Received May 13, 2020, accepted June 24, 2020, date of publication June 29, 2020, date of current

version July 8, 2020.

Digital Object Identifier 10.1109/ACCESS.2020.3005296

Applied Behavior Analysis for the

Treatment of Autism: A Systematic
Review of Assistive Technologies
FÁBIO JUNIOR ALVES 1,2 , EMERSON ASSIS DE CARVALHO 1,2 , JULIANA AGUILAR 3 ,
LUCELMO LACERDA DE BRITO 4 , AND GUILHERME SOUSA BASTOS 2 , (Member, IEEE)
1 Department of Computing, Federal Institute of Education, Science, and Technology of South of Minas Gerais (IFSULDEMINAS), Machado 37750-000, Brazil
2 Institute
of Systems Engineering and Information Technology, Federal University of Itajubá, Itajubá 37500-903, Brazil
3 Department of Special Education and Rehabilitation, Utah State University, Logan, UT 84322, USA
4 Department of Psychology, Federal University of São Carlos, São Carlos 13565-905, Brazil

Corresponding author: Fábio Junior Alves ([email protected])

ABSTRACT Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that may lead to significant
impairment in social communication, repetitive patterns of behavior, and possible fixed and restricted
interests. Applied Behavior Analysis (ABA) is a well-supported and evidence-based treatment for the delays
attributed to ASD. Assistive technologies, such as gamification, software apps, computer-based training
(Web), and robotics; provide a standardized method of implementing ABA techniques. This review provides
a synthesis of the main characteristics of these technologies. The assessed proposals focused on technologies
such as Distributed Systems, Image Processing, Gamification, and Robotics. The primary objectives of
these tools sought to improve social behavior, attention, communication, and reading skills. Some common
limitations found in the literature was a failure to accurately define their target audience, and a failure to
comply with the dimensions of ABA as defined by Baer, Wolf, and Risley in 1968.

INDEX TERMS Autism spectrum disorder, applied behavior analysis, assistive technologies.

I. INTRODUCTION non-verbal languages, socioemotional reciprocity, etc.) and

The psychiatrist Eugene Bleuler first described the term repetitive and stereotyped behaviors with fixed and restricted
autism in 1911, to designate the loss of contact with reality, interests (such as simple motor stereotypes, echolalia, etc.)
which caused a great difficulty or impossibility of communi- [4]. In addition, the DSM-V classifies these deficits according
cation [1]. In 1943, psychiatrist Leo Kanner again used the to different levels of intensity: mild, moderate or severe [4].
term in his work presenting eleven cases of children who had Studies show that in the last few years, the number
an innate inability to establish affective and interpersonal con- of children diagnosed with ASD has been causing global
tact [2]. In addition to the inability to establish socio-affective concern. According to the Centers for Disease Control and
connections and excessive resistance to changes in the envi- Prevention - CDC,1 a body linked to the United States govern-
ronment, the individuals with ASD observed by Kanner also ment, in 2004 it was estimated 1 case for every 125 children
had severe difficulties in using language to communicate [2]. in the United States (USA), in 2020 was estimated 1 case
A year later, Hans Asperger described cases of children with for every 54 American children, this represents an increase
some characteristics similar to autism who also manifested of 131%. A worldwide statistical data estimate that 1 out
difficulties in social communication, however, they presented of 64 children in the United Kingdom, 1 out of 38 children
the differential of having normal intelligence [3]. in South Korea, and over 10 million of the general population
Currently, according to the Diagnostic and Statistical in India have been diagnosed with ASD [5]–[7].
Manual of Mental Disorders DSM-V, autism spectrum dis- In recent decades, Applied Behavior Analysis (ABA) has
orders (ASD) belongs to the category of neurodevelopmental become a well-supported and evidence-based treatment for
disorders. Individuals within the spectrum may have deficits the behaviors associated with ASD [8], [9]. Baer, Wolf,
in communication and social interaction (as in verbal or and Risley (1968) [10] published a paper suggesting seven
The associate editor coordinating the review of this manuscript and
approving it for publication was Francesco Piccialli. 1 https://www.cdc.gov/ncbddd/autism/data.html

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
118664 VOLUME 8, 2020
F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

dimensions for studies conducted in the scope of ABA:

applied, conceptually systematic, analytical, behavioral,
effective, generalized, and technological. The applied dimen-
sion involves a significant behavior change for the subject,
directed towards socially relevant behaviors. Regarding the
behavioral dimension, it describes that the focus of this
dimension is on the observable events of behavior that can be
directly or indirectly measured. In the analytical dimension,
the guarantee is obtained that the intervention will pro-
duce behavior change, as it proves the relationship between
behavior and the environment. The technological dimension
describes the procedure clearly and objectively, making it
descriptive for therapists, teachers, parents to ensure its repli-
cation. The conceptually systematic dimension refers to the
FIGURE 1. Phases of this SR.
use of well-defined concepts of applied behavior analysis,
relating procedures according to the principles of behavior.
The effective dimension, on the other hand, demonstrates
that the intervention guaranteed a change in behavior in a such as ABA, as well as assistive technologies. Assistive
socially appropriate way because if the change did not occur, technologies refer to the range of equipment, services, strate-
it means that the intervention was not adequate. Finally, gies, and practices designed to improve the functional skills
generalization is a dimension that requires the newly obtained faced by people with disabilities, promoting independence
behaviors to occur in different environments, have lasting and inclusion [27].
effects, and ensure that behaviors are adaptable for life. These Thus, taking into account the above information that:
seven dimensions are fundamental to define and qualify ABA 1) there is an increase in the ASD prevalence nowadays;
intervention processes [11]. 2) the effectiveness of ABA in improving autistic behaviors;
In addition to ABA techniques, [6] and [12] point out 3) the positive effect of using well-designed technology on
that another way to contribute to the intervention process of the development and learning process of ASD children; 4) the
individuals with ASD is to employ the use of technologies. well-known interest of ASD individuals in dealing with tech-
According to [13], since 1970, researchers have sought to nological devices; and 5) the need for studies focused on
insert computers in the treatment of children with ASD. specific intervention methods such the ABA; this systematic
Technology provides a positive effect on development and review aims summarize and analyze the existing literature to
learning compared to other instructional methods. Research provides a map for the development of assistive technologies
reports some potential benefits in using technology for the for the treatment of ASD based on the principles of ABA.
treatment of ASD to be: controllable, structured, adaptable,
and stimulating [14]–[16]. II. RESEARCH METHODOLOGY
According to [17], individuals with ASD are generally A systematic review (SR) is a scientific research method
interested in dealing with technology that includes images, that uses rigorous criteria and a well-defined methodolog-
audio, and videos. Recent research reports that a technology ical sequence to develop a central question [28]. Through
can contribute to the intervention process by improving learn- this approach, we developed a research protocol to assist in
ing, communication, social interaction, and other subgroups the search for papers in well-defined scientific databases,
of behaviors, through a positive and favorable environment enabling other researchers to follow the same methodology
[16]–[21]. However, if the technology is not well designed, or to evaluate the adequacy of the defined standards.
it can generate interaction barriers for the individual with According to [29], an SR is arranged in five consecutive
ASD to use it autonomously or, in extreme cases, it can cause stages: 1) question formulation; 2) search of articles;
discomfort and unnecessary stress to the individual [22], [23]. 3) evaluation and selection; 4) analysis and synthesis; and
To employ technologies that support learning in individuals 5) report of results. Reference [30] describe the main phases
with ASD, it is essential to understand how information of a SR as: 1) planning; 2) conducting the review; and 3) anal-
should be presented to them, consider their cognitive profiles, ysis of the results, which after being included in a more
behavioral developmental deficits, strengths and preferences, significant phase (packaging) generate the results expected
and to provide an environment suited to their style of learning by its executor. The conduction flow of our SR followed
[24], [25]. the steps proposed by [30], Figure 1(a), together with the
A systematic literature mapping on review studies about corresponding steps proposed by [29], Figure 1(b).
technologies to ASD people carried out by [26] has identified
some requirements that have not yet been satisfactorily met. A. RESEARCH QUESTIONS
They identified some areas that are possibly worth research- To cover every topic of interest in this SR, we formulated four
ing are studies focused on specific intervention methods, research questions. These questions considered relevant and

VOLUME 8, 2020 118665

 F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

TABLE 1. Research questions. TABLE 2. Inclusion criteria.

TABLE 3. Exclusion criteria.

general aspects important for this study and guide the devel-
opment of this SR to meet the proposed objective. Table 1
shows our research questions.
B. DATA SOURCES AND SEARCH STRATEGIES
We conducted this SR with scientific papers published in
the most comprehensive electronic databases of scientific
research in health and technology: ACM Digital Library,
ERIC Institute of Education Sciences, IEEE Xplore, Sage
Journals, PubMed, and Scopus. We selected articles written in
English because it is the internationally adopted language in
the scientific area. Additionally, we selected articles without TABLE 4. Quality criteria.
defining an initial year of publication, but 2019 was the end
date.
C. ARTICLE SELECTION
Once we chose the databases to search, we determined the
specific search strings to find articles to answer the research
questions and defined the inclusion, exclusion, and quality
criteria to refine and filter the articles found.
1) SEARCH STRINGS
presented in Table 3, and the Quality Criteria (QC) contained
A search string was created based on the keywords in Table 4. The QCs were based on standards in [31]. An asso-
(applied behavior analysis, autism, and technology and ciation of these criteria was used to define when a particular
its synonyms). The final search string was:("applied work would be included or excluded from the SR.
behavior analysis") AND
,→ ("autism") AND ("software" OR D. DOCUMENT SELECTION
,→ "technology" OR "digital
We found a total of 86 papers, as shown in Figure 2. The
,→ information and communication
first step was to export them to Start,2 a software that
,→ technologies" OR "computational
helped us to manage the classification process. Secondly,
,→ environment" OR "computational
we removed the duplicate papers, leaving 68 articles. After
,→ System" OR "computational Systems"
reading the abstract (and other sections, when necessary),
,→ OR "computer" OR "app" OR "game" OR
we applied the inclusion and exclusion criteria, resulting in
,→ "games" OR "gamification" OR "game
a total of 26 papers. Once a study adhered to an inclusion
,→ play" OR "artificial intelligence"
criterion with a consensus among the researchers, we clas-
,→ OR "machine learning" OR
sified it as acceptable for a full reading. Otherwise, when
,→ "deep learning" OR "robot")
a work adhered to an exclusion criterion with consensus
Due to the individual characteristics of each database,
among the researchers, we classified it as excluded. In the
we adopted different strategies to execute the search string.
absence of consensus, we placed such works on hold, and its
In IEEE, ACM, and PubMed databases, we performed
inclusion or exclusion was defined thought meetings between
searches without filtering, searching the entire text. In Sco-
the researchers. Finally, we performed a complete reading of
pus, Sage Journals, and ERIC databases, we conducted
all the papers.
searches by filtering titles and abstracts.
A snowballing process included two new articles.
Five works were excluded due to not meeting the quality crite-
2) STUDY SELECTION CRITERIA
ria. These proposals were excluded since they belonged to the
Based on the research questions, we defined the Inclusion
Criteria (IC) listed in Table 2, the Exclusion Criteria (EC) 2 http://lapes.dc.ufscar.br/tools/start_tool

118666 VOLUME 8, 2020

F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

TABLE 5. Demographic characteristics.

FIGURE 2. Steps of the selection process.

same authors and described the development stages of a same

tool. Thus, we considered only the full work of these authors,
which described the entire development and testing process
of the tool. This SR was conducted with 23 articles that were
successful in the selection and quality evaluation processes.
E. DATA SYNTHESIS
After reading all the selected papers, we tabulated them and
filled out a data extraction form for each one. This form
contains an overview of the work, some essential observa-
tions regarding its content and conclusions, as well as some
basic information such as bibliographic data, publication
date, summary, etc. Table 5 describes the main characteristics
of the selected papers.
III. RESULTS AND DISCUSSIONS
This section presents the results of the analyzed works
and answers the research questions. First, we performed an
analysis of titles and abstracts to find the most relevant
terms described in the papers. For this purpose, we used
VOSviewer3 software. The most relevant term was child;
FIGURE 3. Year of publication.
followed by autism and ABA.
Figure 3 shows the publication years’ frequency
distribution from 2008 to 2019. ABA-base technologies were 77% of the works designed to interact directly with children
proposed only after 38 years from the first proposal [13] to were based on Robots, Gamification, and Image Processing.
insert technology in the treatment of individuals with ASD. These systems were primarily intended to train children with
73% of the studies were published in the last six years. This ASD on skill acquisition.
percentage is explained by the growth of approximately 130% Figure 5 denotes what behavioral or developmental areas
in the average number of annual publications from 2013. the technological formats addressed. All of the web based
Answering the first research question (RQ1 - What programs (training programs and SSABAA) were intended
technologies (technological formats) are being employed to train therapists, psychologists, behavior analysts, and care-
to create ABA-based tools for ASD treatment?), most givers (parents and family members) on the implementation
proposals were based on the following technologies: Robots, of a behavior analytic techniques. 66% of the works related
Gamification, Image Processing (webcam), Storyboards, to the SSABAA (Support Systems for ABA Application)
Augmented Reality, and Web systems. Figure 4 presents the focused on providing tools to assist professionals in defin-
leading technologies used in the selected studies. Nearly 43% ing therapy activities, providing resources to monitor and
of the proposals were based on Web systems. These web collect session data, as well as providing follow-up. The
applications were focused on online training programs and remaining technologies were intended to be used by ASD
Support Systems for ABA Application (SSABAA). About children. These programs focused on techniques/programs
to improve specific skills such attention, social behavior,
3 http://www.vosviewer.com/ communication, and/or reading.

VOLUME 8, 2020 118667

 F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

FIGURE 4. Technological basis of the proposed tools.

FIGURE 6. Commonly addressed domains.

FIGURE 5. Addressed domains vs. technological basis.

FIGURE 7. Number of children included in the experiments.

Approximately 61% (see Table 5) of the tools were fication is related to how much support a child with ASD
developed for the direct use by a child with ASD dur- may need. Children diagnosed as mild need little support
ing the intervention process. The children in these studies to perform basic tasks; those diagnosed as moderate need
ranged from 2 to 15 years old. Most studies were con- some support; those diagnosed as severe need high levels of
ducted at schools, homes, or clinics (typically controlled support.
environments). In regards to the child’s current functioning (RQ3 - Did
As mentioned above, the technological tools developed for the studies seek to identify the Intelligence Quotient (IQ)
direct implementation with children with ASD focused on and ASD level of the investigated people?), studies were
skill acquisition (RQ2 - What behavior(s) did the technol- evaluated based on their inclusion of intelligence quo-
ogy seek to optimize for individuals with ASD?). Figure 6. tient information or validation of ASD level determined by
shows that approximately 87% of the proposed technologies DSM-V. Only 25% of studies described information about
sought improvements in communication, attention, or social the ASD level or current level of functioning of the children
behavior. The heightened focus on these behaviors may be involved. Not all studies followed the standards defined in
related to their significant impact on a child’s physical, social, the DSM-V. Reference [40] did not use a particular diag-
emotional, and intellectual functioning. nostic tool; however, they described with the children they
We identified some points in the proposed tools that may worked with as having moderate to severe levels of ASD.
compromise their quality, replicability, and use. Among them, Reference [46] used diagnostic data drawn from the Social
factors such as sample size and participant demographics Communication Questionnaire (SCQ) [53]. Their diagnos-
such as the participant’s skill deficits and IQ levels. Once tic scored children from 18 to 32 points, on a scale from
there were not single-subject design surveys, Figure 7 denotes 0 to 39. Reference [43] reported that their participants scored
that number of articles under each sample size. Approxi- greater or equal to the cutoff point in the Autism Diag-
mately 87.5% of works whose experiments were performed nostic Observation Schedule-Generic (ADOS-G) [54]. [49]
directly with children cited the sample size. A significant did not use any specific tool. The children involved in
portion of the articles (71.5%) cited for the review had a their study were previously evaluated and demonstrated con-
sample size under seven participants. Only 28.5% of the cerning levels of speech, extremely passive behavior, and
studies contained a sample size of ten to twelve participants. reduced involvement in spontaneous interaction and commu-
According to the APA (2013), the DSM-V classifies the nication. The lack of ASD-related information in most studies
ASD in three levels: mild, moderate, and severe. This classi- (75%), together with the lack of a standard to describe this

118668 VOLUME 8, 2020

F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

FIGURE 8. ABA professionals supported by the proposed tools. FIGURE 9. Attendance in training programs ( [5] proposed an online
training program but did not inform the attendance).

information, makes it difficult to properly analyze the target

audience for each tool, as well as a general analysis of them. TABLE 6. Analysis of works in relation to ABA dimensions.
The cognitive profile of ASD children is distinct from
those with typical development, as well as other develop-
mental anomalies with similar cognitive ability [55]. Thus,
we attempted to identify if the researchers were concerned
about determining the IQ level of the children selected to par-
ticipate in their researcher. Just 25% of the studies gave atten-
tion to IQ identification. References [46] and [47] described
that children had High IQ levels. References [43] and [39]
described that their participants had low IQ levels.
Adult participants were included in studies whose
programs were focused on training applicators of ABA
techniques. About 78% of the overall studies included in
our corpus were intended for use by adults to be trained in
ABA techniques. Figure 8 shows the distribution of intended
participants supported by the proposed tools. Approximately data is present to ascertain whether the behaviors targeted
50% of these tools were proposed for parents or caregivers, were deemed socially significant by the participants, care-
critical subjects in the ABA intervention process. The vast givers, or teachers. The behaviors selected for treatment are
majority of the tools, about 83%, were intended for behavior pivotal behaviors in the development of communication [33],
analysts, therapists, psychologists, or teachers. [40], [44], [49], [50], social participation [39], [46], [47], and
Considering works providing training programs, the academic skills [40], [42], [44], [49], [50]. The other five
attendance (sample sizes) was vastly different. Figure 9 dimensions were not followed entirely, with 91%, 45%, 18%
shows that almost 67% of the studies involved less than and 9% of the works complying with the conceptually sys-
30 participants, although approximately 71% of these works tematic, technological, analytical, and effective dimension,
used only distance learning resources. This small number respectively. All works performed their experimental proce-
of participants indicates a problem of scale, as one of the dures mostly in controlled environments, not being possible
purposes of online courses is to conduct training involving to analyze their adequacy to the generalized dimension.
a more significant number of participants. The primary focus of the studies evaluated were pilot
We attempted to assess the effectiveness of the technolog- studies or initial evaluations of technological behavioral inter-
ical tools through the seven dimensions of ABA (RQ4 - Did ventions. Often the measures used to evaluate the technolog-
the intervention process implemented by the technology ical interventions were qualitative measures on the ease of
follow the principles of ABA?). The seven dimensions were implementation or acceptability of the technology by chil-
used to evaluate ten studies that involved direct behavioral dren or adults. A single study [46] measured the behavior
intervention with children with ASD. We evaluated all works targeted by the intervention. This study reported an increase
to determine if their intervention processes were in accor- in self-initiated questions for participants who interacted with
dance with the primary objective proposed by each dimension their robots. Reference [34] also recorded data on participant
of ABA. Table 6 shows our analysis of the adequacy of such target behavior; however, their intervention was not effective
works concerning each dimension. in increasing emotional discrimination. Due to the limited
All of the studies met the dimensions of applied and data on the targeted behavior, combined with the already
behavioral. The eleven studies targeted observable and mentioned sample sizes, it is difficult to confirm the authors’
measurable behavior for children with autism. While no analysis of effectiveness for the remaining studies.

VOLUME 8, 2020 118669

 F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

Limited data on targeted behavior also limits the analysis ior selection, requirements, interface design, testing, and
of the analytical dimension. In order for an intervention to evaluation.
be considered analytical there must be an analysis between
the behavior and the environment. With the exception of [46] IV. CONCLUSION
and [34], it is impossible to assess whether these interven- In this work, we mapped the main proposals for ABA-based
tions meet the analytical domain without any data on the assistive technologies for the treatment of ASD. We aimed
targeted behavior. It was difficult to assess what measures to identify their technological basis, the behavior domains
of the behaviors were being collected by the technological addressed, and the target audience, as well as to describe their
system to assess if analysis beyond the pilot phase would be main characteristics. Several types of research in robotics
possible. are being performed. The use of robots allows the execution
There was no formal standard in the development of these of specific, repetitive, and motivating tasks. Such technolo-
technologies, which limits our ability to assess if the technol- gies can be tailored according to the particular needs of the
ogy is grounded in the principles of behavior analysis. If a individuals.
study claimed to implement an established behavior analytic Gamification is being employed to increase the learning
technique, such as discrete trial instruction [39], [40], [44], rate in computer-mediated environments, ensuring effective
Picture Communication Exchange System [42], or Pivotal monitoring and improvements in the pedagogical, social, and
Response Training [47], then the study met the dimension behavioral sense. Technologies employed in eye tracking
of conceptually systematic. A study could also be considered have been helpful to understand visual interests, supervise
conceptually systematic if the creators based their technology reading activities, and act as a communication device. Online
on the three-term contingency (antecedents, behaviors, and platforms are being used as a method to teach ABA to
consequences) [56]. The existence of a contingency for the caregivers and health professionals. Finally, computational
target behavior could be explicitly described by the authors technologies to assist the application and monitoring of ABA
[6], [39] or identified through the procedures described in the interventions aim to provide robust management of informa-
study [34], [46], [50]. tion, which can guide better decision making regarding the
As mentioned previously, analysis of the behavioral treatment.
interventions was very difficult due to the limited information It has been observed that research should investigate which
within the majority of the studies. The limited descriptions design resources are critical for the production of therapeutic
of the development and procedural aspects of the behavioral effects and how these resources create their impact (that is,
elements, excluded many studies from meeting the criteria an understanding of the mechanisms of change). Although
under the technological dimension. However, a few studies the existing literature offers some suggestions, additional
did provide significant detail of the behavioral aspects of the research is needed to establish guidelines for the development
intervention, as well as their method of implementation [6], and use of technology to provide behavioral interventions to
[34], [39], [46], [49]. Thus, these works may be considered children with ASD. The use of technology in interventions
technological and may be replicable by future scientists. generally requires technical or programming knowledge that
While all of the studies evaluated claimed to include ABA many clinicians lack, making it necessary to promote multi-
techniques and principles, it was often difficult to assess if disciplinary research and clinical teams. In the same manner,
the interventions developed in these studies abided by the development of technology-based behavioral interventions
dimensions of ABA delineated by [10], since none of them should adhere to the science of ABA. Board Certified Behav-
explicitly cited such dimensions. ior Analysts (BCBAs) should be consulted in the develop-
Regarding other factors, 95% of the papers presented user ment of the behavioral interventions to ensure adherence to
tests. The proposal validation is indispensable as it helps to the principles and dimensions of ABA. It is imperative that
verify user acceptance and whether the intervention had pos- other key stakeholders, such as caregivers, clinicians, and
itive results. The study that did not present test with the user educators, be consulted in determining what target behaviors
were in the prototype stage. User testing includes, in most should be the primary focus of an intervention and how tech-
cases, child observations, video recording, and interviewing. nology can be individualized for a child with ASD. We hope
When the test was applied to education professionals, health that this study can contribute to the state of the art research
professionals, and caregivers, they included observations, involving technology and ABA. By synthesizing the current
interviews, and questionnaires. All the technological tools literature, future researchers may be able to identify innova-
tested showed positive results, according to the evaluations tive research opportunities and provide ABA professionals
of their authors. Regarding the involvement of stakeholders with a summary of current and future behavioral technolo-
in the development stage, only 16% had direct participation. gies. As future work, we intend to develop a protocol to guide
The rest did not show any stakeholder involvement. The ther- and validate the adequacy of technological tools with regard
apists, doctors, and educators providing direct services to the to ABA principles. Consequently, tools developed following
target population have valuable knowledge and experience this protocol may ensure the correct execution of behavior
with the needs of this population. It is critical to involve analysis practices and would be appropriate to identify a
these stakeholders in all development stages, such as behav- profile of the target audience (age, ASD level, IQ, etc.).

118670 VOLUME 8, 2020

F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

REFERENCES [23] C. Putnam and L. Chong, ‘‘Software and technologies designed for people
with autism: What do users want?’’ in Proc. 10th Int. ACM SIGACCESS
[1] P. Szatmari, ‘‘The classification of autism, Asperger’s syndrome, and Conf. Comput. Accessibility, 2008, pp. 3–10.
pervasive developmental disorder,’’ Can. J. Psychiatry, vol. 45, no. 8, [24] S. Powell and R. Jordan, ‘‘Rationale for the approach,’’ in Autism and
pp. 731–738, Oct. 2000. Learning. Evanston, IL, USA: Routledge, 2011, pp. 11–22.
[2] L. Kanner, ‘‘Autistic disturbances of affective contact,’’ Nervous Child, [25] M. Tuedor, F. Franco, A. White, S. Smith, and R. Adams, ‘‘Testing
vol. 2, no. 3, pp. 217–250, 1943. literacy educational software to develop design guidelines for children
[3] M. Rutter and E. Schopler, ‘‘Classification of pervasive developmental with autism,’’ Int. J. Disability, Develop. Edu., vol. 66, no. 1, pp. 19–35,
disorders: Some concepts and practical considerations,’’ J. Autism Develop. Jan. 2019.
Disorders, vol. 22, no. 4, pp. 459–482, Dec. 1992. [26] J. C. Epifânio and L. F. Da Silva, ‘‘Scrutinizing reviews on computer
[4] Diagnostic and Statistical Manual of Mental Disorders (DSM-5), science technologies for autism: Issues and challenges,’’ IEEE Access,
American Psychiatric Association, Philadelphia, PA, USA, 2013. vol. 8, pp. 32802–32815, 2020.
[5] M. Kohli and S. Kohli, ‘‘Electronic assessment and training curriculum [27] A. M. Cook and J. M. Polgar, Assistive Technologies-E-Book: Principles
based on applied behavior analysis procedures to train family members of and Practice. Amsterdam, The Netherlands: Elsevier, 2014.
children diagnosed with autism,’’ in Proc. IEEE Region 10 Humanitarian
[28] J. Biolchini, P. G. Mian, A. C. C. Natali, and G. H. Travassos, ‘‘Sys-
Technol. Conf. (R10-HTC), Dec. 2016, pp. 1–6.
tematic review in software engineering,’’ Dept. Syst. Eng. Comput. Sci.,
[6] S. Shamsuddin, H. Yussof, F. A. Hanapiah, S. Mohamed, N. F. F. Jamil,
COPPE/UFRJ, Rio de Janeiro, Brazil, Tech. Rep. ES 679/05, 2005,
and F. W. Yunus, ‘‘Robot-assisted learning for communication-care in
vol. 679, no. 5, p. 45.
autism intervention,’’ in Proc. IEEE Int. Conf. Rehabil. Robot. (ICORR),
[29] D. Denyer and D. Tranfield, ‘‘Producing a systematic review,’’ in The
Aug. 2015, pp. 822–827.
[7] G. Xu, L. Strathearn, B. Liu, and W. Bao, ‘‘Prevalence of autism spectrum Sage Handbook of Organizational Research Methods, D. A. Buchanan and
disorder among us children and adolescents, 2014–2016,’’ J. Amer. Med. A. Bryman, Eds. Thousand Oaks, CA, USA: Sage, 2009, pp. 671–689.
Assoc., vol. 319, no. 1, pp. 81–82, 2018. [30] B. Kitchenham and S. Charters, ‘‘Guidelines for performing systematic lit-
[8] R. Landa, ‘‘Early communication development and intervention for chil- erature reviews in software engineering version 2.3,’’ Engineering, vol. 45,
dren with autism,’’ Mental Retardation Develop. Disabilities Res. Rev., no. 4, p. 1051, 2007.
vol. 13, no. 1, pp. 16–25, 2007. [31] B. Kitchenham, O. Pearl Brereton, D. Budgen, M. Turner, J. Bailey, and
[9] J. Virués-Ortega, ‘‘Applied behavior analytic intervention for autism in S. Linkman, ‘‘Systematic literature reviews in software engineering—A
early childhood: Meta-analysis, meta-regression and dose–response meta- systematic literature review,’’ Inf. Softw. Technol., vol. 51, no. 1, pp. 7–15,
analysis of multiple outcomes,’’ Clin. Psychol. Rev., vol. 30, no. 4, Jan. 2009.
pp. 387–399, Jun. 2010. [32] M. Ogino, A. Watanabe, and M. Asada, ‘‘Detection and categorization of
[10] D. M. Baer, M. M. Wolf, and T. R. Risley, ‘‘Some current dimensions of facial image through the interaction with caregiver,’’ in Proc. 7th IEEE Int.
applied behavior analysis 1,’’ J. Appl. Behav. Anal., vol. 1, no. 1, pp. 91–97, Conf. Develop. Learn., Aug. 2008, pp. 244–249.
1968. [33] M. Pistoia, M. Pistoia, and P. Casacci, ‘‘ASTRO: Autism support therapy
[11] E. K. Morris, N. G. Smith, and D. E. Altus, ‘‘B.F. Skinner’s contributions by RObot interaction,’’ in Italian Forum of Ambient Assisted Living. Cham,
to applied behavior analysis,’’ Behav. Analyst, vol. 28, no. 2, pp. 99–131, Switzerland: Springer, 2016, pp. 303–309.
2005. [34] D. E. Hughes, E. Vasquez, and E. Nicsinger, ‘‘Improving perspective taking
[12] R. Urbano, S. da Rocha Rosso, C. Á. Salomão, M. de Fátima Bernhardt, and empathy in children with autism spectrum disorder,’’ in Proc. IEEE Int.
‘‘ABAcadabra: Protótipo de aplicativo para auxílio no desenvolvimento Conf. Serious Games Appl. Health (SeGAH), May 2016, pp. 1–5.
cognitivo de crianças autistas,’’ in Proc. 2nd Congr. Technol. Educ. Paraíba, [35] D. Granpeesheh, J. Tarbox, D. R. Dixon, C. A. Peters, K. Thompson,
Brazil: Federal Univ. of Paraíba, 2017, pp. 676–682. and A. Kenzer, ‘‘Evaluation of an eLearning tool for training behavioral
[13] M. V. Panyan, ‘‘Computer technology for autistic students,’’ J. Autism therapists in academic knowledge of applied behavior analysis,’’ Res.
Develop. Disorders, vol. 14, no. 4, pp. 375–382, Dec. 1984. Autism Spectr. Disorders, vol. 4, no. 1, pp. 11–17, Jan. 2010.
[14] K. Dautenhahn, ‘‘Design issues on interactive environments for children [36] J. Jang, D. R. Dixon, J. Tarbox, D. Granpeesheh, J. Kornack, and
with autism,’’ in Proc. 3rd Int. Conf. Disability, Virtual Reality Associated Y. de Nocker, ‘‘Randomized trial of an eLearning program for training
Technol. (ICDVRAT), 2000, pp. 1–8. family members of children with autism in the principles and procedures
[15] K. Higgins and R. Boone, ‘‘Creating individualized computer-assisted of applied behavior analysis,’’ Res. Autism Spectr. Disorders, vol. 6, no. 2,
instruction for students with autism using multimedia authoring software,’’ pp. 852–856, Apr. 2012.
Focus Autism Other Develop. Disabilities, vol. 11, no. 2, pp. 69–78, [37] J. R. Khonglah and A. Khosla, ‘‘A low cost webcam based eye tracker
May 1996. for communicating through the eyes of young children with ASD,’’ in
[16] D. K. Murray, ‘‘Autism and information technology: Therapy with com-
Proc. 1st Int. Conf. Next Gener. Comput. Technol. (NGCT), Sep. 2015,
puters,’’ in Autism and Learning. Evanston, IL, USA: Routledge, 2011,
pp. 925–928.
pp. 98–113.
[38] C. D. C. Heath, H. Venkateswara, T. McDaniel, and S. Panchanathan,
[17] K. Khowaja, S. S. Salim, A. Asemi, S. Ghulamani, and A. Shah, ‘‘A sys-
‘‘Detecting attention in pivotal response treatment video probes,’’ in
tematic review of modalities in computer-based interventions (CBIs) for
Proc. Int. Conf. Smart Multimedia. Cham, Switzerland: Springer, 2018,
language comprehension and decoding skills of children with autism spec-
pp. 248–259.
trum disorder (ASD),’’ Universal Access Inf. Soc., no. 19, pp. 213–243,
[39] M. C. Buzzi, M. Buzzi, B. Rapisarda, C. Senette, and M. Tesconi, ‘‘Teach-
2020, doi: 10.1007/s10209-019-00646-1.
[18] J. A. Kientz and G. D. Abowd, ‘‘When the designer becomes the user: ing low-functioning autistic children: ABCD SW,’’ in Proc. Eur. Conf.
Designing a system for therapists by becoming a therapist,’’ in Proc. Technol. Enhanced Learn. Berlin, Germany: Springer, 2013, pp. 43–56.
Extended Abstr. Hum. Factors Comput. Syst. (CHI), 2008, pp. 2071–2078. [40] L. A. Dickstein-Fischer, R. H. Pereira, K. Y. Gandomi, A. T. Fathima, and
[19] D. F. Trevisan, L. Becerra, P. Benitez, T. S. Higbee, and J. P. Gois, G. S. Fischer, ‘‘Interactive tracking for robot-assisted autism therapy,’’ in
‘‘A review of the use of computational technology in applied behavior Proc. Companion ACM/IEEE Int. Conf. Hum.-Robot Interact., Mar. 2017,
analysis,’’ Adapt. Behav., vol. 27, no. 3, pp. 183–196, Jun. 2019. pp. 107–108.
[20] R.-Y. Tseng and E. Yi-Luen Do, ‘‘The role of information and computer [41] C. D. Hamad, R. W. Serna, L. Morrison, and R. Fleming, ‘‘Extending
technology for children with autism spectrum disorder and the facial the reach of early intervention training for practitioners: A preliminary
expression wonderland (FEW),’’ Int. J. Comput. Models Algorithms Med., investigation of an online curriculum for teaching behavioral intervention
vol. 2, no. 2, pp. 23–41, Apr. 2011. knowledge in autism to families and service providers,’’ Infants Young
[21] C. Whalen, D. Massaro, and L. Franke, ‘‘Generalization in computer- Children, vol. 23, no. 3, p. 195, 2010.
assisted intervention for children with autism spectrum disorders,’’ in Real [42] K. Daouadji Amina and B. Fatima, ‘‘MEDIUS: A serious game for autis-
Life, Real Progress for Children With Autism Spectrum Disorders: Strate- tic children based on decision system,’’ Simul. Gaming, vol. 49, no. 4,
gies for Successful Generalization in Natural Environments. Baltimore, pp. 423–440, Aug. 2018.
MD, USA: Brookes, 2009, pp. 105–148. [43] J. R. Cardinal, T. P. Gabrielsen, E. L. Young, B. D. Hansen, R. Kellems,
[22] R. Muñoz, T. Barcelos, R. Noël, and S. Kreisel, ‘‘Development of software H. Hoch, T. Nicksic-Springer, and J. Knorr, ‘‘Discrete trial teaching inter-
that supports the improvement of the empathy in children with autism ventions for students with autism: Web-based video modeling for para-
spectrum disorder,’’ in Proc. 31st Int. Conf. Chilean Comput. Sci. Soc., professionals,’’ J. Special Edu. Technol., vol. 32, no. 3, pp. 138–148,
Nov. 2012, pp. 223–228. Sep. 2017.

VOLUME 8, 2020 118671

 F. J. Alves et al.: ABA for the Treatment of Autism: A SR of Assistive Technologies

[44] L. Walsh and M. Barry, ‘‘Demystifying the interface for young learn- EMERSON ASSIS DE CARVALHO was born
ers with autism,’’ in Proc. IADIS Int. Conf. (IHCI), Amsterdam, in Campos Gerais, Brazil, in June 11, 1982.
The Netherlands, Jul. 2008. He received the B.S. degree in computer science
[45] L. Roll-Pettersson and S. Ala’I-Rosales, ‘‘Using blended and guided tech- from José do Rosário Vellano University, Alfenas,
nologies in a university course for scientist-practitioners: Teaching applied Brazil, in 2005, and the M.S. degree in computer
behaviour analysis to autism professionals,’’ J. Intellectual Disabilities, science and technology from the Federal Univer-
vol. 13, no. 2, pp. 113–142, Jun. 2009. sity of Itajubá, Itajubá, Brazil, in 2013, where he
[46] B. Huskens, R. Verschuur, J. Gillesen, R. Didden, and E. Barakova, ‘‘Pro-
is currently pursuing the Ph.D. degree in electrical
moting question-asking in school-aged children with autism spectrum dis-
engineering. His researches focus on automation
orders: Effectiveness of a robot intervention compared to a human-trainer
intervention,’’ Develop. Neurorehabilitation, vol. 16, no. 5, pp. 345–356,
systems, distributed systems, and artificial intel-
Oct. 2013. ligence. He is an expert on software development and programming with
[47] E. Barakova and T. Lourens, ‘‘Interplay between natural and artificial intel- ten years of industrial experience, developing solutions in mining, Linux
ligence in training autistic children with robots,’’ in Proc. Int. Work-Conf. technologies, and mobile and web applications. Since 2014, he has been
Interplay Between Natural Artif. Comput. Berlin, Germany: Springer, an Adjunct Professor with the Department of Computing, Federal Insti-
2013, pp. 161–170. tute of Education, Science, and Technology of South of Minas Gerais
[48] V. Bartalesi, M. C. Buzzi, M. Buzzi, B. Leporini, and C. Senette, ‘‘An (IFSULDEMINAS). He has authored eight conference papers/journal arti-
analytic tool for assessing learning in children with autism,’’ in Proc. cles. His awards and honors include the Best Undergraduate Student from
Int. Conf. Universal Access Hum.-Comput. Interact. Springer, 2014, the Brazilian Computer Society (SBC), in 2005.
pp. 209–220.
[49] C. A. da Silva, A. R. Fernandes, and A. P. Grohmann, ‘‘STAR: Speech ther-
apy with augmented reality for children with autism spectrum disorders,’’ JULIANA AGUILAR was born in Cali, Colombia,
in Proc. Int. Conf. Enterprise Inf. Syst. Cham, Switzerland: Springer, 2014, in April 17, 1985. She received the B.A. degree in
pp. 379–396. psychology and sociology and the M.Sc. degree in
[50] J. Leaf, A. Preston, D. Richter, and R. Gerlick. (Jun. 2017). An applied behavior analysis from the University of
Undergraduate Service Learning Research Project Using a Humanoid Missouri, Columbia, MO, USA. She is currently
Robot to Enhance Treatment for Children With Autism Spectrum Disorder. pursuing the Ph.D. degree in disability disciplines
[Online]. Available: https://www.scopus.com/inward/record.uri?eid=2- with an emphasis in applied behavior analysis with
s2.0-85030555966&partner%ID=40&md5= Utah State University (USU), under the mentor-
a257e84ec6db64b12b908d01c213ca09 ship of Dr. T. Higbee, BCBA-D, LBA. She has
[51] G. Presti, M. Scagnelli, M. Lombardo, M. Pozzi, and P. Moderato, been providing behavior analytic services to young
‘‘SMART SPACES: A backbone to manage ABA intervention in autism children with ASD, since 2006. She is currently a board-certified behav-
across settings and digital learning platforms,’’ AIP Conf. Proc., vol. 2040,
ior analyst and a licensed behavior analyst in the state of Utah. She cur-
no. 1, 2018, Art. no. 140002.
rently works as the Case Manager of Autism Support Services, Education,
[52] L. S. Heitzman-Powell, J. Buzhardt, L. C. Rusinko, and T. M. Miller,
‘‘Formative evaluation of an ABA outreach training program for parents Research, and Training (ASSERT) Program with the USU. She is a Graduate
of children with autism in remote areas,’’ Focus Autism Other Develop. Assistant with the USU and focuses her research on skill acquisition in young
Disabilities, vol. 29, no. 1, pp. 23–38, Mar. 2014. children with ASD, caregiving, and staff training, as well as the effects of
[53] M. Rutter, A. Bailey, and C. Lord, The Social Communication Ques- culture on behavioral interventions.
tionnaire: Manual. Torrance, CA, USA: Western Psychological Services,
2003.
LUCELMO LACERDA DE BRITO was born in
[54] C. Lord, S. Risi, L. Lambrecht, E. H. Cook, B. L. Leventhal, P. C. DiLavore,
A. Pickles, and M. Rutter, ‘‘The autism diagnostic observation schedule—
Teófilo Otoni, Brazil, in 1982. He received the
Generic: A standard measure of social and communication deficits associ- degree in history from UNIVAP, São José dos
ated with the spectrum of autism,’’ J. Autism Develop. Disorders, vol. 30, Campos, Brazil, in 2006, and the M.S. degree in
no. 3, pp. 205–223, 2000. social history and the Ph.D. degree in education
[55] B. A. Corbett, L. J. Constantine, R. Hendren, D. Rocke, and S. Ozonoff, from Pontifícia Universidade Católica (PUC), São
‘‘Examining executive functioning in children with autism spectrum dis- Paulo, Brazil, in 2009. He is currently pursuing
order, attention deficit hyperactivity disorder and typical development,’’ a postdoctoral training in special education with
Psychiatry Res., vol. 166, nos. 2–3, pp. 210–222, Apr. 2009. the Federal University of São Carlos (UFSCar),
[56] B. F. Skinner, Science and Human Behavior, no. 92904. New York, NY, São Paulo. His research focuses on teacher training
USA: Simon and Schuster, 1965. processes for evidence-based apractices for special education. He acts as a
teacher of basic education and higher education. He coordinated and taught
the Applied Behavior Analysis Postgraduate Course with the CBI of Miami.
He has authored the book Transtorno do Espectro Autista: uma brevíssima
introdução and several articles in magazines and periodicals on education
and school inclusion of the person with autism.
FÁBIO JUNIOR ALVES was born in Campo Belo,
Brazil, in February 17, 1983. He received the
B.S. degree in information systems from Pontifical GUILHERME SOUSA BASTOS (Member, IEEE)
Catholic University, Arcos, Brazil, in 2006, and the was born in Volta Redonda, Brazil, in Decem-
M.S. degree in education from the University of ber 22, 1977. He received the M.Sc. degree in
Vale do Sapucaí of Pouso Alegre, Pouso Alegre, electrical engineering from the Federal University
Brazil, in 2016. He is currently pursuing the Ph.D. of Itajubá (UNIFEI), Itajubá, Brazil, in 2004, and
degree in electrical engineering with the Federal the D.Sc. degree in electronics and computer engi-
University of Itajubá, Itajubá, Brazil. Since 2007, neering from the Aeronautics Institute of Technol-
he has been working in the academic area, and he is ogy (ITA), São José dos Campos, Brazil. He is
currently an Adjunct Professor with the Department of Computing, Federal currently an Associate Professor with the UNIFEI.
Institute of Education, Science, and Technology of South of Minas Gerais He has authored over 70 journal articles, book
(IFSULDEMINAS). He has authored ten conference papers/journal articles. chapters, and conference papers. He is carrying out his research on machine
His research focuses on automation systems. His awards and honors include learning, autonomous robotics, power systems, and technology for autism
the Best Undergraduate Student from the Brazilian Computer Society (SBC), spectrum disorders.
in 2006.

118672 VOLUME 8, 2020