Journal of Autism and Developmental Disorders, Vol. 20, No. 1, 1990

Sustained A t t e n t i o n in Children with A u t i s m 1 Helen Bray Garretson 2

Hillcrest Educational Center Deborah Fein Boston University School of Medicine and University of Connecticut Lynn Waterhouse Trenton State College

Although many children with early infantile autism cannot maintain attention to externally imposed tasks, they may continue a repetitive behavior o f their own choosing f o r long periods o f time. This study examined the performance o f autistic and mental age matched normal children on a Continuous Performance Test o f sustained attention. Results suggest that autistic children's difficulties in sustaining attention on imposed tasks may be attributable partly to a developmental delay and partly to the motivational contingencies o f task rather than to a primary impairment in the ability to sustain attention.

A l t h o u g h m a n y c h i l d r e n with e a r l y i n f a n t i l e a u t i s m c a n n o t flexibly a l l o c a t e a n d m a i n t a i n a t t e n t i o n to e x t e r n a l l y i m p o s e d t a s k s f o r even a few m i n u t e s , t h e y m a y b e a b l e to c o n t i n u e a r e p e t i t i v e b e h a v i o r a l r o u t i n e o f t h e i r o w n c h o o s i n g f o r l o n g p e r i o d s o f time. D o e s this e r r a t i c a b i l i t y t o s u s t a i n a t t e n t i o n r e p r e s e n t a p r i m a r y i m p a i r m e n t o f a t t e n t i o n in a u t i s t i c c h i l d r e n o r d o e s it d e p e n d m o r e o n m o t i v a t i o n a l c o n t i n g e n c i e s o r specific t a s k d e m a n d s ? A t least t h r e e h y p o t h e s e s ( n o t m u t u a l l y exclusive) c a n be f o r m u l a t e d :

~This project was supported by an NIMH grant MH 28605 to L. Waterhouse and D. Fein. We appreciate the cooperation of the staff and students of League School, Newton, Massachusetts, and the Nazarene Child Care Center in Beverly, Massachusetts. We gratefully acknowledge the help of Hope Trefry, Dot Lucci, Kaye Cook, and Kathleen Sheher with data gathering and analysis. Ross Rizley, Anne Copeland, Jaqueline Liederman, and Leslie Brody of the Boston University Graduate Program in Psychology provided helpful comments at every stage of this project. 2Address all correspondence to Helen Bray Garretson, Hillcrest Educational Center, 370 North Street, Pittsfield, Massachusetts 01201. 101 0162-3257/90/0300-0101506.00/0 9 1990 Plenum Publishing Corporation

102

Garretson, Fein, and Waterhouse

1. The usual externally imposed tasks are too complex and changing for autistic children who tend to prefer simple, repetitive pursuits. Research with a variety of tasks has documented the inability of autistic children to manage complex tasks. Children with autism are "overselective" in that they respond to limited discriminative cues (Cook, Anderson, & Rincover, 1982; Fein, Waterhouse, & Tinder, 1979; Lovaas, Koegel, & Schreibman, 1979; Wilhelm & Lovaas, 1976), with little flexibility in their scanning to detect other important information, especially on complicated tasks. Some theorists have suggested that an underlying state of hyperarousal leads to the constricted behavioral repertoire and the preoccupations with simple, repetitive stimuli (Hutt & Hutt, 1968). Stereotyped movements may serve as arousal displacement activities to modulate tension (Kinsbourne, 1980). A task of sustained attention such as the Continuous Performance Test (CPT; Rosvoid, Mirsky, Sarason, Bransome, & Beck, 1956) does not require the subject to shift strategies; rather it tests the child's ability to maintain a single strategy over time. Frankel, Freeman, Ritvo, and Pardo (1978) have suggested that a repetitive stimulus may be inherently reinforcing for children with autism. If the everyday difficulty getting autistic children to maintain attention on tasks is due to the complex and changing nature of these tasks, autistic children might be expected to perform within normal limits on a simple, continuous performance task. 2. Autistic children's motivation for mastery or for social reward is insufficient to maintain performance. Motivational abnormalities in autism have been studied primarily to determine what aspects of the learning and/or testing situation improve task performance. The bulk of behavior modification research and clinical techniques (Ferster, 1961; Lovaas, 1977) emphasizes that strong primary reinforcers (Dunlap & Egel, 1982), sensory stimulation (Hung, 1978; Murphy, 1982), or idiosyncratic preferred reinforcers facilitate task-relevant learning in autistic children (Koegel & Egel, 1979). Social reinforcers such as smiles of frowns, do not as predictably elicit the expected results (Howlin, 1978; Wing, 1978) although some children with autism are semitive to interpersonal manipulations of the experimental situation. Clark and Rutter (1981) demonstrated that clear structure and high interpersonal demand increased attention and performance on a model-building task. Most cognitive research with autistic children does not consider that the reinforcement contingencies of the learning situation may significantly affect their performance. Reinforcement parameters are critical to evaluate because motivation is an intrinsic and crucial component of attention, and what may appear as a cognitive or attentional deficit could be related rather to motivational deficits (Kinsbourne, 1983). If motivational deficits are key to the autistic child's difficulty with everyday tasks, then varying incentives on the CPT should alter performance, and the proper incentive conditions might bring performance within normal limits.

Sustained Attention in Autistic Children

103

3. Autistic children have a deficit in maintaining attention on externally imposed stimuli over and above their motivational state and their preferencefor simple stimuli. This is not a single hypothesis but a group of related ideas suggesting fundamental perceptual and attentional anomalies (Cohen & Johnson, 1967; Gold & Gold, 1975; Ornitz & Ritvo, 1976; Rimland, 1964; Schopler, 1965, 1966) affecting the taking in of stimuli from the external environment. If there exists a primary abnormality in such a system, one might expect that although simple stimuli and effective incentives would maximize performance on a signal detection task such as the CPT, such performance would remain defective. In the current study, a modified version of the CPT was run with 23 children with autism and 23 normal children, manipulating rate of presentation (slow vs. fast, reflecting task difficulty) and reinforcement modality (social vs. tangible, reflecting motivational differences). If the third hypothesis which suggests fundamental perceptual or attentional abnormalities is correct, one would expect impaired performance on the CPT even under conditions of high motivation and simple repetitive stimuli and responses. If task complexity underlies the autistic child's impairment in sustained attention, performance on a simple CPT might be most affected by the rate of presention, while if motivational variables are crucial, the reinforcement modality might be the critical factor.

METHODS

Subject Population Subjects for this study were 23 children who attended a day school for children with autism. A review of the children's school and medical records indicated that all of the children had demonstrated symptoms of early infantile autism before the age of 36 months. Data were collected before the publication of DSM-III-R but behavioral data were available from a Behavior Rating Scale (Fein, Walerhouse, Lucci, & Snyder, 1985) filled out by teachers. Each student was evaluated for the presence of behaviors that matched specific DSM-III-R diagnostic criteria. See Table I for examples of common behaviors reported and their corresponding DSM-III-R criteria. All children met DSM-III-R diagnostic criteria for 299.00 Autistic Disorder. The children ranged in chronological age from 4 years 11 months to 19 years 2 months (M = 12 years 4 months, SD = 3 years 10 months). The sample included 5 girls and 18 boys, 5 nonwhite subjects, and a wide SES range. Children were excluded from the study if they had no comprehensible language or if they utilized sign language as their primary means of communication. Mental age on the Peabody Picture Vocabulary Test (PPVT) ranged

104

Garrelson, Fein, and Waterhouse

Table I. Examples of Behavior Rating Scale Items Used to Match Diagnostic Criteria DSM-III-R criterion Rating scale item

A1. Marked lack of awarness of the existence or feelings of others A3. No or impaired imitation B4. Marked abnormalities in the production of speech B5. Marked abnormalities in the form or content of speech including sterotyped and repetitive use of speech, use of "you" when "I" is meant, ideosyncratic use of words or phrases C1. Stereotyped body movements C4. Unreasonable insistence on following routines in precise detail C5. Markedly restricted range of interests and a preoccupation with one narrow interest

Has the child seemed "hard to reach" or "in a shell?" Has the child ignored people as if they did not exist? Has the child imitated other children at play? Has the child used hollow sounding speech? Has the child used words which mean something only to him/her?

Has the child shown repetitive hand or arm flapping, finger posturing, or clapping? Has the child insisted things must be done the same way each time? Has the child been preoocupied with certain objects or topics?

f r o m 3 years 2 m o n t h s to 10 years 2 m o n t h s ( M = 5 years 8 m o n t h s , SD -- 1 year 6 m o n t h s ) . M e n t a l age o n the D r a w - A - D e s i g n Test ( D A D ) f r o m the M c C a r t h y Scales of Children's Abilities (McCarthy, 1972) ranged from 3 years 6 m o n t h s to 9 years 6 m o n t h s ( M = 6 years, SD -- 2 years 1 m o n t h ) . It is evident t h a t these autistic c h i l d r e n r e p r e s e n t e d a wide range o f r e t a r d a t i o n . P e a b o d y IQs r a n g e d f r o m b e l o w 40 to 90 with M = 51.8, SD = 15.72. N o r m a l c o n t r o l children, recruited t h r o u g h a day care center, a h o u s ing project, a n d c h u r c h b u l l e t i n b o a r d , were m a t c h e d o n the basis o f sex a n d m e n t a l age. T h e n o r m a l s a m p l e i n c l u d e d two n o n w h i t e c h i l d r e n a n d represented a wide SES range. T w o c o n t r o l g r o u p s were f o r m e d , o n e o n the basis o f P P V T scores ( P P V T n o r m a l s ) a n d o n e o n the basis o f D A D scores ( D A D normals). F i n d i n g s were c o m p a r a b l e for b o t h c o n t r o l groups; only the c o m p a r i s o n s with the D A D g r o u p are r e p o r t e d here. N o r m a l s r a n g e d f r o m 3 years to 9 years 6 m o n t h s o n the D A D ( M = 5 years 9 m o n t h s , S D = 1 year 11 m o n t h s ) a n d were m a t c h e d case b y case to the autistic children, w i t h i n 2 raw score p o i n t s , r e p r e s e n t i n g a m a x i m u m o f 1 year d i f f e r e n c e in m e n t a l age e q u i v a l e n t s .

Testing A p p a r a t u s T h e C o n t i n u o u s P e r f o r m a n c e Test o f R o s v o l d et al. (1956), which consists o f letters r a p i d l y p r e s e n t e d over a n a p p r e c i a b l e i n t e r v a l o f time, was

Sustained Attention in Autistic Children

105

modified to use more easily discriminable and familiar stimuli. The stimulus material consisted of 23 simple black line drawings of common objects and animals from the Peabody Picture Vocabulary Test. These were filmed in a random but predetermined order of presentation in three epochs of 4 min each (Slow1, Fast, Slow2) on Super-8 films. In order to be sure that the pictures were discriminable for the children, only pictures on an earlier test of the PPVT were included. Pictures that were conceptually unclear, out of date, visually complicated, representing verbs, or which included human or animal body parts were excluded from the stimulus list. The chair was chosen as the target stimulus because it represented an emotionally neutral familiar object of medium figure complexity. Each testing session included three epochs, varying in rate of presentation (Slow1, Fast, Slow2). The slow version of the task introduced a stimulus every 3 sec for a duration of 1.5 sec (27 frames) followed by 1.5 sec (27 frames) of white space. The fast version introduced a stimulus picture every second for a duration of 0.7 sec (13 single frames) followed by 0.3 sec (5 frames) of white space. Films rather than slides were used to avoid the audible click of the changing side which may have served an arousing or cuing function. The response apparatus was a button-press with an audible click mounted on a board and placed in front of the child. The response was recorded by an ongoing event recorder in the next room, which noted onset, duration, and termination of key pressing.

Procedures The original testing design of the research called for four sessions reinforced in a Social, Tangible, Tangible, Social sequence to control for order and practice effects. Data on all four conditions were gathered for children with autism. The children without autism found the task boring to the point of being aversive and so many normal children were unwilling to complete the third or fourth session that it proved impossible to gather enough data for analysis. T-test comparisons of the minute-by-minute scores between the two social and two tangible sessions indicated that there were no significant performance differences for the autistic children between comparably reinforced minutes, suggesting that differences in performance between the social and tangible condition are not due to order effects. A similar analysis on those normal children who did complete all sessions produced the same results. Therefore, only the first and second sessions for both autistic and normal children have been compared. This represents a possible confounding variable which is addressed in our discussion section. For reinforcement in the social condition, the child was enthusiastically told "Good work, [child's name]," after every fifth correct target iden-

106

Garretson, Fein, and Waterhouse

tification (FR = 5) without establishing eye contact or touching the child. In the tangible condition, the child identified a preference for pretzels or pennies and could hear and see the reinforcer dropped into a bowl to the side of the examiner after every fifth correct response (FR = 5) but was not able to touch it until the end of the testing sessions.

Training Procedure Because it was important that the subjects reliably make the perceptual discrimination and accurately perform the task before the test of sustained attention began, the training sessions were as specific, structured, and lengthy as necessary to help each child meet a standard level of accuracy. Phase I o f the discrimination training consisted of an interactional training procedure with a booklet of 36 pictures (12 target stimuli, 24 nontarget stimuli). Mild punishment in the form of a firm "no, only press this when you see the chair" and removal of the child's hand from the button was used when the child responded to a nontarget or during an interval. When the child responded correctly, without prompting, to three targets and three nontarget stimuli, Phase 2 o f the training was started. This consisted o f a filmed standard practice presentation at the 3-sec slow speed constructed with an initially high rate of critical stimuli, gradually decreasing to the 30~ rate used in the experimental trials. Children were reinforced for every correct response with the appropriate reinforcement of that day until the child reached the criterion of 9/12 correct responses to target stimuli and not more than 3/27 errors of commission to nontarget stimuli. Phase 3 of the training procedure utilized the same training film but moved to the standard reinforcement schedule of FR = 5, according to the reinforcement condition o f the day until the child met the criterion o f 9/12 correct responses and not more than 3/27 errors o f commission. Testing on the first slow epoch began immediately after the child met the established criterion. Training to criterion was repeated before every new testing session. Four autistic children were unable to meet the training criterion after two training sessions and were not included in the sample population. Observations of each child during the full length o f both testing sessions were performed by an observer who noted on the event recorder any occurrences of looking away from the screen, self-stimulation, verbal selfmonitoring, other meaningful verbalizations and vocalizations. These data, which reflect an attempt to identify the ongoing relationship between behavior and accuracy on the CPT, will be reported separately.

Sustained Attention in Autistic Children

107

RESULTS

Dependent measures were evaluated with a repeated measure analysis of variance in which the m a j o r between subject factor was categorization as autistic versus normal. Scores were tested for the m a j o r within-subject factors of Reward (Social vs. Tangible), Rate of Presentation (Slow1, Fast, Slow2), and Time (minutes T1-T4) and any interactions of these variables. Any scores that were coded in the form of a proportion (e.g., accuracy rates) were subject to an arc sine t r a n s f o r m a t i o n before the analyses.

Hit Rate The basic measure o f accuracy, hit rate, is defined as the total correct responses out of the total number of target stimuli presented. An overall view of the 24 nested cell design and m a j o r accuracy results can be gained f r o m Figure 1, which shows (autistic vs. normal) mean hit rates, minute by minute.

Social

Reinforcement

Slow

Tangible

Fast

1"1 T2 T3 "r4 I

.92

|

.9C c cL o

.88 .86 .84 .82

.

.80 u

.78

"6

.76

~

.74

o

I

i

|

i

r

i

i

t

x

!

i

I

1

I

|

\/"

\

)

\

.68 .66

.60

I

T1 l"z "1"3 T4

\

.72

.62

.

Slow 2

TI T2 T3 T4

.70 o.

Reinforcement Fast

T1 T2 T3 T4

\

"/~'

Slow

Slow~

V

.58 .56 Autistic

......

DAD N o r m a l s

Fig. 1. Hit rate (group means for each minute): Group x Reward x Rate of Presentation

x Time.

108

Garrelson, Fein, and Wsterhouse

This illustrates how accuracy appears sensitive to the major factors of Time, Rate of Presentation, and Reward. Repeated measures ANOVA revealed that for both autistic and normal children, hit rate deteriorated over time within each epoch (vigilance decrement) and was lower on the fast task, the more difficult task. Although we expected globally poorer performance for the children with autism, group differences were evident only for changes in reinforcement during the third trial (Slow2). For autistic children, Slow~ performance was significantly lower than the Slow, performance but only under social reinforcement. The ANOVA data show the effect of group membership on two interactions: Group • Reward, F(1, 36) = 8.52, p = .006, and Group • Reward • Rate, F(2, 72) = 4.76, p = .011. Figure 2 illustrates the three-

.go ~ A

c

.85

"\

0

/

*\

*2

o

0

s

.8o

t.--

V

>,

o

.75

I. -i O O