Table Of ContentJOURNALOFAPPLIEDBEHAVIORANALYSIS 2005, 38, 549–553 NUMBER4 (WINTER2005)
NONCONTINGENT REINFORCEMENT AS TREATMENT FOR
FOOD REFUSAL AND ASSOCIATED SELF-INJURY
DAVID A. WILDER, MATTHEW NORMAND, AND JULIE ATWELL
FLORIDAINSTITUTEOFTECHNOLOGY
We examined the use of noncontingent reinforcement to decrease self-injury and increase bite
acceptanceinachildwhoexhibitedfoodrefusal.First,abrieffunctionalanalysissuggestedthat
self-injurywasmaintainedbyescapefromfoodpresentation.Next,weevaluatedanintervention
thatinvolvednoncontingentaccesstoavideoduringfeedingsessions.Resultsoftheintervention
showed adecrease in self-injuryandan increase inbiteacceptance.
DESCRIPTORS: brieffunctional analysis, foodrefusal, self-injury
_______________________________________________________________________________
Noncontingent reinforcement (NCR) is an METHOD
effective treatment for many forms of aberrant
Participant and Setting
behavior. However, the effects of NCR have
Raley, a 40-month-old girl, participated in
been inconclusive when implemented in the
the study. Raley had been diagnosed with
context of feeding sessions for children who
autism, gastroesophageal reflux, and food
exhibit food refusal. For example, Cooper et al.
allergies. She ate some select foods and did
(1995) found that noncontingent play facilitat-
notreceiveanycalories fromsupplemental(i.e.,
ed food acceptance in 2 children who exhibited
tube) feedings. Based on a physician’s recom-
food refusal, but they did not examine the role
mendation, her caregivers fed her soft textured
ofNCRduringinitialfoodconsumption.More
or pureed foods only. Her verbal repertoire
recently, Reed et al. (2004) found that NCR
consisted of a few unintelligible sounds. All
without escape extinction did not reduce
sessions were conducted in a therapy room
inappropriate behavior and did not increase
equippedwithaone-waymirroratapsychology
food consumption among participants. Only
clinic affiliated with a university. Sessions took
when combined with escape extinction did
placetwiceperweekforapproximately6 weeks.
NCR produce decreases in inappropriate be-
Due to the intensity of self-injury, session
havior, and even then only for some partici-
termination criteria were established. However,
pants. Escape extinction was necessary to
no sessions were terminated for this reason.
increase consumption among all participants
in the study, regardless of whether NCR was
Data Collection and Reliability
present or absent. The purpose of the present
During the brief functional analysis, the
study was to further examine the use of NCR
dependent variable was self-injurious behavior
(without escape extinction) for the treatment of
(SIB). During the treatment evaluation, de-
inappropriate behavior (self-injury in this case)
pendent variables included SIB and food
and food refusal.
acceptance. SIB was scored when Raley
pinched, scraped (against objects), or scratched
We thank the parents of the participant for their
cooperation. her own skin, or when she hit her head against
RequestsforreprintsshouldbesenttoDavidA.Wilder, the ground or an object. Food acceptance was
Florida Institute of Technology, School of Psychology,
definedasanyinstanceinwhichthefoodonthe
150W. University Blvd.,Melbourne, Florida32901.
doi:10.1901/jaba.2005.132-04 spoon went pasttheplane ofRaley’s lips within
549
550 DAVID A. WILDER et al.
5 s of presentation. Data on food expulsion room during sessions, but she did not interact
were also collected (Cooper et al., 1995), but with Raley during this time. During the ignore
this behavior never occurred throughout the condition, the therapist did not interact with
study.Datawerecollected on laptopcomputers Raley, and there were no programmed con-
using 10-s partial-interval recording (SIB) and sequences for SIB. During the play condition,
event recording (acceptance). Partial-interval Raley had access to preferred items and
recording was used instead of event recording activities identified via a paired-stimulus pref-
for SIB because some topographies of Raley’s erence assessment (Fisher et al., 1992), and no
SIB (e.g., scraping, scratching) did not occur as demands were presented. The therapist de-
a discrete event. Data on SIB were converted to livered attention to Raley on a fixed-time 30-s
apercentagemeasurebydividingthenumberof
schedule, and there were no programmed
intervals with SIB by the total number of
consequences for SIB. During the demand
intervals in the session and multiplying by
condition, Raley sat on the floor of the therapy
100%. Data on acceptance were converted to
room. The therapist presented a bite of food
apercentagemeasurebydividingthenumberof
(i.e., corn and sweet potatoes, Stage 2 baby
occurrencesofacceptancebythenumberofbite
food) on a spoon every 30 s. The therapist
presentations and multiplying by 100%.
delivered brief praise if Raley accepted the bite.
A second independent observer scored 29%
ContingentonSIB,thespoonwasremovedand
of sessions during the brief functional analysis
the therapist moved away from Raley for
and 33% of sessions during the treatment
approximately 15 s. If Raley did not accept
evaluation. Interobserver agreement for SIB
the bite but did not engage in SIB, the spoon
was obtained by dividing agreements by agree-
remained at her lips for 30 s, at which time
ments plus disagreements and multiplying by
a new bite was presented. The next bite of food
100%. Mean total, occurrence, and nonoccur-
was presented either after the escape interval or
rence agreements for SIB during the functional
at the next 30-s interval. During the attention
analysis were 95%(range, 78%to 100%), 81%
condition,thetherapistreadamagazineanddid
(range, 55% to 100%), and 92% (range, 78%
notinteractwithRaley.ContingentonSIB,the
to 100%), respectively. Mean total, occurrence,
therapist delivered a statement of concern and
and nonoccurrence agreements for SIB during
a brief physical touch. A multielement design
thetreatmentevaluationwere88%(range,61%
was used to evaluate the brief functional
to 100%), 81% (range, 50% to 100%), and
analysis. After the first four sessions, the play
86% (range, 64% to 100%), respectively.
condition and the demand condition were
Interobserver agreement for acceptance was
alternated in a brief reversal fashion. The order
obtained on a per-interval basis by dividing
in which the first four sessions were conducted
the smaller frequency by the larger frequency
and multiplying by 100%. Interobserver agree- was randomly determined. All sessions were
ment for acceptance was 90% (range, 80% to 10 min in duration.
100%). Treatment evaluation. Baseline sessions were
identical to the demand condition of the brief
Experimental Design and Procedure functional analysis. During NCR, sessions
Brief functional analysis. During the brief remainedidenticaltobaseline,exceptthatRaley
functional analysis, Raley was exposed to four had continuous access to a children’s video
conditions: ignore, play, demand, and atten- (identified as her most preferred item via
tion. Due to the severity of her SIB and her a paired-stimulus preference assessment). SIB
reported crying when separated from parents, continued to result in a 15-s break. In other
Raley’s mother preferred to stay in the therapy words, extinction was not in place. A reversal
NONCONTINGENT REINFORCEMENT AS TREATMENT 551
design was used to examine the effects of NCR. extinction may not always be necessary to
Because of the severity of SIB, all sessions were reduce inappropriate behavior and to increase
reduced to 5 min in duration. acceptance among children who exhibit food
refusal.
One reason for the discrepancy between the
RESULTS AND DISCUSSION
currentstudyandpreviousresearchcouldbethe
Figure 1 (top) depicts the percentage of
nature of the reinforcer. That is, the video was
intervals with SIB across the various conditions
not delivered in the traditional sense but was
of the brief functional analysis. Raley exhibited
present throughout the session. Although toys
relatively high levels of SIB during the demand
and interaction were continuously available
condition(M553%ofintervals)andlittleSIB
during sessions in previous studies (e.g., Reed
during the ignore (M 5 7%), play (M 5 5%),
et al., 2004), the extent to which there were
and attention (M 5 8%) conditions. The brief
breaks in interaction, toy manipulation, and so
reversal analysis further supports the results of
forth, is not clear. Also, Raley had a great deal
the four initial functional analysis sessions.
of prior exposure to the video and hence it may
Raley exhibited elevated levels of SIB during
have produced more predictable stimulation.
the demand condition (M 5 43%) relative to
For whatever reason, the video was preferred
the play condition (M 5 2%).
tothepointthatitcompetedwiththeeventthat
Figure 1 (middle) shows the percentage of
maintained SIB (i.e., escape from food pre-
intervals with SIB in the treatment evaluation.
sentation).Previousstudieshaveconfirmedthat
During the first baseline phase, SIB occurred
the noncontingent delivery of an arbitrary
during a mean of 44% of intervals. During the
stimulus can decrease aberrant behavior
first NCR phase, SIB decreased to a mean of
(Fischer, Iwata, & Mazaleski, 1997; Fisher,
7% of intervals. During the return to baseline,
O’Connor, Kurtz, DeLeon, & Gotjen, 2000)
SIB increased to a mean of 66% of intervals.
andthatNCRwitharbitraryreinforcersislikely
During the second NCR phase, SIB again
to be most effective when the arbitrary re-
decreased to a mean of 6% of intervals.
inforcers are of higher quality than the main-
Figure 1 (bottom) depicts the percentage of
taining reinforcer (Fischer et al., 1997). Lalli et
trials with acceptance in the treatment evalua-
al. (1999) reinforced compliance with a pre-
tion. During the first baseline phase, Raley
ferred edible item and did not place problem
accepted a mean of 20% of bites. During the
behavior on extinction. They speculated about
first NCR phase, her acceptance increased to
whether their treatment effects were due to the
a mean of 90% of bites. During the return to
quality of reinforcement available in competing
baseline phase, Raley accepted a mean of 23%
ofbites.Finally,duringthesecondNCRphase, schedules or to the altering of the establishing
Raley’s acceptance increased to a mean of 93% operation for escape-maintained behavior. In
of bites. the current study, because the video was
Results suggest that noncontingent reinforce- available on a noncontingent basis, the NCR
ment can be an effective treatment for escape- procedure may have been effective because it
maintained self-injury in some children who eliminated or altered an establishing operation.
exhibit food refusal. In addition, this study Access to the video may have reduced the
suggests that in some children, food acceptance aversiveness of the situation by altering the
can be enhanced with the use of NCR. The establishing operation for escape.
results of this study are in contrast to pre- One limitation of the current study involves
vious research (e.g., Reed et al., 2004), in delivery of attention during the brief functional
that the current study suggests that escape analysis. It is possible that the results of the
552 DAVID A. WILDER et al.
Figure 1. Percentage of intervals with SIB across the conditions of the brief functional analysis (Att 5 social
attention; Dem 5 demand) (top); percentage of intervals with SIB during the treatment evaluation (middle); and
percentageof trialswithbiteacceptance during the treatment evaluation(bottom).
NONCONTINGENT REINFORCEMENT AS TREATMENT 553
analysiswouldhavebeendifferenthadattention identified through such an analysis might be
been delivered by Raley’s mother instead of by rotated during mealtimes to prevent possible
a therapist. A second limitation is the brief habituation.
duration of sessions and the small number of
sessions per phase. It is possible that the results
REFERENCES
would have been different had sessions been
Cooper, L. J., Wacker, D. P., McComas, J., Brown, K.,
longer, phases been extended, or both. Howev-
Peck, S. M., Richman, D., et al. (1995). Use of
er, Raley’s mother reported that the interven-
component analysis to identify active variables in
tionwaseffectiveduring30-minmealsathome. treatment packages for children with feeding dis-
Thus, the brevity of the assessment and orders. Journal of Applied Behavior Analysis, 28,
139–153.
observation periods did not compromise the
Fischer, S. C., Iwata, B. A., & Mazaleski, J. L. (1997).
ultimate clinical outcome. A final potential Noncontingent delivery of arbitrary reinforcers as
limitation of the study is that it is possible that treatment for self-injurious behavior. Journal of
Applied BehaviorAnalysis,30, 239–249.
during the NCR phases of the treatment
Fisher,W.W.,O’Connor,J.T.,Kurtz,P.F.,DeLeon,I.
evaluation, nonremoval of the spoon (in the
G., & Gotjen, D. L. (2000). The effects of
absence of SIB) may have functioned as noncontingent delivery of high- and low-preference
extinction, thus limiting the extent to which stimulionattention-maintaineddestructivebehavior.
Journal ofAppliedBehavior Analysis, 33,79–83.
the results are attributable to noncontingent
Fisher, W., Piazza, C. C., Bowman, L. G., Hagopian, L.
access to the video. However, if this were the P., Owens, J. C., & Slevin, I. (1992). A comparison
case, a steady increase in food acceptance across of two approaches for identifying reinforcers for
personswithsevereandprofounddisabilities.Journal
all phases of the treatment evaluation would be
of AppliedBehaviorAnalysis,25,491–498.
expected, but such a data pattern was not
Lalli, J. S., Vollmer, T. R., Progar, P. R., Wright, C.,
obtained. Borrero, J., Daniel, D., et al. (1999). Competition
Given the mixed results of this and prior between positive and negative reinforcement in the
treatment of escape behavior. Journal of Applied
studies, future research should examine the
Behavior Analysis,32,285–296.
conditions under which NCR is and is not Reed, G. K., Piazza, C. C., Patel, M. R., Layer, S. A.,
likely to be effective for food refusal. Also, Bachmeyer, M. H.,Bethke, S. D., et al. (2004). On
therelativecontributionsofnoncontingentreinforce-
variations in the current treatment approach
ment and escape extinction in the treatment of food
could be explored. For example, a competing refusal. Journal of Applied Behavior Analysis, 37,
stimulus assessment might be conducted prior 27–42.
to treatment of food refusal to determine if
Received September 7,2004
stimuli can be identified that may success-
Final acceptanceJuly 14,2005
fully compete with aberrant behavior. Stimuli Action Editor,TimothyVollmer
