenerally higher frequency level of correct infer-ences in cycle 3 compared to cycle 1; althoughmemorization of answers cannot be ruled out,there were at least 17 calendar days between anytwo successive presentations of a problem fromcycle 1 to 2, and at least 38 days between thepresentations of problems from cycle 1 to 3, andthis patient was assessed as having severe memorydeficits.

Figure 3 shows frequency of correct infer-ences on one of the two problems presented


35

each day of treatment plotted on a computer-drawn SC chart. (Computer-drawn charts aresubstituted for the original hand-plotted SC chartsto facilitate print reproduction, Note that overdays 65-77, the task was not presented.) Eachday�s representative performance was chosen byselecting the best performance on one of the twoproblems presented that day. The best was usedbecause theoretically the impact of most extra-neous variables, such as distracting noises fromoutside of the treatment area etc., would degraderather than enhance performance. Summary datareflecting both performances were also routinelyplotted and monitored, but are omitted fromthis report for the sake of brevity. Errors oc-curred rarely, so error frequencies are omittedfrom Figures 3 trough 5. It is conventional inPrecision Teaching to plot and show �floors� (theamount of time spent on a task) and to connectthe data points on adjacent days. These conven-tions are omitted here to reduce visual clutter.

Figure 3 is presented with �trend-following�celeration lines. These were drawn because pat-

terns of change in the patient�s frequency of cor-rect inferences were occurring in the absence ofplanned changes in the intervention. Visualizingthe pattern led to a closer inspection of data onthe cueing behavior of the clinician.

Figure 4 shows cues per minute given by theclinician, with �trend-following� celerations plot-ted as in Figure 3. Note the pattern of variabilityin weeks 2 and 5 as contrasted with weeks 1, 3and 4 in Figure 4. Week 2 in particular showspeak cue frequencies on Monday, Wednesday, andFriday. After day 49, cues were given at a maxi-mum of one per problem and a change wasmade to counting by the clinician instead of anobserver.

Figure 5 shows the superimposed �trend-fol-lowing� celeration lines of the patient�s correctinferences and the clinician�s cues, and thus re-veals the striking relationship between the twotypes of behavior. Note that the pairs of celerationlines reversed direction multiple times betweenday 1 and day 49 while maintaining the inverserelationship. Frequency of cues decreased while

Figure 2. Frequencies of correct inferences on the first and third cycles through Mindbender problems 1 - 22.


36

Figure 3. Overall and trend-following celerations on correct inferences. After day 49 (denoted by the vertical phase line) theclinician substantially reduced cueing. Note the reversals in correct celerations.

frequency of correct responses increased, and viceversa. In the two weeks immediately followingcue reduction on day 49, rate of correct infer-ences fluctuated. In the last four weeks of treat-ment, correct inferences showed a steady accel-eration while the clinician continued to restrictcueing.

Figure 6 displays data from baseline behavioron another type of verbal performance of thispatient, collected during the same speech-languagesessions. It shows frequency of correct responsesgiven by the patient during controlled oral wordassociation (�categorical naming�) in which the cli-nician gave the name of a category or class ofobjects and the patient provided as many ex-amples of class membership as possible withinthe time limit (one minute). Also shown are theerrors (category examples repeated) during thetask, and the reciprocal of the amount of timetaken for the task each session (i.e. the �floor�).

Halfway through the study, the floor was raised(i.e. the patient was timed for a shorter period)because it was suspected that performance wasaffected by fatigue and problems with sustainedattention. Performance improved slightly afterthis, with a �jump up� in frequency but no changein celeration. On two occasions, marked witharrows, the floor was dropped to test endur-ance; rate of performance dropped again dur-ing these probes. Aside from testing the effectsof a shorter versus longer timing period, no in-tervention was implemented for this task.

The data shown in this figure were examinedas a �reference behavior� for each day of the study,to examine the possibility of events producing acommon source of variability across behaviortypes, and to search for �overall� celeration trendsthat might suggest spontaneous recovery or prac-tice effects, in the absence of intervention. Vari-ability may arise after events such as medication


37

Figure 4. Trend-following celerations and cue frecuencies. After day 49 (denoted by the vertical phase line) the cliniciansubstantially reduced cueing.

adjustments, illness, or some change in the patient�senvironment that affected him or her on a givenday of days. An examination of this patient�s datain Figures 3 and 6 shows no apparent correlationin patterns of variability to suggest a commonsource of effect. Furthermore, no overall accel-eration was observed in this reference behavior,indicating that it was not affected by spontane-ous recovery or by the opportunity to repeat thetask multiple times.

Reading comprehension was targeted for in-tervention during the same time period, but noeffects of treatment were shown, and the dataare not presented here in the, interest of brevity,

Discussion

Since the present study was not a rigorouslycontrolled experiment, interpretations must bemade with caution, and, since it was a single-sub-

ject study, limits to generality must be noted.However, examination of these data reveals thatthis 1 year post-injury patient with TBI improvedhis performance on �Mind Bender� logic prob-lems. We did not examine logic in other contextsand thus have no measure of generalization ofthe effect across settings or to �real life� prob-lems. See Parente, Twum, and Zottan (1994) fora discussion of problems encountered in cogni-tive rehabilitation research related to transfer andor generalization of learning. In the present study,measurement and SC charting of frequency datafacilitated the testing of an- intervention strategy,and future research can use frequency data tocompare patients� performances with differentmaterial and in different settings, to test transferand generalization.

Three possibilities for the observed improve-ment of this patient in this setting are; (1) increasedlogical reasoning as a result of spontaneous re-


38

Figure 5. Trend-following celerations of correct inferences and cues. After day 49 the clinician substantially reduced cueing.Note the inverse relationship between cue and correct inference celerations.

covery; (2) increased logical reasoning as a resultof the intervention; and (3) practice effects spe-cific to this task and its set of problems and an-swers. While the data do not permit us to choosebetween these alternatives without ambiguity, thepatient�s performance on another verbal behav-ior (the word association task) does not indicatea benefit from spontaneous recovery or prac-tice.

Some further interpretations of these data(presented in the following paragraph) rest onthe accuracy of the celeration lines. With largerdata sets, a least-squares line of best fit can becalculated (Backman et al., 1997). When celerationlines are derived from too small a number ofdata points to perform this calculation (e.g. threeto five data points in a week), celeration lines canbe estimated. More confidence may be felt whenthe data points fit closely on a straight line thanwhen there is wide variability. �Trend-following�

celeration lines are especially problematic, sincejudgements as to the place where the trends breakmust be made. Since the �trend-following�celerations were estimated for the present data, itis suggested that readers modify any of theceleration lines, and then form alternative inter-pretations and check them against the data.

Frequency measurement provides a frame-work in which to explore the responses of, a givenpatient to a given set of antecedents and conse-quences, including patterns of cueing. With re-spect to cueing, Figures 3 through 5 showed thecelerations in the patient�s correct performancefrequency as having a generally inverse relation-ship to the celerations in clinician frequency. Thepatient-therapist dyad appeared to move as a unitin patterns that extended over multiple sessions.Two variables may fluctuate as a function ofsome third variable related to both or by coinci-dence. However, in this case the pattern of data


39

Figure 6. Standard Celeration Chart showing the frequencies of correct items given by the patient in the Categorical Namingexercise over the course of treatment. Floors are shown as dashes. Arrows indicate probe days testing the efects of returning tothe one minute performance.

on the two variables seemed to have been spe-cific to the �Mind Bender� logic exercises sincethe celerations in Figure 6 for the categorical nam-ing activity show no such relationship to eitherof the other two. It also seems unlikely that thesemultiple reversals of direction were producedby chance. Thus, the behavior does not appearto have been driven by some general third vari-able, or by random inconsistencies, but ratherseems to have been a function of events in the�Mind Benders� treatment activities. Correct in-ferences and cueing appear to have had recipro-cal effects. Modifying the celeration lines in anattempt to improve their fit to the data does notseem to substantially change the patterns of re-versals.

That the behavior of a patient with severetraumatic brain injury varied systematically ac-cording to antecedents and consequences is of

special interest given the oft-noted difficulties ofTBI patients in profiting from experience (e.g.Damasio & Anderson, 1993). The inverse rela-tionship between cues and patient�s responsessuggests that clinician cues may have controlledpatient correct responses, and/or that the patient�sresponses controlled the therapist�s cueing behav-ior. Circumstances prevented the experimentalapplication of cueing and cue reduction severaltimes to definitively address this question, but onepossible scenario for the inverse cueing-perfor-mance pattern is that if the patient is silent andlooks puzzled long enough, or if he offers anincorrect answer, a cue will be forthcoming. Inthis scenario, the patient�s silence or incorrect re-sponse is reinforced by the clinician�s cueing, whilein turn this cueing is put on a progressively leanerschedule of reinforcement by correct responses.As days progress the clinician cues more and


40

more in an effort to improve the patient�s per-formance or be more helpful, while the patientresponds correctly less and less frequently. Even-tually the clinician may recognize overcueing andreduces cues, driving the pair in the opposite di-rection.

It has been observed (Parker & Crawford,1992) that patients with severe TBI may be ableto complete complex activity only with cues ateach step. However, without measurement andanalysis of celerations for both patient and clini-cian, it seems that it would be difficult to deter-mine when overcueing occurs for a particularpatient. The results of this study suggest thatcounting and graphing the clinician�s daily cue fre-quency and patient�s behavior frequencies wouldallow the clinician to determine empirically whencueing is assisting versus hindering patient per-formance. An experimental analysis of treatmentvariants, such as cueing, with each patient couldvalidate the utility of treatment for that patient.Clinicians could then titrate their treatment toproduce the maximum sustained positive move-ment in patient achievement.

The Precision Teaching methods of data col-lection and display make routine experimentalanalysis practical and feasible; note that in thepresent study only a few minutes� of data collec-tion and plotting per day were required for stabledata and precise prediction. If changes (such asmedication changes) programmed by the mem-bers of the rehabilitation team have an effect onthe behavior in question, a change will be seen inthe data: then the clinicians can look for and con-trol events that impact that behavior. Furthermore,if several behaviors are routinely monitored si-multaneously, conditions that cause changes acrossmany behaviors can be more easily separated fromthose that perturb only one of the measured be-haviors. In the present study, the stability of thereference behavior data in Figure 6, collectedduring a different part of the therapeutic session,supported the interpretation of patient behaviorchanging in response to clinician behavior, withgradual improvement in performance on logicalproblem-solving for a selected type of task. Theimmediate implication of this study is that thecollection and display of data using the PrecisionTeaching method helped the clinician to use an

experimental analysis approach in clinical patientmanagement. In this clinical-experimental analy-sis model, multiple-baseline and other designs canbe deployed to assess treatment efficacy with eachpatient. The standardized SC chart facilitated thecomparison of data across behaviors and pro-vided a uniform framework within which thefrequency, celeration, and variability associatedwith each condition or behavior could be seen.

As those in the field of medical rehabilitationstrive toward greater accountability, their effortshave been hampered by the state of measure-ment as currently practiced. Ordinal, subjectiverating scales are virtually the sole means by whichdecisions of resource allocation are made. Foranalysis of outcomes across consumers, it wouldseem desirable to investigate the use of frequencydata, instead of ordinal data requiring extensivestatistical manipulation with the accompanyinglogical problems (Merbitz et al., 1989b; Wright,1992). In contrast to Likert-type functional as-sessment scales, such as the Functional Indepen-dence Measure (FIM, Granger, Ottenbacher, &Fiedler, 1993) or �Rehabits� (Fisher, Harvey, Tay-lor, Kilgore, & Kelly, 1995) or similar ordinallybased measures, frequency data are immediatelyavailable and do not require computerized statis-tical processing, for error-correction, utility, ormeaning. Many other measures used within reha-bilitation by individual practitioners, such as neuro-psychological tests, were not designed for thepurpose of ongoing assessment and detectionof subtle change. Frequency measurement, withits focus on testing distinct skills via rate as well asaccuracy of performance, is ideally suited forfollowing a patient�s progress over short or longerperiods of time.

The Precision Teaching method for display-ing behavioral data is usable across wide varietiesof disciplines and behaviors; note Carr and Wil-liams (1982) for an example in Physical Therapy,Binder (1981) for one in Occupational Therapy,Roth and colleagues (1990) for an explorationof gait, and Johnson and Layng (1992) for a pro-vocative example of use across several academicskills. Thus, the possibility exists for a commondata language for the teaching and therapeuticdisciplines involved in the measurement of be-havior, in a way similar to how the International


41

System of Units (SI) is used as a common datalanguage by physical scientists. Such a develop-ment could facilitate cost-benefit analysis since thebehavior is measured and quantified in physicalunits, and the charges, cost, or time of treatmentcan be known. Amount of behavior change perdollar or per unit time in treatment may be cal-culated and different therapeutic options may becompared to optimize outcomes for each pa-tient, within the course of treatment for that pa-tient. A team could experimentally analyze vari-ous approaches to individual patient issues withina relatively short period of time. In the presentstudy, the data were collected and plotted byhand, but automated methods have been de-scribed (Merbitz, Cherney, & Marqui, 1992;Merbitz, Grip, Halper, Mogil, Cherney, & Bellaire,1989; Merbitz, Cherney, & Marqui, 1991; Merbitz,King, Cherney, Marqui, Grip, & Markowitz,1992), and such methods are likely to increasethe practical appeal of Precision Teaching to re-habilitation clinicians.

As experience is accrued with frequency-basedmeasurement systems used across disciplines, thequestions of what behaviors are measured andthe extent to which those chosen are useful andempowering to the patient can be answered ex-perimentally. Hanlon (1994) has noted the rela-tively recent application of learning principles andbehavior management techniques in cognitiverehabilitation; the frequency measurement anddata display methods of Precision Teaching willfacilitate the development and testing of thesetechniques for individual patients. Of course, sub-stantial additional work remains to measure thegeneral outcomes desired in rehabilitation(Johnston, Keith, Hinderer, & Gonnella, 1992;Fuhrer, 1987; Whiteneck, 1994), but ultimately afrequency-based system with a comprehensive listof behaviors may serve as one of its founda-tions.

AcknowledgementsThis work was supported in part by Grant

#G008003034 from the National Institute forDisability and Rehabilitation Research, Depart-ment of Education, in part by Grant

#H129E30004-94 from the Rehabilitation Ser-vices Administration, and in part by the authors.

References

Backman, C. L., Harris S. R., Chisholm, J. M., &Monette, A. D. (1997.) Single-subject researchin rehabilitation: a review of studies using AB,withdrawal, multiple baseline, and alternatingtreatment designs. Archives of Physical Medicineand Rehabilitation, 78, 1145-1153.

Bernard, C. (1957/1863). An Introduction to theStudy of Experimental .Medicine. Dover: NewYork.

Binder, C. (1981) Occupational therapy with theChart: a 10-day assessment of grasp-releasepatterns with the severely handicapped. Paperpresented at the Precision Teaching Winter Confer-ence. Orlando, FL.

Carr, B. & Williams, M. (1982). Analysis of thera-peutic techniques through the use of the Stan-dard Behavior Chart. Physical Therapy, 62, 117-183.

Damasio, A. R. & Anderson, S. W. (1993). Thefrontal lobes, In Clinical Neuropsychology (3rded.), K. M. Heilman & E. Valenstein (Eds.),New York: Oxford University Press.

Dixon, T. M. & Layton, B. S. (1999). Traumaticbrain injury, In Afedical Aspects of Disability, M.G. Eisenberg, R. L. Glueckauf & H. H.Zaretsky (Eds.) (pp. 210-242). New York:Springer.

Fisher, W. P.; Harvey, R. F., Taylor, P., Kilgore, K.M., & Kelly, C. K. (1995). Rehabits: a com-mon language of functional assessment. Ar-chives of Physical Medicine and Rehabilitation, 76,113-122.

Fuhrer, M. J.(1987). Rehabilitation Outcomes: Analy-sis and Measurement. Baltimore, MD: Brookes.

Gonnella, C. (1939). Single-subject experimentalparadigm as a clinical decision tool. PhysicalTherapy, 69, 6 1-609.

Granger, C. V., Ottenbacher, K. J., & Fiedler, R.C. (1993). The Uniform Data System formedical rehabilitation: report of first admis-sions for 1993. American Journal of PhysicalMedicine and Rehabilitation, 74, 62-66.


42

Hanlon, R. (1994). Neuropsychological rehabili-tation, In Neuropsychology, D. Zaidel (Ed.), SanDiego: Academic.

Hansen, N. K. (1999, September). Selection: de-fining and supporting target behaviors. Behav-ior frequencies for process and outcome measures inSCI rehabilitation, Merbitz C (chair), Sympo-sium conducted at the annual meeting of theAmerican Association of Spinal Cord InjuryPsychologists and Social Workers, Las Vegas,NV.

Harnadek, A. (1978). Mindbenders. Pacific Grove,CA: Midwest.

Johnson, K. R., & Layng, T. J. (1992). Breakingthe structuralist barrier: fluency. American Psy-chologist, 47, 1475-1490.

Johnston, M. V., Keith, R., Hinderer, S., &Gonnella, C. (1992). Measurement standardsfor interdisciplinary medical rehabilitation.Archives of Physical Medicine and Rehabilitation, 73,S1-S23.

Johnston, J., Pennypacker, H. S. (1980). Strategiesand Tactics of Human Behavioral Research.Hillsdale, NJ: Erlbaum.

Koenig, C. H. (1972). Charting the future course ofbehavior. Dissertation, Univrsity of Kansas,Lawrence, KS.

Lezak, M. D. (1978). Living with the charactero-logically altered brain injured patient. Journalof Clinical Psychiatry, 39, 592-598.

Lindsley, O. R. (1992). Precision Teaching: dis-coveries and effects. Journal of Applied BehaviorAnalysis, 25, 51-57.

Lindsley, O. R. (1994). Performance is easy tomonitor and hard to measure. In Handbookof Human Performance Systems, R. Kaufman, S.Thiagarajan, & P. MacGillis, (Eds.). San Di-ego, CA: University Associates.

Merbitz, C. T. (1996). Frequency measures ofbehavior for assistive technology and reha-bilitation. Assistive Technology, 8, 121-130.

Merbitz, C. T., Cherney, L. R., & Marqui, H.(1991): The Communication Analysis System. StateCollege, PA: Parrot.

Merbitz, C.T., Cherney, L. R., & Marqui, H.(1992a). The Communication Analysis System.In Johns Hopkins National Search for ComputingApplications to Assist Persons with Disabilities,IEEE. Laurel, MD: Computing Society Press.

Merbitz, C., Grip, J., Halper, A., Mogil, S.,Chemey, L., & Bellaire, K. (1989a). The Com-munication Analysis System. Archives of Physi-cal Medicine and Rehabilitation, 70, 118-123.

Merbitz, C., King, R., Cherney, L., Marqui, H.,Grip, J., & Markowitz, T. (1992b). Comput-erized behavioral data collection and analysisfor improved clinical outcomes in rehabilita-tion. Behavior Research Methods, Instruments, andComputers 24: 366-372,

Merbitz, C. T., Morris J, Grip JG. 1989b, Ordi-nal scales and foundations of misinference.Archives of Physical Medicine and Rehabilitation, 70,118-123.

Michaud, L. (1995). Evaluating efficacy of reha-bilitation after pediatric traumatic brain injury.In Traumatic Head Injury in Children, S. Broman& M. Michel (Eds.) (pp. 319-345). New York:Oxford University Press.

Mozzoni, M. P. & Bailey, J. S. (1996). Improvingtraining methods in brain injury rehabilitation.Journal of Head Trauma Rehabilitation, 11, 1-18.

Ottenbacher, K. (1990). Clinically relevant designsfor rehabilitation research: the idiographicmodel, American Journal of Physical Medcine andRehabilitation, 70, 5144-5150.

Parente, R., Twum, M., & Zoltan, B. (1994). Trans-fer and generalization of cognitive skill aftertraumatic brain injury. NeuroRehabilitation, 4, 25-35.

Parker, D. M., & Crawford, J. R. (1992) Assess-ment of frontal lobe dysfunction, In A Hand-book of Neuropsychological Assessment, J. R.Crawford, D. M. Parker & W. W. McKinlay(Eds.) (pp. 267-294). Hove: Erlbaum.

Pennypacker, H. S., Koenig, C., & Lindsley, O. R.(1972). Handbook of the Standard Behavior Chart.Gainesville, FL: Precision Teaching of Florida.

Prigatano, G. P. (1992). Personality disturbancesassociated with traumatic brain injury. Journalof Consulting and Clinical Psychology, 60, 360-368.

Roth, E., Merbitz, C., Grip, J., Bogolub, M.,Mroczek, K., & Dugan, S., et al., (1990). TheTimer-Logger-Communicator gait monitor:recording temporal gait parameters using aportable computerized device. InternationalDisability Studies, 12, 10-16.


43

Sohlberg, M. M., & Mateer, C. A., (1989). Intro-duction to Cognitive Rehabilitation: Theory and Prac-tice. New York: Guilford.

Whiteneck, G. G. (1994). Measuring what mat-ters: key rehabilitation outcomes. Archives ofPhysical ,Medicine and Rehabilitation, 75, 1073-1076

Willer, B. S., Allen, K. M., Liss, M., & Zicht, M. S.(1991). Problems and coping strategies ofindividuals with traumatic brain injury and theirspouses. Archives of Physical Medicine and Reha-bilitation, 72, 460-464.

Williamson, D. J. G., Scott, J. G., & Adams, R. L.(1996). Traumatic brain injury. In Neuropsychol-ogy for Clinical Practice: Etiology, Assessment, andTreatment of Common Neurological Disorders, R.

L. Adams, O. A. Parsons, J. L. Culbertson, &S. J. Nixon (Eds.) (pp. 268-289). Washington,DC: American Psychological Association.

Wilson, B. A., Emslie, H., Quirk, K., & Evans, J.(1999). George: learning to live independentlywith NuroPage. Rehabilitation Psychology, 44,284-296.

Wood, M. M. (1994). Cognitive rehabilitationof the brain-injured patient. In Neuropsychol-ogy in Clinical Practice, S. Touyz, D. Byrne, & A.Gilandis (Eds.).(pp. 113-131). Sydney: Aca-demic.

Wright, J. (1992). Statistical considerations in theinterpretation of functional outcome scales.NeuroRehabilitation, 2, 21-30.


Documents

Cueing and Logical Problem Solving in Brain Trauma ...binde1.verio.com/wb_fluency.org/Publications/MerbitzMillerHansen... · 33 individual treatment sessions with an experienced speech-language