Abstract
BACKGROUND:
Videomodeling is an effective technological tool for intervention and rehabilitation of children with autism spectrum disorders in different contexts.
OBJECTIVE:
The purpose of this paper is to verify whether a videomodeling intervention is effective for the acquisition of social and emotional behaviour and skills of children with ASD using tablet PC.
METHODS:
One 6-year-old child with ASD, who does not use verbal communication, was involved in a multiple baseline across behaviours design: the independent variable was the videomodeling intervention, instead the dependent variable was the participant behaviour. Intervention was preceded by a pre-session of Functional Communication training. An intersubject analysis was used to evaluate participant percentage of response through the sessions.
RESULTS:
The participant showed a marked improvement in his emotional skills. The intersubjective analysis carried out shows the sharp increase in response in many target behaviours from the baseline to the intervention phase. A limitation of the study was the lack of time: there were only two inversion sessions, no follow-up evaluation, no generalization and no double-blind study.
CONCLUSIONS:
It would be interesting for future research to extend it to everyday life scenarios, to enhance generalization in other contexts, involving a larger number of people with ASD.
Background
Children with Autism Spectrum Disorders (ASD) show difficulties in social interactions and mutual communication, as well as repetitive and restricted behavioral patterns [1]. ASD are heterogeneous disorders and, as the term “spectrum” suggests, have different levels of severity, which may be mild, medium or severe. The prevalence of the disorder is 1 out of 68 children [2].
The technological development of the last 30 years has allowed people with disabilities to reach a higher level of autonomy: a lot of interventions based on technological tools have enabled or rehabilitated abilities, competences, behaviors of people with a condition of ASD [3, 4, 5]. A particularly effective methodology, integrated with technology, for ASD is videomodeling, an educational methodology whose functioning is based on learning by observation and imitation. According to the social learning theory [6, 7, 8], individuals learn, not only through practical experience and their own contact with the object of learning, but also through the observation of people and their behavior.
Videomodeling, therefore, consists in the observation of a video, in which a model shows a target behavior, and in the subsequently imitation of that model. A lot of individuals with autism benefit from visually cued instruction [9] so videomodeling is widely implemented in interventions for ADHD people, especially with children and adolescents, not only in contexts of education and rehabilitation, but also at home, at school, and in the community. This methodology provides meaningful and functional learning concerning areas such as social skills, interaction and communication skills, adaptive behavior and functional autonomy skills, academic skills, motor skills [10, 11].
Moreover, studies carried out that videomodeling is effective for emotional skills learning: this methodology can support children with autism to recognize emotions and to percept and respond appropriately to facial expressions [16, 17]. Researchers implemented videomodeling to teach children with autism to respond to two similar perspective-taking tasks [18] and to respond appropriately to the context [19]. Thanks to the use of videomodeling children with autism, therefore, can learn complex empathic responses that include multiple components and can generalize those responses to different stimuli [20].
People with ASD tend to imitate without difficulty what they see in a video, rather than a person doing an action directly in front of them: many children with ASD show poor orientation to social stimuli already in the first year of life, they do not always turn when they hear their name, they have reduced eye contact and social detachment. For example, children look more at the background than the characters, prefer movie clips focusing less on people, faces and eyes than on other regions [12].
Videomodeling methodology, moreover, offers numerous advantages: the video can recreate a wide variety of settings; individuals with ASD have a strict selective attention and better process visual information rather than verbal ones, and videomodeling allows the attention to focus on a visual object. The vision can be repeated several times until the individual begins to reproduce the target behaviour. A skill that could be complex for children is broken down into several phases and analyzed step by step, to make the task more clearly defined and facilitate learning; scientific evidence [13] shows that watching videos leads to a maintenance of what is learned over time and can successfully promote a good generalization capacity in different contexts and environments. According to the current guidelines the video modeling technique is fully included in the strategies considered effective and valid, based on scientific evidence (or evidence based-practice).
The video can be played on devices such as a computer or tablet. The functionality of a technology such as tablets and portable devices is demonstrated by the flexibility and the ease with which they can be used: a portable device is easily transportable, not only in the field of learning and rehabilitation, but also in domestic and community contexts of people’s lives and is accessible to different age groups [14, 15]; it allows adaptability to situations and stimulates motivation.
Our study
Individuals with a condition of ASD present difficulties in the expressive behavior of their emotions, especially in modulating the expression of their face based on the affective state experienced. The expression of emotions is sometimes absent, sometimes ambiguous; in addition, individuals with ASD experience major difficulties in recognizing and interpreting emotional expressions in others [1]. Consequently, the difficulties of expression, but even before recognition, of one’s own emotional state can lead to frustration, maladaptive and dysfunctional behaviors. Despite previously cited studies, there is no enough research about the efficacy of videomodeling in teaching emotional skills in the field of ASD, especially through tablet PC.
The current paper presents a single case study in which skills of emotional communication were taught to a child with ASD using the video modeling methodology through a tablet PC. The study’s design was a multiple baseline across behaviors, A-B-A type. The research hypothesis is that the use of videomodeling is effective for the acquisition of social and emotional behaviour and skills of children with a condition of ASD.
Most researchers, who documented the effectiveness of videomodeling for the development of emotional skills, involved in their studies children or adolescents with good language skills. The present research work has had the purpose of helping a non-verbal child to learn and develop emotional skills: particularly, the communication of their own and others’ emotional states. The goal of this study is not only to evaluate whether videomodeling is a suitable technique or not to facilitate the acquisition of emotional skills, but also to display a conscious and profitable use of a technological tool like a tablet PC, within the intervention context.
Methods
Research design and participant
The experimental project involved a multiple baseline across behaviors design, A-B-A type [21]. The participant was a 6-year-old child with a mild-level condition of Autism Spectrum Disorder. The previous functional assessment through VB-MAPP, Verbal Behavior Milestones Assessment and Placement Program [22], highlighted the presence of significant problem behaviors (such as frequently saying “no”, escape and avoidance), the absence of spontaneous request (mand) and naming (tact), an intraverbal difficulties, the presence of eco-behavior or vocal imitation, a strong dependence on the reinforcer. The results, however, confirmed a good level of receptive and listening, play, social behavior and imitative behavior. It was also verified that the child knew how to use technologies that allow video playback, such as smartphones and tablets.
The child was following a cognitive-behavioral therapy inspired by the Applied Behavior Analysis – ABA principles. One intervention planned for him was particularly the Functional Communication Training [23], which stimulates the association of words with manual signs to facilitate communication, especially the spontaneous request.
Target task
To verify the research hypothesis, a task analysis was created and the following target behaviors were identified:
To show correctly an emotional state through the change of face expressions; for emotion facial expression, reference was made to the theorization of Ekman and Friesen [24]. To properly label an own emotional state, therefore recognizing and naming an own emotional state. To correctly label an emotional state experienced by others, therefore recognizing and naming the emotional state shown by others through facial expression and vocalizations.
The categories of emotions selected for the intervention were joy, anger and surprise: these emotions can vary in intensity but are universally recognized and expressed in the same way.
The setting in which the intervention was carried out was a large and luminous room of a rehabilitation centre; one wall consisted of large windows, the other walls were light colored. The environment was essential: there was a small table with two seats, a larger table with two larger chairs, a large shelf where many games were stored, a locker and a rug resting on the floor. In the room, during the sessions, there were always the child, the therapist and two students.
Before starting the experimentation, an assessment the activities and the objects preferred by the child was carried out: the preferences assessment was a necessary practice to understand which material to use in the different sessions in order to involve and motivate the participant as much as possible.
The assessment was performed in two sessions. The type of evaluation of the preferences used was the Single Stimulus Preference Assessment [25]. First, it was chosen to measure the percentage of interaction of the child with 11 stimuli in the therapy room, based on the suggestion of the therapist. The stimuli, presented randomly, were: the game of hide and seek, sweets such as donuts and biscuits, dolls, rocket, felt-tip pens, soap bubbles, musical instruments, train and elastic spring. If the child did not approach the stimulus after 5 seconds from the presentation the therapist reinforced and encouraged to try it for 5 seconds; if he did not approach the stimulus again, the therapist presented the next stimulus. The participant interacted with all the stimuli presented. All stimuli produced a high percentage of interaction; therefore, they were included within the experimentation (Table 1).
Example of data sheets. First session of preference assessment data recording
Example of data sheets. First session of preference assessment data recording
s
Videos were created based on the task-analysis. 3 videos for each emotion (total of 9) were made. In videos two people, a girl and a boy, interacted in a series of scenarios. One of the situations was the following: the boy expressed an emotion based on what was happening, for example playing a game, and the girl asked “How do you feel?”; then he named the correct emotional state (happy, angry, surprised). Other videos showed the girl expressing an emotional state through her face, then the boy looked at her, she asked “How do I feel?”; the boy then named the correct emotional state (happy, angry, surprised). Each video had a duration of about 30 seconds. The videos were recorded by a tablet, in particular a 10” Huawei Mediapad T3, and were edited through a program inside the tablet itself. The same portable device was used to show the videos to the participant.
Procedure
The entire intervention lasted for 3 months, two days a week of two hours each. According to the experimental research project the independent variable was the video’s vision, the dependent variable were the participant’s answers. Intervention was preceded by Functional Communication pre-session. Three images of emoticons (a happy, an angry and a surprised face) were shown to the child; furthermore, the therapist named the corresponding emotion with the appropriate intonation, in order to create an association between emotion, sign and vocalization. Once the child independently acquired and associated emotions through the signs, the intervention started.
The baseline session of the research consisted in 4 sessions with a 2-day weekly schedule. During the sessions, situations were created for each type of emotion. For example: the therapist started a game or activity that was pleasant and fun for the child (from the assessment of preferences), such as hide and seek game. Face’s expression of joy was observed during the game. After the end of the game/activity the therapist asked him “how do you feel?”: If the participant expressed the joy status (with the sign established during the pre-training) within 3 seconds of the question, the target behavior was reached, vice versa it was not. An inadequate response to the experienced emotional state was considered an error.
As for the other categories of emotions: for the category of anger the therapist and the child started an activity or a game of interest, and suddenly she took a game from his hand, or did not give him a game he wanted. For the surprise category, while the therapist and the child were playing in the room, a person knocked on the door and came in with a gift pack for the child, or the child accidentally found an object under the carpet (previously hidden from the therapist).
To evaluate the behavior of labeling others’ emotional state the therapist has shown, through the expression of the face and with vocalizations (for example “How beautiful!”, “How I enjoy myself!”, “Ugh, it doesn’t work!”, “Wow!”), an emotional state of joy, anger or surprise, then asked the participant “How do I feel?”: if the child expressed the correct emotional state within 3 seconds of the question the target behavior was reached, vice versa it was not. An inadequate response to the experienced emotional state was considered an error.
In each session the participant was given 3 response opportunities for each emotion category and for each target behavior to be evaluated. The scenarios were presented in random order during the sessions. Each scenario had a minimum duration of 30 seconds and a maximum of 1 minute.
Example of data sheets. Target behavior: to show the emotional state of joy
Example of data sheets. Target behavior: to show the emotional state of joy
During phase B of intervention, which took place in 6 sessions twice a week for the first two weeks and once a week for the last two weeks, the training consisting of videomodeling was introduced. At the beginning of each session, the participant was asked to sit on the seat and look at the tablet. The therapist was always sitting next to the child to make sure he paid attention to the video. After watching the videos and before the beginning of the real scenarios, the therapist always introduced a short interactive activity to avoid too strong conditioning of the intervention. At this point, the therapist started the interaction by presenting the same baseline scenarios. The participant was always given a time of 3 seconds to issue the appropriate answers. As in the baseline, if there was no response within 3 seconds, the scenario was concluded. Where the child has shown the right behavior and issued the right answer has been reinforced with an applause, with a vocalization (for example “Very good!”) or with a required game. The scenarios were presented intermittently between various activity for a total of 9 in each session.
A week after the end of the last session of the intervention, in order to avoid a possible dragging or carryover effect, phase A of inversion began, which took place in 2 sessions, one a week. At this stage the treatment with video modeling was withdrawn and there was a return to baseline conditions, in order to assess how the participant was able to reproduce the target behaviors without any type of training or input. In all these sessions the child’s behavior was recorded to facilitate data collection and subsequent evaluation.
The procedure of the Momentary Time Sample Recording (MTSR) [26] was followed for collecting data of the first target behavior, that is to correctly show the emotional state of joy through the change of the face. MTSR is a data-recording technique that is usually used when the observed behavior is not easily quantifiable.
Thus, the expression of the emotion of joy was analyzed by counting the number of time intervals in which the target behavior occurred. A total observation time of 30 seconds was divided into 10 intervals of 3 seconds. At the end of each time interval, therefore of the 3 seconds, and only at the end of that interval, the presence of the target behavior was observed, neither before nor after. If the target behavior was present at the end of that interval, an X was marked in the corresponding box. If it was not present, an O was marked. In the last column, the total number of times the target behavior occurred was entered. The total was then divided by 10 (number of intervals) and the result multiplied by 100, to obtain a percentage of expression of the behavior to be included in the data taking table. Subsequently the average percentage of expression of each single session was calculated by dividing the obtained percentages of each scenario of the session by the total number of occasions (therefore 3) (Table 2).
Regarding the expression of anger and surprise through the face, it was not possible to take advantage of the MTSR as the expression through the face of these two emotional states is unlikely to last more than a few seconds in a 30-second time interval. For this reason a table was created in which to insert the data relating to each emotion: where the expression of the emotional state was present within 3 seconds of the stimulus and for at least 3 seconds “Yes” was marked in the cell corresponding to the emotional state and at the session; in the event that the expression did not appear it was marked “No”; if instead the child expressed an emotion inappropriate to the context, it has been marked “Wrong”.
Data collection of the other two target behaviors, or to correctly label their emotional state and correctly label the emotional state of others, was obtained with the same data sheets described above for anger and surprise.
After data recording, these were inserted on scatter charts using the Microsoft Excel program: as regards the first target behavior, particularly the joy category, a graph was created in which have been inserted on the X axis all the sessions and on the Y axis the average percentage of expression calculated for each single session.
For the other target behaviors, scatter plots were created in which on the X axis the intervention sessions were always inserted and on the Y axis the response percentages for number of opportunities of each session: these percentages were obtained by dividing the number of responses with a positive outcome (Yes) by the number of occasions given in each session (3 for anger and 3 for surprise) and multiplying the result by 100.
Mean percentage (on the Y-axis) of manifestation of the target behavior between sessions (on the X-axis). Target behavior: showing the emotional state of joy.
Mean percentage (on the Y-axis) of manifestation of the target behavior between sessions (on the X-axis). Target behavior: to label the emotional states of anger experienced by others.
The research with single subjects traditionally involves an analysis of the data with repeated measurements, therefore an intersubjective comparison within and between the various sessions of the study. Therefore, to examine the effects of the intervention, a comparison is made between the performance of the subject before and after the introduction of the independent variable (video modeling). The intersubjective analysis carried out shows the sharp increase in response in many target behaviors from the baseline to the intervention phase. Thanks to the introduction of video modeling training, the participant quickly reached the results that were hypothesized at the beginning of the research.
During the baseline the child has already shown a good competence of expression and labeling of the emotional category of joy (Fig. 1). During the baseline sessions, the target behavior of labeling an own emotional state was inhomogeneous (between 30% and 100%); during the intervention instead, the percentage increased to about 80% of the correct answer for the expression through the face and to 100% for all the other target behaviors. During the inversion sessions these percentages were maintained.
Regarding anger (Fig. 2) the percentages of expression of the child emotional state during the baseline were medium-low (between 0% and 30%); the ability of label his anger was equally medium-low and inhomogeneous (between 0% and 30%) and those of label others emotional state was absent (0%). During the intervention sessions there was a clear and immediate increase in the response rates of all target behaviors, at 100% for face expression, 100% for labeling his emotional state and still 100% for labeling others’ emotional state. During the inversion phase, the previous percentages were maintained by the child.
Regarding the surprise, the first target behavior during the baseline was completely absent (0%), during the intervention there was a gradual improvement and increase in response rates up to the maximum (from 30% to 100%). The high percentage was maintained even during the inversion phase.
Mean percentage (on the Y-axis) of manifestation of the target behavior between sessions (on the X-axis). Target behavior: to label the emotional states of surprise.
Regarding the ability of label his surprise emotional state (Fig. 3), during the baseline was completely absent (0%), during the intervention there was a gradual increase in the response rate until the fifth session in which there was a decrease in the percentage of correct answer (from 100% to 60/70%) because the child often mistakenly labeled the emotion as “joy”; during the inversion phase the percentage was about 30/40%. Similarly, the ability of label the surprise emotional state of others was even more heterogeneous between the fourth and sixth session of intervention (100%/30%/100%) and during inversion sessions the percentage of correct answer fell to 30%. This decline was probably due to a weakness in the scenarios of the surprise that, after some time, could have created in the participant a certain expectation and therefore decreased the state of surprise.
The starting hypothesis of the research was that the videomodeling technique was effective for the acquisition of social and emotional behavior skills of children with ASD: from the general results, it can be concluded that the hypothesis was confirmed, as the child showed a marked improvement in his emotional skills, both in correctly showing his emotional state through the change of face expressions, in recognizing and giving a name to his emotional state, and in labeling an emotional state experimented and shown by others. The child successfully learned the target behaviors for each category of emotions, above all joy and anger. These results are in accord with literature on videomodeling that highlights the potential of this methodology in supporting children with autism to recognize emotions and to percept and respond appropriately to facial expressions.
In addition, the intervention was interesting thanks to the involvement of a child who does not use a verbal communication channel, if not a few words repeated in echoic. It was interesting, especially for Functional Communication intervention, because the child learned the ability to communicate an emotional state through the sign language: this means, first, that he expanded his repertoire of manual signs and, secondly, that he created a mode of expression of emotions that suited his needs, his characteristics and his abilities.
It is necessary to consider some limitations. A first limitation of the study is the fact that just 1 child was included; it would be better to include a larger sample of children to validate the intervention. Another limitation is certainly the length of the intervention, that is short compared to the most of multiple baseline across behavior designs. In fact, there were only two inversion sessions; no follow-up evaluation was possible after a certain period and generalization could not be assessed with new people and in new contexts. In studies and interventions with a single subject, as in this case, we must consider risk factors such as unpredictable variables that may compromise the progress of the experimentation, such as a disease of the participant, or any event that prevents continuous participation.
Because of the lack of time of the research, another limitation was the lack of a double-blind study to verify the actual results of the intervention, i.e. a data analysis by a second experimenter, in order to avoid a possible evaluative subjectivity of visual intersubjective analysis.
Another possible limit, as already underlined, is the difficulty in planning and constructing a scenario for the surprise emotion. Surprise is a difficult emotion to predict; It is possible that after a few sessions the child become accustomed to the scenario, to gifts presented and to object founded, so the need is to improve the construction of suitable scenarios that are more naturalistic.
Conclusions
It would be interesting for future research to extend the study to other types of emotions, such as sadness, fear and disgust. It would be also useful to extend it in everyday life scenarios, so that emotional skills can also be generalized in other contexts, involving people who take care of the child such as parents, teachers, but also peers and others. Moreover, one future direction could be to implement the video modeling methodology with the aim of supporting emotional self-regulation of children with ASD, to indicate strategies for managing their emotional state. Researchers in the future may also reflect on the possibility to extend this study on videomodeling, involving a larger number of people with autism spectrum disorders.
Footnotes
Conflict of interest
None to report.