Acquisition of Domain-Specific Knowledgein Patients with Organic Memory Disorders

Elizabeth L. Glisky and Daniel L. Schacter

Patients with organic memory disorders are typically unable to acquire and retain newinformation and are therefore often unable to lead independent lives. The present paperoutlines an approach to memory remediation that attempts to teach memory impairedpatients domain-specific knowledge relevant to their everyday functioning. Severalstudies describe the successful use ofa training technique, the method ofvanishing cues,in teaching patients information and skills associated with the operation ofa micro-computer. Additional studies show that knowledge acquired in the laboratory can beapplied in at least one important domain of everyday life—the workplace. Suggestionsfor the use of these techniques with other cognitively disabled populations and in otherdomains are considered.

I emory disorders are among themost prevalent consequences of

various kinds of cerebral trauma (forreviews, see Cermak, 1982; Hirst, 1982;Parkin, 1987; Schacter & Crovitz,1977; Squire, 1987; Weiskrantz, 1985).Although memory impairments aretypically associated with poor per-formance on standard laboratory testsof recall and recognition, they canalso seriously affect an individual'sability to lead an independent life(Schacter, Glisky, & McGlynn, inpress). With increasing numbers ofpeople surviving closed head injuryand other brain insults that producememory problems, the need for effec-tive rehabilitation techniques has in-creased and research activity has ex-panded correspondingly. Yet memorydeficits have remained largely resis-tant to remedial efforts.

One reason for this is that mostresearch concerning remediation oforganic memory disorders has at-tempted either to restore memoryfunction through the use of repetitivedrills and exercises, or to teach pa-tients how to use mnemonic strategies(i.e., visual imagery) that can bebroadly applied in everyday life. Un-fortunately, neither of these ap-proaches has yet produced any clin-ically significant effects: The use of

repetitive drills and exercises, which isbased on the misguided notion thatmemory is like a muscle that can bestrengthened through practice (e.g.,Harris & Sunderland, 1981), has failedto provide any evidence of memoryimprovement; and attempts to teachmnemonic strategies have founderedbecause brain damaged patients haveproven unable to use such strategiesspontaneously in their daily lives (formore detailed discussion, see Glisky& Schacter, 1986; O'Connor & Cer-mak, 1987; Schacter & Glisky, 1986;Wilson, 1987).

In view of the lack of evidence forgeneral improvements of memoryfunction in various patient groups, wedecided to focus on a more limitedgoal that seemed somewhat more real-istic and perhaps useful: to help pa-tients acquire specific knowledge andskills that could enable them to func-tion more productively in a particulararea or domain of everyday life. Ac-cordingly, our approach to memoryremediation is concerned with the ac-quisition of domain-specific knowl-edge—knowledge pertaining to a spe-cific task, function, or subject that is orcould be relevant to a patient's abilityto function independently in the realworld (Schacter & Glisky, 1986). Thisapproach is based in part on recent

theoretical and empirical develop-ments concerning preserved memoryabilities in amnesic patients, whichhave provided new insights into thenature of both normal and pathologi-cal memory processes.

The main purposes of the presentpaper are to review briefly the foun-dations of our approach to memoryremediation, to describe a number ofstudies that have yielded positive out-comes both in the laboratory and inthe real world, and to consider somepromising new directions for futureinvestigations.

EMPIRICAL ANDCONCEPTUALFOUNDATIONS OF THEDOMAIN-SPECIFICAPPROACH

Preserved Learning inAmnesic Patients

In recent years, it has become clearthat memory impaired patients arecapable of acquiring new facts abouttheir everyday world. For example,patients have learned names of peoplein their environment (Dolan & Nor-ton, 1977; Glasgow, Zeiss, Barrera, &Lewinsohn, 1977; Jaffe & Katz, 1975),appropriate hospital behaviors (Do-lan & Norton, 1977; Seidel & Hodgkin-son, 1979), details about their illness(McGlynn & Schacter, in press; Wil-son, 1987), and information aboutdaily living activities (Cermak, 1976).For the most part, however, knowledgeacquired in these studies has beenrelatively simple, consisting of just afew discrete facts. More critically, inmost instances, although they wereable to leam the information, patientswere unable to apply it in their dailylives. The newly acquired informationseemed to remain isolated from therest of their world knowledge (cf.O'Connor & Cermak, 1987).

Other studies have demonstratedthat severely amnesic patients canacquire various perceptual, motor,and cognitive skills in a normal ornearly normal fashion. For example,patients have learned to performrotary pursuit tasks (Milner, Corkin, &

Volume 21. Number 6. June/July 1988 333

Teuber, 1968), read mirror-invertedtext (Cohen & Squire, 1980; Mos-covitch, 1984), and solve jigsaw puz-zles (Brooks & Baddeley, 1976) at thesame rate as normal subjects althoughthey may have no explicit recollectionof the leaming experience.

Preserved learning by amnesic pa-tients has also been demonstrated inrepetition priming tasks (e.g., Cer-mak, Talbot, Chandler, & Wolbarst,1985; Diamond & Rozin, 1984; Graf &Schacter, 1985; Jacoby & Wither-spoon, 1982; Schacter & Graf, 1986;Warrington & Weiskrantz, 1968,1974),in which performance with respect to astimulus is facilitated as a result of asingle prior exposure to the samestimulus. For example, Warringtonand Weiskrantz (1974) reported thatafter studying a list of words such aschair, flame, and so forth, amnesicpatients were as likely as normal sub-jects to complete word stems such ascha and fla with wordsfrom the studied list even though theycould not explicitly remember that thewords had been previously presented.

The foregoing findings have impor-tant implications for the remediationof memory disorders. The variousdemonstrations of preserved learningby amnesic patients suggest that itmay be possible for patients to acquireknowledge in domains relevant toeveryday life even though the knowl-edge may not be explicitly retrievable.Additionally, the results of the prim-ing studies suggest that a possiblyeffective teaching method may be onethat capitalizes on patients' positiveresponses to fragment cues. The goalof our research program was to ex-plore these possibilities—to determinewhether amnesic patients could ac-quire complex domain-specific knowl-edge that was applicable to everydaylife and to assess the effectiveness offragment cuing techniques in facilitat-ing the acquisition of that knowledge.In addition, we hoped to establishwhether such knowledge could be re-tained over long periods of time, andmost importantly, whether the knowl-edge acquired in the laboratory couldultimately be used in real-world con-texts.

Method of Vanishing Cues

Initial studies focused on the do-main of knowledge relevant to theunderstanding and operation of amicrocomputer. We selected this do-main both for its practical relevanceand for its theoretical significance.The computer has the potential toserve as a powerful compensatorydevice if memory impaired patientscan leam how to operate it (e.g.,Harris, 1984; Jones & Adam, 1979;Skilbeck, 1984; Wilson & Moffat,1984). Additionally, computer knowl-edge represents a well-specified do-main of interrelated facts and infor-mation that is well suited to the experi-mental study of complex learning (cf.Norman, 1978).

Our first study (Glisky, Schacter, &Tulving, 1986b) investigated whethermemory impaired patients could ac-quire some of the vocabulary associ-ated with the use of a microcomputer.Relatively little is known about vo-cabulary learning by amnesic patientsand the little evidence that exists isnegative (Gabrielli, Cohen, & Corkin,1983). For this reason, learning ofvocabulary seemed like an ideal taskin which to test the effectiveness of anew teaching technique. The teachingmethod that we developed was de-signed to tap patients' preserved mem-ory abilities, specifically the ability toproduce previously studied materialin response to fragment cues. Thebasic idea of this technique, which wehave called the method of vanishingcues, is to provide sufficient cue infor-mation on an initial leaming trial toensure elicitation of the correct re-sponse, and then to withdraw thatinformation gradually across leamingtrials until the target is produced in theabsence of cues.

The method of vanishing cues issimilar to a procedure that Skinner(1958) found useful in developing pro-grams for his teaching machines andto techniques that have been appliedto teach normal and autistic childrendifficult discriminations (e.g., Schreib-man & Charlop, 1981). Cue fadingmethods have also been used in cogni-

tive rehabilitation programs to facili-tate problem-solving behavior (e.g.,Ben-Yishay et al., 1985) and to teachnew behaviors to mentally handi-capped individuals (e.g., Schaefer &Martin, 1975). There is also at least onereport of the successful use of a first-letter cuing and fading technique withan amnesic Korsakoff patient wholeamed the location of his locker andthe names of two hospital staff mem-bers (Jaffe & Katz, 1975). Many ofthereports, however, are anecdotal, andadministration of the techniques haslacked systematic control.

Our study compared the method ofvanishing cues with a standard repeti-tion procedure in which definitionsand words were simply presentedrepeatedly without cues. Four memoryimpaired patients (three with closedhead injuries and one posten-cephalitic) attempted to leam thedefinitions of 30 computer-relatedvocabulary items. In the vanishingcues condition, a definition was pre-sented on Trial 1, for example, to store aprogram, and was followed by an ever-increasing fragment ofthe target word(i.e., ,y , 5a , sav ) until thesubject was able to respond correctlywith the word save. On subsequenttrials, letters were gradually with-drawn from the fragment (i.e., sav ,sa , 5 , ) until even-tually the definition was presentedalone without any letter cues. If at anytime patients were unable to producethe correct response, letters were addedone at a time until the target wasgenerated.

Results of this study demonstratedthat (a) memory impaired patientswere able to acquire 20 to 30 items ofnew vocabulary and produce themwithout letter cues, (b) leaming wasfaster by the method of vanishing cuesthan by the standard repetition pro-cedure, (c) the rate of learning formemory impaired patients was muchslower than that for normal subjects,and (d) leaming was durable with littleforgetting over a 6-week retentioninterval. This evidence for learningand retention of vocabulary is par-ticularly impressive in view of the factthat one of our patients was so severely

334 Joumal of Learning Disabilities

amnesic that he could not even re-member ever having performed thetask—yet he was able to leam andretain almost all of the vocabularyitems.

Two positive features ofthe methodof vanishing cues are noteworthy.First, the procedure permits a sys-tematic tracking of performance overtime in terms of the number of lettercues required for correct responding.In the present study, as training con-tinued, subjects needed progressivelyfewer letters to identify targets. Thissteady reduction of letter cues impliedthat learning was taking place eventhough patients were unable in theearly trials to produce the correct re-sponses when no cues were present.Second, all patients enjoyed learningby this method. Because a correct re-sponse is generated on every trial,training is a positive experience that ismuch less frustrating for patients thansimple repetition.

Having demonstrated that memoryimpaired patients could acquire thebasic vocabulary associated with amicrocomputer, we embarked on aseries of studies to explore whetherpatients could use this knowledge tolearn how to operate a microcom-puter. In all of these experiments, themethod of vanishing cues served asthe principal training technique.

Computer Learning in MemoryImpaired Patients

Three interactive computer lessonswere constructed to teach patientssome of the basic operating pro-cedures for an Apple He microcom-puter. In the initial study (Glisky,Schacter, & Tulving, 1986a) four headinjured patients with memory disor-ders of varying severity constituted theexperimental group; a matched con-trol group also participated. The threelessons were graded in complexity andrequired patients to leam and usecorrectly a total of nine different com-puter commands. For example. Lesson1 focused on the PRINT command,which is used to display information

on the computer screen, and theHOME command, which is used toclear the screen. Patients were in-structed regarding the meanings ofthese commands, and then performedvarious tasks that required use ofthem. In Lesson 2, patients had tolearn many new commands such asLIST, SAVE, CATALOG, and others;they were also introduced to the ideaof a "program" and were taught towrite brief programs that increased incomplexity as training progressed.Lesson 3 introduced additional com-mands and taught patients how to edittheir programs by adding, deleting, orchanging lines. In each of the threelessons, learning proceeded via themethod of vanishing cues. If patientsfailed to type the appropriate com-mand, they pressed the RETURN keyon the computer and were presentedwith letter-fragment cues. As in thevocabulary study, patients' progresswas charted by tracking the number ofletter cues required for successfulperformance.

All four head injured patients in ourinitial computer learning study suc-cessfully completed the three com-puter training programs. At the end ofthat time they could display messageson the screen, clear the screen, per-form disk storage and retrieval func-tions, and write and edit simple pro-grams. Even the most severely amnesicpatient, who never remembered everworking on a computer despite morethan 100 hours of training, nonethe-less was able to leam all of the basicoperations. Furthermore, the trainingtechniques seem to be broadly appli-cable. In a subsequent experiment(Glisky & Schacter, in press) similarfindings were obtained with memoryimpaired patients of other etiologies,including ruptured aneurysm andencephalitis. In addition, it was dem-onstrated that the computer knowl-edge acquired by patients in thesestudies was retained, with little evi-dence of decay, over long periods oftime (7 to 9 months).

Although patients learned to per-form all of the procedures within thecontext ofthe training programs, theirlearning was far from normal. They

required many more training trials (26to 154) to achieve criterial levels of per-formance than did controls (10 to 28)and the knowledge that they acquiredseemed to be qualitatively different.Patients' new learning could be char-acterized as "hyperspecific": it wasbound to the stimulus characteristicsof the training situation and wasrelatively inaccessible to cues otherthan those present during originalleaming. Patients had considerabledifficulty responding to open-endedquestions about what they had learnedand could not specify appropriatecomputer commands when the word-ing of instructions was altered. Con-trol subjects, on the other hand, hadno particular problems with thesetasks.

The results of these experimentsprovide some reasons for optimismconcerning possibilities for the reha-bilitation of memory disorders butthey also suggest some caveats. First,on the positive side, patients withmemory disorders of varying severityand etiology can acquire some com-plex knowledge and skills in thelaboratory, including procedures nec-essary to perform basic operations ona microcomputer. It therefore seemsreasonable to speculate that patientscould be trained to use a microcom-puter as an extemal aid outside ofthelaboratory to assist them in everydayliving. Second, the vanishing cuestechnique could be used to teachpatients complex knowledge in otherdomains (e.g., educational, vocational,domestic) that might enable them tohandle problems of daily life indepen-dently. However, the infiexibility orhyperspecificity of learning achievedby patients implies that difficulties oftransfer from the laboratory to the realworld are likely. It is important there-fore to test directly whether knowledgeacquired in the laboratory can be usedby patients in the real world and underwhat conditions. The next section de-scribes a case study in which weattempted to teach a memory im-paired patient knowledge and skillsthat had to be applied in an importantdomain of everyday life: the work-place.

Volume 21, Number 6, June/July 1988 335

VOCATIONAL TRAINING:EMPIRICAL CASE STUDIES

Computer-Related Work

Numerous studies of brain injuredpatients have cited memory impair-ment as a major obstacle to re-entryinto the work force (Bond, 1975;Bruckner & Randle, 1972; Levin,Grossman, Rose, & Teasdale, 1979;Prigatano & Fordyce, 1986; Weddell,Oddy, & Jenkins, 1980). Because oftheir inability to remember the pro-cedures required in even the simplestjobs, patients are often unable toobtain any kind of meaningful em-ployment. We recently had the oppor-tunity to investigate whether the van-ishing cues procedure could be used tohelp a patient with serious memoryproblems acquire domain-specificknowledge necessary to perform acomplex job (Glisky & Schacter,1987).

The patient, H.D., was a 32-year-oldwoman who became severely amnesicas a result of herpes simplex en-cephalitis. Following her illness, H.D.was retained by her employer in a sim-ple clerical position. When this jobwas subsequently eliminated from thecompany, however, the patient wasconsidered to be untrainable for anyof the remaining available jobs. Onour recommendation, the employeragreed to allow us to attempt to trainthis patient for a part-time job withinthe corporation.

H.D. has a severe memory disorder;her memory quotient on the WechslerMemory Scale is 65, which reflectssubstantial impairment. In additionshe has a somewhat depressed IQ (84on the Wechsler Adult IntelligenceScale-Revised) (Wechsler, 1981) and aslight dysnomia. However, she pos-sesses excellent attentional abilitiesand is strongly motivated to leam newinformation and skills. H.D. had par-ticipated in our computer leamingstudies and had demonstrated that shewas capable of learning complexmaterial using the method of vanish-ing cues; she also had very goodkeyboard skills that she had acquiredmany years earlier as a typist. For

these reasons we recommended acomputer data-entry job as being wellsuited to her capabilities.

The job required a data-entry clerkto extract information from companydocuments called "meter cards" andenter it into a computer display con-sisting of nine columns with codedheadings. Meter cards were all similarin appearance and relevant informa-tion appeared on all cards in the samelocation. However, varying amountsof irrelevant information also ap-peared on the cards and had to beignored, and the mapping betweencards and computer display was notstraightforward. For example, the lastsix numbers ofthe eight-digit SERIALNO. on the meter card had to beentered into the DOC # column ofthecomputer display, and the date wasentered in the reverse order to which itappeared on the card. The task wasthus relatively complex, but oncelearned it remained invariant acrosstrials and cards.

Performance of the job requiredleaming of (a) the general terminologyassociated with the job, (b) the mean-ing of each ofthe nine coded headingsin the computer display, (c) the docu-ment and operation codes, (d) thelocation and meaning of the informa-tion on the cards, (e) the mapping be-tween cards and display, (f) the use ofspecific function keys on the computerkeyboard, (g) the meaning of the cor-rect response to error messages, and(h) the integration of all of these com-ponents into the correct sequence. Inaddition, the actual data-entry taskhad to be performed quickly; experi-enced clerks at the company requiredapproximately 10 to 15 seconds toenter data from a single card intothe computer.

The first three phases of job trainingwere conducted in the laboratory,where we constructed a detailed sim-ulation of the real-world job. In thefirst phase of training, H.D. attemptedto leam, using the method of vanish-ing cues, the basic facts and conceptsassociated with the task. During thisknowledge acquisition phase of train-ing, 28 incomplete sentences were pre-sented on the computer screen and

H.D. was required to type the correctcompletions on the keyboard. Forexample. Information from one docu-ment is entered into a single row and iscalled a (blank) was to be completedwith the word record. Sentences de-fined terms and concepts, explainedcolumn headings, and outlined card/computer mappings. Hints were pro-vided, when needed, in the form ofinitial letters of the target word andwere gradually vanished over trials.

On the first training trial, H.D.needed 60 hints and took 55 minutes tocomplete all sentences correctly. Bythe end of the 25th trial, however, shewas performing perfectly withouthints in approximately 9 minutes. Shethen entered the skill acquisition phaseof training during which she actuallyperformed the data-entry task withreal meter cards. The purpose of thisphase of training was to improvespeed and efficiency of data entry.Across 15 training sessions (2 to 3 perweek) and 2,440 meter cards, H.D.'stime to enter data from a single carddeclined from 63 seconds to 13 sec-onds, well within the range of speedsachieved by experienced operators.The third segment of laboratory train-ing was concerned with the teachingof special procedures such as error-handling routines. Once again, H.D.showed steady learning across sixtraining sessions.

The final phase of the experimenttook place in the real world, where weassessed whether and to what extentthe knowledge and skills acquired inthe laboratory could be used on thejob. Following an initial adjustment tominor differences between the labora-tory simulation and the actual task,H.D. rapidly achieved error-free per-formance at speeds equivalent to thoseattained in the laboratory. Further-more, despite spending only 2 to 3 daysper month on the task, her speed con-tinued to improve on the job over 3months until performance asymp-toted at approximately 10 seconds percard. H.D., at the time of this writing, isstill performing this job on a regularpart-time basis at a high level ofefficiency. Interestingly, although herability to recount some aspects of the

336 Journal of Learning Disabilities

job has improved with time and expe-rience, she is still unable to reportmany details of the task despite herability to perform it flawlessly.

Extending the Domain ofVocation^ Training: AMicrofilming Job

Because of our success in trainingH.D. for the part-time data-entry joband because of her skill and efficiencyin performing the task, H.D.'s em-ployers were interested in having usteach H.D. another part-time job sothat she could eventually be employedon a full-time basis. The new jobrequired that H.D. learn to operate amicrofilming machine for the purposeof filming several thousand medicalrecords that were occupying much-needed space in filing cabinets.

This job differed from the data-entry job in that laboratory simulationof the task was impossible because ofequipment requirements. For thistask, therefore, we used labeled pic-tures of the apparatus and attemptedto teach all aspects of the proceduresin propositional form—similar to theknowledge acquisition phase of theprevious study. Forty-five incompletesentences that explained information,concepts, and operations required bythe task were presented on the com-puter screen one at a time. H.D. had totype the correct completions on thecomputer keyboard v̂ dth initial lettercues ofthe target responses being pro-vided as cues if needed. Some samplesentences are illustrated in Figure 1.

H.D. once again started slowly; sherequired 73 minutes to complete thefirst trial and needed 129 hints to pro-duce the correct responses. However,as shown in Table 1, performanceimproved steadily and after 31 trialsacross 11 training sessions, H.D.achieved error-free perfonnance in21 minutes.

At the same time as H.D. was learn-ing the facts and procedures associ-ated with the perfonnance of the mi-crofilming job, she was also involvedin another aspect of job preparationthat attempted to teach her the namesand occupations of her new co-

Picture A shows the machine that you wiil beoperating. The machine is caiied a(MiCROFILMER)Picture B is a ciose-up diagram of the

(COPYBOARD)At the right front of the copyboard, there are threeswitches or (BUTTONS)The button iabeted # 1 is caiied an(iNDICATOR LIGHT)The indicator iight is (coior)(RED)When the film has reached the end of the roll, thered indicator iight (FLASHES)Picture C is a ciose-up diagram of the

(CAMERA)The part of the camera indicated by the number 1is caiied the . (WiNDINGHANDLE)When ioading or unloading fiim from the camera,you turn the winding handie(CLOCKWISE)The number of turns you make is at least

(TEN)

Figure 1. Sample sentences from the microfilm-ing training sessions with the correct completionsshown in parentheses.

workers. Having observed in our pre-vious study that H.D. was somewhatembarrassed by her inability to re-member details about her work en-vironment, we included this segmentof training in an attempt to facilitateher social adjustment to her newsurroundings. For the same reason, weincluded three sentences in the basictraining program concerning the de-partment and building in which thejob was located.

To teach H.D. information abouteach of seven fellow employees, wepresented her with a company pho-tograph of the person, and three sen-tences on the computer screen. Sen-tences were as follows: Thefirst name ofthe person in photograph # 5 is

. When this sentence hadbeen completed conectly with thename Anne, for example, the followingsentence was displayed: ANNE'S lastname is . After conectcompletion of this sentence, the finalsentence read, ANNE BLACK is the

, which was to be com-pleted with the response nurse. Themethod of vanishing cues was againused to provide letter hints for the cor-rect responses. The procedure wasmodified slightly from its usual for-mat. To prevent random guessing of

proper names on the initial trial, com-plete names were provided on Trial 1and vanished a letter at a time on sub-sequent trials. The minimum possiblenumber of hints on the first trial wastherefore 78, which included all lettersin the names. Occupations were gen-erated as in previous experiments withletters being added on the first trialuntil a correct response was producedand then withdrawn across trials.

Results of the name/face learningstudy are presented in Table 2. Steadylearning over trials is again evidentwith hints decreasing from 94 on Trial1 to zero by the end of 30 trials. Timesdecreased conespondingly. H.D. hadsome difficulty discriminating be-tween three similar female faces,which she could not remember untilshe was given the first letter ofthe first

Table 1Number of Hints and Time per Trial to

Complete Correctly 45 SentencesConcerning the Microfilming Job

Sessions(3 trials/session) Hints Time (min)

123456789

1011

129-10997- 6452- 2823-12-7-4-2-2-2-0

13842111

73.2-41.439.8-29.240.0-34.636.7-30.529.1-27.429.0-26.628.1-24.228.4-24.624.6-22.021.8-20.321.0

Table 2Number of Hints and Time Taken toComplete 21 Sentences Concerning the

Names of Fellow Employees andTheir Jobs

Session(3 trials/session) Hints Time (min)

123456789

1011

94-6557-3932-1815-1213- 58-6-4-4-2-0

15.1-8.412.2-8.810.4-7.29.9-9.18.1-7.27.7-7.19.0-7.57.9-6.87.8-6.86.6-5.95.0

Volume 21, Number 6, June/July 1988 337

name. This pattern of responding canbe seen in Table 2 by the flat levels ofperformance in Sessions 8 and 9. Untilthat point in training, photographshad always been presented in a dif-ferent random order on each trial.During the 10th session, this pro-cedure was changed in an attempt toovercome her problems. On each ofthe three trials of Session 10, pho-tographs were displayed on the tablein a hierarchical order according tojob description. This organizationalframework appeared to eliminate theconfusion that she had been experi-encing. In Session 11, photographswere again presented in a randomorder, and performance was perfect.

It remained to be demonstratedwhether such laboratory trainingwould benefit on-the-job performancegiven that the actual job had neverbeen performed except in the labora-tory. Because the laboratory and real-world tasks were so different, we werenot able to obtain direct measures oftransfer. However, we provided twodays of on-the-job practice on the mi-crofilming task and then monitoredH.D.'s performance continuously for 6days and twice a week for 2 weeksthereafter. Our observations indicatedthat the patient was initially confusedby the actual job situation and did notrecognize various aspects ofthe task asbeing related to those learned in thelaboratory. Forexample, she appearednot to be familiar with the microfilm-ing equipment although it was identi-cal to the apparatus in the photo-graphs with which she had beentrained. In addition, many features ofthe job were new to H.D.: the locationof her work space, the proceduresinvolved in removing and replacingfiles from the cabinets, some record-keeping chores, and so on.

However, she appeared to learn thetask relatively quickly. Across the first6 days, H.D.'s filming speed increasedfrom approximately 12.3 seconds to7.3 seconds per document. By thistime, performance appeared to berelatively automatic and the patientwas filming approximately 2,500 rec-ords per day, a number that the com-pany considered to be more than

satisfactory. She also learned to mon-itor for the end-of-film signal and toload and unload film from the camerawithout help. Although we could notbe certain that on-the-job perform-ance was enhanced by prior labora-tory training (no control conditionswere possible), the fact that shelearned the task so readily is stronglysuggestive of that conclusion.

actual job situation as closely as possi-ble in order to minimize transfer prob-lems. Although much has yet to bediscovered about the factors affectingtransfer of training, we know thatmemory impaired patients experienceserious difficulties when called uponto apply information learned in onesetting to similar problems that occurin a different setting.

Summary of Training Studies CONCLUDING COMMENTS

The two foregoing studies demon-strate that it is possible to teach apatient with a severe memory disorderdifferent kinds of complex domain-specific knowledge that can be appliedin an important segment of everydaylife. Although we have demonstratedsuccess with two different vocationaltasks, we have as yet trained only asingle patient. Generalizability of ourfindings to other patient groups, there-fore, is uncertain until further inves-tigations are conducted. On the basisof our research to date, however, wehave reached some tentative con-clusions concerning the types of jobsthat are most likely to be appropriatefor memory impaired patients andsome of the potential pitfalls thatmight be encountered in vocationalcontexts.

First, suitable jobs are likely to bethose that require a set of relativelyinvariant procedures. Patients appearable to leam complex tasks but areunable to cope with novel situations.Any variations on basic proceduresthus disrupt performance unless themeans of handling such variations aretaught specifically. Second, it appearsthat complex tasks can be most easilylearned if they are broken down intocomponent steps with each compo-nent being taught explicitly and dir-ectly. This approach is well docu-mented as an effective instructionaltechnique with learning disabled indi-viduals (e.g.. Smith & Robinson, 1986)and similar task analysis approacheshave been used in vocational trainingof severely retarded persons (e.g., Mit-haug, 1979; Wacker & Hoffman, 1984;Wehman & Hill, 1981). Third, lab-oratory training should mimic the

The research reported in this paperrepresents an initial attempt to exploresome of the conditions under whichpatients with memory disorders canacquire complex knowledge and skillsin domains relevant to everyday life.Only a few domains have been inves-tigated as yet and those only with a fewpatients. Nevertheless, the results areencouraging. In the domains that wehave studied, we have not yet observedany limit on the amount or complexityof knowledge that can be acquired byamnesic patients. It is conceivable thatalthough the learning process is slow,a great deal of knowledge can beacquired and retained by memoryimpaired patients that could signifi-cantly improve their ability to func-tion independently in the real world.

Many directions for further researchcan be envisioned. Other domains ofknowledge need to be explored—vocational, educational, social, and soon, and opportunities for non-sheltered employment might be con-sidered for individuals with otherkinds of handicaps (cf. Brown et al.,1986). The method of vanishing cueshas proven to be an extremely effectivetechnique for teaching new informa-tion to patients with memory deficits.Other intellectually or cognitively dis-abled populations, including childrenwith cerebral trauma, may benefitsimilarly.

Problems involved in transfer willcontinue to occupy investigators con-cerned with training handicappedindividuals to function in the realworld. Until more is known about thefactors affecting generalization to newcontexts, the methods of training thatseem most likely to be successful are

338 Journal of Learning Disabilities

those that focus on teaching each skill,fact, or procedure directly and expli-citly in a context as similar as possibleto that in which the information willultimately be used. A major challengefor research will be to discover meth-ods by which such learning can bemade more flexible and thereby moreuseful in the endless variety of situ-ations presented in everyday life.

ABOUT THE AUTHORS

Elizabeth L. Glisky is a visiting assistant professorin the Department of Psychology at the University ofArizona. Daniel L. Schacter is an associate pro-fessor of psychology at the University of Arizona.Both authors received their PhDs in experimentalpsychology from the University of Toronto wherethey were subsequently associated with the Unit forMemory Disorders. They are currently conductingresearch at the Amnesia and Cognition Unit,University of Arizona, concerning the nature oforganic memory disorders and possibilities forrehabilitation. Address: Elizabeth L. Glisky. De-partment of Psychology. University of Arizona. Tuc-son. AZ 85721.

REFERENCES

Ben-Yishay. Y.. Rattock. J.. Lakin. P., Piasetsky,E.G.. Ross. B.. Silver. S.. Zide. E.. & Ezrachi. O.(1985). Neuropsychological rehabilitation: Questfor a holistic approach. Seminars in Neurology, 5,252-259.

Bond. M.R. (1975). Assessment of the psychosocialoutcome afier severe head injury. In Outcome ofsevere damage to the C.NS. Symposium 34(pp. 141-153). London: Ciba Foundation.

Brooks. D.N.. & Baddeley. A.D. (1976). What canamnesic patients learn? Neuropsychologia. 14,111-122.

Brown. L.. Shiraga. B., Eord. A.. Nisbet. J.. Van-Deventer. P.. Sweet. M.. York. J.. & Loomis, R(1986). Teaching severely handicapped studentsto perform meaningful work in nonshelteredvocational environments. In R.J. Morris & B. Blatt(Eds.). Special education: Research and trends(pp. 131-189). New York: Pergamon Press.

Bruckner. EE., & Randle, A.P.H (1972). Return towork after severe head injuries. Rheumatologyand Physical Medicine. 11, 344-348.

Cermak. L.S. (1976). The encoding capacity of apatient with amnesia due to encephalitis. Neuro-psychologia. 14. 311-322.

Cermak. L.S. (Ed.). (1982). Human memory andamnesia. Hillsdale. NJ: Erlbaum.

Cermak. L.S., Talbot, N. Chandler. K.. & Wotbarst.L.R. (1985). The perceptual priming phenomenonin amnesia. Neuropsychologia. 23. 615-622.

Cohen. N.J.. & Squire, LR. (1980). Preserved learn-ing and retention of pattern-analyzing skill inanmesia: Dissociation of "knowing how" and"knowing that." Science. 210. 207-209.

Diamond. R.. & Rozin. P. (1984). Activation of exist-

ing memories in the amnesic syndrome. Journalof Abnormal Psychology. 93, 98-105.

Dolan, M.P.. & Norton, J.C. (1977). A programmedtraining technique that uses reinforcement tofacilitate acquisition and retention in brain-damaged patients. Journal of Clinical Psychol-ogy, 33. 496-501.

Gabrielli. J.D.E.. Cohen. NJ, & Corkin. S (1983).The acquisition of lexical and semantic knowl-edge in amnesia. Society for Neuroscience Ab-stracts, 9. 238.

Glasgow. R.E., Zeiss. RA.. Barrera, M., & Lewin-sohn. P.M. (1977). Case studies on remediatingmemory deficits in brain-damaged individuals.Journat of Clinical Psychology. 33. 1049-1054.

Glisky, E.L.. & Schacter. D.L. (1986). Remediationof organic memory disorders: Current status andfuture prospects. Journal of Head TraumaRehabilitation. 1(3), 54-63.

Glisky, E.L.. & Schacter. D.L. (1987). Acquisition ofdomain-specific knowledge in organic amnesia:Training for computer-related work. Neuro-psychologia. 25. 893-906.

Glisky. E.L., & Schacter, D.L. (in press-b). Long-term retention of computer learning by patientswith memory disorders. Neuropsychologia.

Glisky. E.L.. Schacter, D.L, & Tulving E. (1986a).Computer learning by memory-impaired pa-tients: Acquisition and retention of complexknowledge. Neuropsychologia. 24, 313-328.

Glisky. E.L.. Schacter. D.L. & Tulving E. (1986b).Learning and retention of computer-relatedvocabulary in memory-impaired patients: Meth-od of vanishing cues. Journal of Clinical andExperimentl Neuropsychology. 8, 292-312.

Graf P.. & Schacter, D.L. (1985). Implicit andexplicit memory for new associations in normaland amnesic subjects. Journal of ExperimentalPsychology: Learning Memory, and Cognition.11. 501-518

Harris. J. (1984). Methods of improving memory. InBA. Wilson & N. Moffat (Eds.). Clinical manage-ment of memory problems (pp. 46-62). Rockville.MD: Aspen Systems.

Harris. J.E.. & Sunderland. A. (1981). A brief surveyof the management of memory disorders inrehabilitation units in Britain. InternationalRehabilitation Medicine, 3. 206-209.

Hirst. W. (1982). The amnesic syndrome: Descrip-tions and explanations. Psychological Bulletin,91, 435-460.

Jacoby, L.L.. & Witherspoon. D. (1982). Remember-ing without awareness. Canadian Journal of Psy-chology. 36, 300-324.

Jaffe. P.G.. & Katz, A.N. (1975). Attenuatinganterograde amnesia in Korsakoffs psychosis.Journal of Abnormal Psychology. 34. 559-562.

Jones. G.H. &Adam. J.H. (1979). Towards a pros-thetic memory. Bulletin of the British Psychologi-cal Society. 32. 165-167.

Levin. H.S., Grossman, R.G.. Rose, J.E.. & Teasdale.G. (1979). Long-term neuropsychological out-come of closed head injury. Journal of neuro-surgery. 50. 412-422.

McGlynn, S.M.. & Schacter. D.L (in press). Un-awareness of deficits in neuropsychological syn-dromes. Journal of Clinical and ExperimentalNeuropsychology.

Milner. B., Corkin. S. & Teuber. H.L (1968).Eurther analysis of the hippocampal amnesicsyndrome: 14 year follow-up study of H.M.Neuropsychologia. 6. 215-234.

Mithaug D. (1979). The relation between pro-grammed instruction and task analysis in the pre-vocational training of severely and profoundlyhandicapped persons. American Association forthe Education ofthe Severely/Profoundly Handi-capped Review. 4. 162-178.

Moscovitch. M. (1984). The sufficient conditions fordemonstrating preserved memory in amnesia: Atask analysis. In LR. Squire & N. Butters (Eds.).Neuropsychology of memory (pp. 104-114). NewYork: Guilford Press.

Norman. DA. (1978). Notes toward a theory of com-plex learning. In A.M. Lesgold. J. Pellegrino, S.Fokkema. & R. Glaser (Eds.). Cognitive psychol-ogy and instruction (pp. 39-48). New York:Plenum Press.

0 Conner. M.. & Cermak. LS. (1987). Rehabilita-tion of organic memory disorders. In MJ. Meier.A. Benton. & L Diller (Eds.). Neuropsychologicalrehabilitation (pp. 260-279). New York: Guil-ford Press.

Parkin. A.J. (1987). Memory and amnesia. London:Basil Blackwell.

Prigatano. G.P. & Fordyce. D.J (1986). Cognitivedysfunction and psychosocial adjustment afierbrain injury. In G.P. Prigatano and Others.Neuropsychological rehabilitation afier braininjury (pp. 1-17). Baltimore: Johns HopkinsUniversity Press.

Schacter. D.L. & Crovitz. HF. (1977). Memoryfunction afier closed head injury: A review ofthequantitative research. Cortex. 13, 150-176.

Schacter, D.L. & Glisky. E.L (1986). Memoryremediation: Restoration, alleviation, and theacquisition of domain-specific knowledge. In B.Uzzell & Y. Gross (Eds.). Clinical neuropsychol-ogy of intervention (pp. 257-282). Boston: Mar-tinus Nijhoff

Schacter. D.L. Glisky. E.L. & McGlynn. S.M. (inpress). Impact of memory disorder on everydaylife: Awareness of deficits and return to work. InD. Tupper & K. Cicerone (Eds.), The neuro-psychology of everyday life. Volume I Theoriesand basic competencies. Boston: MartinusNijhoff.

Schacter, D.L. & Graf P. (1986). Preserved learningin amnesic patients: Perspectivesfrom research ondirect priming. Journal of Clinical and Exper-imental Neuropsychology. 8. 727-743.

Schaefer. H.H. & Martin. P.L (1975). Teaching newbehaviors to the mentally retarded. In H.H.Schaefer & P.L. Martin (Eds.), Behavioral ther-apy (pp. 242-253). New York: McGraw-Hill.

Schreibman, L, & Charlop. MJ. (1981). S+ versusS— fading in prompting procedures with autisticchildren. Journal of Experimental Child Psychol-ogy. 31. 508-520.

Seidel. H, & Hodgkinson. P.E. (1979). Behaviormodification and long-term learning in Kor-sakoffs psychosis. Nursing Times, 75, 1855-1857.

Skilbeck. C. (1984). Computer assistance in themanagement of memory and cognitive impair-

(continued on p. 351)

Volume 21. Number 6. June/July 1988 339

Journal of Disorders of Communication. 20.289-296.

Sarno. M.T. (1980). The nature of verbal impair-ment after closed head injury. Journal of Nervousand Mental Disease. 168. 685-692.

Sarno. M.T. (1984). Verbal impairment after closedhead injury: Report ofa replication study. Jour-nal of Nervous and Mental Disease. 172, 475-479.

Sarno. M.T. Buonaguro. A.. & Levita. E. (1986).Characteristics of verbal impairment in closedhead injured patients. Archives of PhysicalMedicine and Rehabilitation. 67. 450-455.

Schiller. F. (1947). Aphasia studied in patients withmissle wounds. Journal of Neurology, Neuro-surgery. and Psychiatry. 10. 183-197.

Scherzer. B. (1986). Rehabilitation following severehead trauma: Results of a three year program.Archives of Physical Medicine and Rehabilita-tion, 67. 366-373.

Sklar, M. (1966). Sklar aphasia scale: Protocolbooklet. Beverly Hills. CA: Western PsychologicalServices.

Smith. R. (1984). Treatment of communication dis-orders. In M. Rosenthal E. Griffith, M. Bond. &JMiller (Eds.). Rehabilitation ofthe head injuredadult (pp. 355-366). Philadelphia: FA. Davis.

Spreen, 0., & Benton, A.L. (1977). Neurosensory

center comprehensive examination for aphasia(rev. ed.). Victoria, BC: University of Victoria.Neuropsychology Laboratory.

Szekeres, S.F.. Ylvisaker. M.. & Holland. A.L.(1985). Cognitive rehabilitation therapy: A frame-work for intervention. In M. Ylvisaker (Ed.), Headinjury rehabilitation: Children and adolescents(pp. 219-246). San Diego: College-Hill Press.

Taylor, M. (1963). A measurement of functionalcommunication in aphasia. Archives of PhysicalMedicine and Rehabilitation, 46, 101-107.

Thomsen. I (1975). Evaluation and outcome ofaphasia in patients with severe closed headtrauma. Journal of Neurology. Neurosurgery. andPsychiatry. 38. 713-718.

Thomsen, I. (1981). Neuropsychological treatmentand longtime follow-up in an aphasic patientwith very severe head trauma. Journal of ClinicalNeuropsychology, 3, 43-51.

Thomsen, I. (1984). Late outcome of very severeblunt head trauma: A 10-15 year second follow-up. Journal of Neurology, Neurosurgery, and Psy-chiatry, 47, 260-268.

Vogel. Af., & von Cramon, D. (1983). Articulatoryrecovery after traumatic mutism. Folia Phoniat-rica. 35, 294-309.

von Cramon, D. (1981). Traumatic mutism and thesubsequent reorganization of speech functions.

Neuropsychologia. 19. 801-805.Wechsler, D. (1948). Wechsler memory scale. New

York: Psychological Corp.Wepman, J. (1951). Recovery from aphasia. New

York: Ronald Press.Wertz, R. (1978). Neuropathologies of speech and

language: An introduction to patient manage-ment. In D.F. Johns (Ed.), Clinical managementof neurogenic communicative disorders (pp. 1-101). Boston: Little, Brown.

Ylvisaker. M.. & Szekeres, S. (1986). Management ofthe patient with closed head injury. In R. Chapey(Ed.), Language intervention strategies in adultaphasia (pp. 474-490). Baltimore: Williams &Wilkins.

Yorkston. K.. & Beukelman, D. (1981). Assessmentof intelligibility of dysarthric speech. Tigard, OR:C.C. Publications.

Yorkston, K., Stanton. K., & Beukelman. D. (1981).Language-based compensatory training forclosed head injured patients. In R. Brookshire(Ed.), Proceedings of the Clinical AphasiologyConference (pp. 293-299). Minneapolis: BRKPublishers.

Zazula, T. (1984). Perception and production ofintonation in moderate and severe head injuries.Dissertation Abstracts International, 45, 3648.

Continued from p. 332

Wilson & N. Moffatt (Eds.), Clinical managementof memory problems (pp. 89-112). London:Croom Helm.

Wood, R.L. (1986). Rehabilitation of patients withdisorders of attention. Journal of Head Trauma

Rehabilitation, 1(3), 43-54.Wood, R.L (1987). Brain injury rehabilitation: A

neurobehavioral approach. London: CroomHelm.

Wood. R.L.. & Fussey. L (1987). Computer-based

cognitive retraining: A controlled study. Inter-national Disability Studies (formerly Inter-national Journal of Rehabilitation Medicine). 9.149-153.

Continued from p. 339

ment. In BA. Wtlson & N Moffat (Eds.). Clinicalmanagement of memory problems (pp. 112-133).Rockville. MD: Aspen Systems.

Skinner. B.F. (1958). Teaching machines. Science,128, 969-977.

Smith, D.D.. & Robinson. S. (1986). Educating thelearning disabled. In R.J. Morris & B. Blatt (Eds.),Special education: Research and trends (pp. 222-248). New York: Pergamon Press.

Squire, L.R (1987). Memory and brain. New York:Oxford University Press.

Wacker. D.P.. & Hoffman, R.C. (1984). Vocationalrehabilitation of severely handicapped persons. InC.J. Golden (Ed.), Current topics in rehabilitationpsychology (pp. 139-171). Orlando, Fl: Grune &Stratton.

Warrington, E.K., & Weiskrantz. L (1968). Newmethod of testing long-term retention with specialreference to amnesic patients. Nature, 217, 972-974.

Warrington. E.K.. & Weiskrantz, L. (1974). Theeffect of prior learning on subsequent retention inamnesic patients. Neuropsychologia. 12, 419-428.

Wechsler, D. (1981). Wechsler adult intelligencescale-Revised (Manual). New York: PsychologicalCorp.

Weddell, R., Oddy, M., & Jenkins. D. (1980). Socialadjustment after rehabilitation: A two yearfollow-up of patients with severe head injury. Psy-chological Medicine. 10, 257-263.

Wehman, P., & Hill, J.W. (1981). Competitiveemployment for moderately and severely handi-

capped individuals. Exceptional Children, 47,338-345.

Weiskrantz, L (1985). On issues and theories ofthehuman amnesic syndrome. In N. Weinberger, J.McGaugh. & G. Lynch (Eds.), Memory systems ofthe brain: Animal and human cognitive processes(pp. 380-415). New York: Guilford Press.

Wtlson, B. (1987). Rehabilitation of memory. NewYork: Guilford Press.

Wtlson. B., & Moffat, N. (1984). Rehabilitation ofmemory for everyday life. In J.E. Harris & P.Morris (Eds.), Everyday memory: Actions andabsent-mindedness (pp. 207-233). London: Aca-demic Press.

Volume 21. Number 6. June/July 1988 351