The Reinforcing Effect of Social Robots in Applied Behavior Analysis of Autistic Children
Institutional Affiliation
Author’s Name
Date
Table of Contents
Applied Behavior Analysis as a Science. 3
Proposed Research Methodology. 6
Discussion and Analysis of expected observations. 11
ABSTRACT
Comprehending the organic basis for autism spectrum disorder (ASD) is the foundation upon which experiments leading to possible treatment could be developed. Developing a robust treatment plan will, therefore, enable autistic individuals to have not only a chance to lead productive lives, but also independent lives. Of the various treatment options available, behavioural therapy is often commonly used due to the better results it provides compared to other alternatives. However, a challenge exists to determine the best therapy treatment out of the several types of behavioural therapies available. In this evidence-based research proposal, a quantitative study is proposed on the possible use of social robots in the application of the applied behaviour analysis (ABA). Social robots are intended to play a reinforcing effect to help children diagnosed with ASD acquire skills in comparison to the use of affection, toys, escapement, chocolates, and candies. Using seven social robots and twelve subjects, the proposed study hypothesizes that social robots could be used in real-word as viable reinforces in the implementation of ABA among children.
Applied Behavior Analysis as a Science
The concept of applied behaviour analysis (ABA) traces its origin from the findings of B.F. Skinner, the concept’s developer. According to Skinner, behaviours that lead to favourable or desirable outcomes tend to be reinforced. In contrast, those whose results or outcomes are not profitable are likely to decrease within a prolonged period (Roane & Carr, 2016). Skinner, therefore, argued that desirable outcomes could be developed or shaped by providing reinforcing effects to ensure the successful performance of the coveted results. This concept is similarly applied in the clinical treatment of the autism spectrum disorder among children, which are mostly affected by the psychological disorder. Ivar Lovaas first developed the first application of intensive applied behaviour analysis in the treatment of ASD in what was initially known as intensive behavioural intervention (EIBI). At its conception, the procedure provided 5-7 days of comprehensive treatment. It depended on discrete-trial teaching (DTT), which emphasized the repetitive teaching of skills in discrete but straightforward methods. Currently, advanced ABA studies exist that have further developed the concept into more sophisticated and productive forms.
Introduction
Advancements in technological innovations have led to the mass production of robots with enabled capacities of social interactions. Depending on the developers’ target, social robots’ scope of interaction range from toddlers to adults. Furthermore, more advanced robots are programmed to communicate in a variety of languages and also engage in a variety of social activities. Social robots that are designed to interact with children diagnosed with autism spectrum disorder are likely to be useful gadgets in the active employment of the applied behaviour analysis therapy among ASD individuals (Pennisi et al., 2016). The proposal follows previous successful, positive result recorded upon the use of social robots in care settings. According to Hung et al. (2019), social robots can be programmed to perform a range of clinical care activities which include cognitive training, companionship, affective and physiological therapy, and social facilitation.
In another study conducted to evaluate the interactions of autistic individuals with human beings, a touchscreen computer, and a robotic dinosaur, the ASD individuals tended to express themselves most verbally when the interaction partner was a social robot in comparison to human beings and other artificial tools (Hudson &Lewis, 2020). Other studies have also assessed the physical look of robots, among other parameters in their relation to autism spectrum disorder research (Palestra, & Esposito, 2017). In most of the ASD and social robots related studies, the use of robots as possible therapeutic tools have affirmed the possibility of robots producing more desirable results compared to physical human involvement as well as other artificial machines and tools.
Scientific literature reviews by scholars on the subject hold that the use of robots as therapy agents on ASD individuals could cause a range of both positive and negative behaviours. The systematic literature reviews argue that in as much as social robots may present promising therapeutic methods in the treatment of ASD patients, still more research work to ascertain the incremental benefits for the preference for social robots is necessary (Pennisi et al., 2016). On a critical perspective, other past studies have also been conducted that raise the concern for the need to bridge the gaps that exist between clinical researchers and social robots, also referred as Socially-Assistive Robots (SARS) (Hudson &Lewis, 2020).
In past clinical research, Perez and Rodriguez (2019) reported that the use of social robotics in real-life is not only a learning tool for children living with autism but also works to improve and develop their joint attention (JA). Zheng et al. (2017) also confirmed the possibility of designing, developing, and evaluating a noninvasive autonomous robot-mediated JA intervention system on fourteen (14) young children diagnosed with the autism spectrum disorder. According to Zheng’s results, the participants showed significant improvement in their joint attention skills as they maintained a constant interest in the social robot throughout the assessment periods.
Proposed Research Methodology
The proposed study is to be conducted at the Lieber Recovery Clinic in New Jersey. The targeted participants are the children in the facility’s autism unit. Being that New Jersey state leads among states with high ASD cases, selecting a care centre in this state would be most appropriate. The objective of this study is, therefore, to ascertain the effect of social robots as reinforcing agents in administering the applied behaviour analysis among autistic children. The proposed research seeks to evaluate the variance in the impact of ABA therapy on children when social robots alongside other recognized reinforcement tools such as toys, escapement, chocolates, candies, and also human affection are applied. This proposed research aims at answering three research questions. The three hypotheses are:
- What robotic style is a child diagnosed with autism spectrum disorder attracted to?
- What autism spectrum disorder profile in children is enticed to a social robot?
- Does the alternative use of social robots in ASD therapy among children produce better results in their ability to acquire skills in comparison to other commonly known reinforcement tools and techniques?
Research protocol
To respond to the three hypotheses, a three-step protocol is to be conducted according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-V) standards and provisions. Additionally, the study is to be endorsed by Lieber Recovery Clinic Autism Spectrum Disorder ethics research committee. The health facility has to cross-check to ensure that all the research standards as well as necessary scientific code of ethics are met to the latter. Lieber Recovery Clinic’s research and ethics committee are to approve the seven selected social robots. The robots are to be named: Jumbo Robot (SR1), All-amazing Kitten Robot (SR2), Happy-tot Robot (SR3), Disney Robot (SR4), Joyous-for-kids Robot (SR5), Brighter-days Kitty Robot (SR6), and You-are-special Fancy Robot (SR7).
Robot selection criteria
Key characteristics define the selected robots approved for use in this proposed study. The seven robots chosen for the study are to be commercially acquired and tested beforehand to ensure that they do not impose any form of physical or psychological harm on the selected children. Besides the assessment for safety, the robots are also supposed to be reactive to sound as well as be independent. Out of the seven robots, Disney Robot, Jumbo Robot, and Happy-tot Robot, are neither responsive to touch nor animated. On the ability to converse, Disney Robot, Happy-tot Robot, and You-are-special Fancy Robot should be conversational. The estimated purchase prices of the robots on online robots shops such as overstock.com, target.com, store.irobot.com, amazon.com, and jameco.com among other online shops range from as low as $75 to $4,500. However, they can as well be acquired on physical walk-in robot stores located in New Jersey. The estimated prices for specific robots are Jumbo Robot ($75), All-amazing Kitten Robot ($850), Happy-tot Robot ($250), Disney Robot ($300), Joyous-for-kids Robot ($2,000), Brighter-days Kitty Robot ($4,500), and You-are-special Fancy Robot ($450).
Participants’ Characteristics and Inclusion Criteria
To attain the (fourteen)14 required participants (subjects) for the proposed study, the researcher(s) is (are) to liaise with Lieber Recovery Clinic ASD unit in sending invitations to parents interested in having their children involved in the study. The invitation requests are to be accompanied with links to an online questionnaire that parents are to answer and evaluated to obtain the sample size. Alternatively, the researcher(s) could liaise with the facility’s ASD unit in identifying the neediest autistic patients within the care centre in liaison with the approval of the selected children’s parents or guardians. The inclusion mechanism to be used to identify the sample participants is 6-8 years. Additionally, the participants should be children enrolled for 28 therapeutic hours per week ASD program at the Lieber Recovery Clinic autism unit. The participants ought to be clients at the facility for at least the past twelve (12) months. The 14 children/subjects are to be labelled from C1 – C14 for data collection and recording for step 1 and step 2 activities.
Research Procedure
The results of the observations from the three-steps are to be tabulated, analyzed, and the hypotheses confirmed or annulled. Of the three steps, only the first two steps are tabulated since they both perform the same though slightly varied assessment technique.
Stage 1
The fourteen participants are to be subjected to a free operant observation preference evaluation (Chazin, & Ledford, 2016). Operant observation assessment, in this case, is used to capture in writing the quantity of time taken by the 14 children in responding to stimuli they are introduced to or subjected to. The seven robots; Jumbo Robot, All-amazing Kitten Robot, Happy-tot Robot, Disney Robot, Joyous-for-kids Robot, Brighter-days Kitty Robot, and You-are-special Fancy Robot are to perform seven known and easily recognized stimuli of the researcher (s)’s choice. The frequency of the exposure to the randomly selected stimuli should be four times in the first week of the study and should last for exactly 10 minutes. Notably, the activity is to be undertaken in a room with a bright but not distractive colour. Therefore, a total of 40 minutes per child is to be spent during the first week of the study.
To record the 10 minutes session, each session is to be subdivided into 10 intervals of 1 minute each. During the 10 1-minute intervals, the researcher or observer is to take note or even record the items that the subjects touch or even speak to or with. Staring at an item without conversing with it or touching it should not be considered as engagement. At this stage, the autistic children are either expected to contact, speak, or just stare at items without response. It is required of the observer to remain silent during the entire session. To get the results of the engagement for Analysis, the observer or researcher is needed to divide the recorded number of engagement intervals by 10 and multiplied by 100 to express the rate to percentage. For a subject to proceed to step 2, they have to interact with at least two of the robots and also show interest in at least two of the items of engagement. The affirmation or qualification for proceeding to step 2 should be jointly agreed upon by 5 observers. The observers are to check the autistic child’s activities with the robots as well as the stimuli.
Stage 2
Step 2 of the research procedure shall require the researcher(s) to administer a paired stimulus preference evaluation with an assessment of each of the 14 participants (Chazin, & Ledford, 2016). Similar to step 1, this step likewise, will demand that participants engage in the therapeutic investigation four times during the second week with each session lasting 10 minutes. Like step 1, the second step too will also take a total of 40 minutes for each participant. The difference between this step and the initial step is that in step 2, the robots are paired, and the participants are expected to either choose a pair or none. Upon choice of a particular pair, the participants are accorded 1-minute playtime with the selected pair. This step aims to identify the subjects’ preferences. These preferences are likely to be formed by week one’s interaction with social robots.
Like step 1, this step as well requires the joint approval of five observes for participants to progress to the next step (step 3). However, participants are subjected to more restrictions for them to progress to step 3. Progressing to step 3 requires that for a participant to proceed to the next step, then they need to have interacted with any of the robots for a minimum of 80% of the allotted time for a minimum of one session during the first step in the first week. This requirement, therefore, implies that subjects that participate in step 2 are only those that met the 80% minimum engagement requirement in step 1. Notably, the choice of the 80% minimum required qualification is not research-based but rather a requirement of applied behaviour analysis (ABA) practice (Bishop et al., 2019).
Stage 3
This final stage is expected to be comprised of fewer participants due to the elimination that occurs in the previous two steps. Contrary to the other two phases, this step is not tabulated as is the case with the two preference assessments. This step occurs in the third week. The engagement period in stage 3, like the first two stages, is also 10 minutes. However, different from the other two phases, the third stage employs the use of a high preference social robot that brings in a reinforcing effect on the engagement. Like in the other two stages, stage 3 will also require the joint participation of five observers in the selection of the high-preference social robot. Finally, the observers are tasked with the role of selecting an appropriate expressive or receptive active acquisition program which is to be used in the comparison of the response rates between high-preference robots and non-robotic items.
Discussion and Analysis of expected observations
The engagement of 14 participants/subjects and 7 social robots in real-world interactions proves the argument that robot-human interactions can produce positive social impacts. It is expected from the study that ten (10) of the selected subjects will engage actively in the interaction, whereas (four) 4 are expected to exhibit minimum human-robot interaction. From the study, there is no expectation of the emergence of any general profile among the 14 participants which is likely to distinguish the participants that engage in social interaction from those that fail to engage in robot-human interaction actively.
The study is also expected to prove that none of the 7 social robots is likely to be preferred over the other. On the reverse, none of the 7 social robots is also expected to be least preferred in comparison to the other robots. Therefore, this neutrality in preference of robots means that any of the 7 randomly selected social robots are possible reinforcement agents in the development of cognitive and social skills. These two observations thus affirm the first two research questions. However, on the endorsement of the best autism spectrum disorder profile, the study results fail to provide any recommendation. Instead, the study only emphasizes cost, safety, and hygiene. Such observations and conclusions are due to the small sample size selected. To provide more comprehensive and detailed findings, there is a need for future research studies to expand the sample size and composition. Concentrating on New Jersey, a state with high cases of ASD individuals may provide a myopic view of the impact of social robots as reinforcement in employing the applied behaviour analysis in the development of cognitive and social skills among autistic children aged between 6-8 years.
Conclusion
Since the confirmation of the first autism spectrum disorder case in 1943, numerous research activities to develop the most effective ASD therapy have held that behavioural modification therapies present more positive results compared to other available and commonly used ASD treatment methods. A majority of sociology and psychological scholars hold that the use of applied behavioural Analysis alone as the most productive method in the treatment of ASD. On the other hand, other scholars argue that ABA alone is not sufficient but instead needs to used alongside other treatment methods. Therefore, the use of social robots to reinforce ABA in the treatment of autism spectrum disorder among children provides an opportunity for further research advancements since existing studies prove that their use over human beings offers more positive results.
References
Hung, L., Liu, C., Woldum, E., Au-Yeung, A., Berndt, A., Wallsworth, C., … & Chaudhury, H. (2019). The benefits of and barriers to using a social robot PARO in care settings: a scoping review. BMC geriatrics, 19(1), 232.
Hudson, C., & Lewis, L. (2020, March). Social Robots as Reinforcement in Applied Behavior Analysis. In Companion of the 2020 ACM/IEEE International Conference on Human-Robot Interaction (pp. 266-268).
Palestra, G., De Carolis, B., & Esposito, F. (2017). Artificial Intelligence for Robot-Assisted Treatment of Autism. In WAIAH@ AI* IA (pp. 17-24). http://ceur-ws.org/Vol-1982/paper3.pdf
Pennisi, P., Tonacci, A., Tartarisco, G., Billeci, L., Ruta, L., Gangemi, S., & Pioggia, G. (2016). Autism and social robotics: A systematic review. Autism Research, 9(2), 165-183.
Zheng, Z., Zhao, H., Swanson, A. R., Weitlauf, A. S., Warren, Z. E., & Sarkar, N. (2017). Design, development, and evaluation of a noninvasive autonomous robot-mediated joint attention intervention system for young children with ASD. IEEE transactions on human-machine systems, 48(2), 125-135.
Perez, I. A., & Rodriguez, I. R. (2019, March). Joint-attention development in children with ASD by using social robotics as an educational tool. In Conference Proceedings EDUNOVATIC 2018: 3rd Virtual International Conference on Education, Innovation and ICT (p. 21). Adaya Press.
Chazin, K. T., & Ledford, J. R. (2016). Evidence-based instructional practices for young children with autism and other disabilities. Retrieved November, 10, 2019.
Bishop, S. K., Moore, J. W., Dart, E. H., Radley, K., Brewer, R., Barker, L. K., … & Toche, C. (2019). Further investigation of increasing vocalizations of children with autism with a speech‐generating device. Journal of applied behavior analysis.
Roane, H. S., Fisher, W. W., & Carr, J. E. (2016). Applied behavior analysis as treatment for autism spectrum disorder. The Journal of pediatrics, 175, 27-32.