REVIEW

CURRENT
OPINION ICT and autism care: state of the art
Charline Grossard a,b, Giuseppe Palestra b, Jean Xavier a,b,
Mohamed Chetouani b, Ouriel Grynszpan b, and David Cohen a,b

Purpose of review
Over the past 10 years, the use of information and communication technologies (ICTs) has increased in
regard to the treatment of individuals with autism spectrum disorders (ASDs). ICT support mechanisms (e.g.
computers, laptops, robots) are particularly attractive and are adapted to children with ASD. In addition,
ICT algorithms can offer new perspectives for clinicians, outside direct apps or gaming proposals. Here,
we will focus on the use of serious games and robots because of their attractiveness and their value in
working on social skills.
Recent findings
The latest knowledge regarding the use of ICT in the forms of serious games and robotics applied to
individuals with ASD shows that the field of serious games has already achieved interesting and promising
Downloaded from http://journals.lww.com/co-psychiatry by BhDMf5ePHKbH4TTImqenVI60I5lI//uKGULtht4sA3JNqz7kZtNuRhsnV52uK0a7 on 10/12/2018

results, although the clinical validations are not always complete. In the field of robotics, there are still
many limitations on the use of ICT (e.g. most interaction are similar to the wizard of Oz), and questions
remain concerning their eventual effectiveness.
Summary
To describe the implications of the findings for clinical practice or research, we describe two large projects,
namely, JEMImE and Michelangelo, as examples of current studies that are aimed at enhancing social skills
in children with ASD by including novel algorithms with clinical insights in robots or serious games.
Keywords
autism spectrum disorder, information and communication technologies, serious games, social robots, social skills

INTRODUCTION categories [7]. First, iPod and iPad apps aim to facili-
Autism spectrum disorders (ASDs) are developmen- tate specific aspects of social life. Despite some
tal disorders characterized by deficits in social inter- promising apps (e.g. Rubycube or the Social Detec-
action, communication and repetitive behavior. tive app), most available apps have received limited
Among these deficits, all social skills, which include empirical clinical validation [7]. Second, serious
social initiation, social interaction rules, emotion games can be described as ‘digital games and equip-
production and recognition, can be affected. As a ment with an agenda of educational design and
consequence, integration into society is a major ASD beyond entertainment’ [8]. Finally, ICT interven-
burden [1]. ASD patients may also present with tions include the use of robots with children with
different difficulties that can affect attention and ASD [7,9–11]. In this subdomain, clinical validation
executive learning [2], general cognition or oral is also limited, as many studies have focused on the
language [3]. Academic skills are often impaired, development of novel social skills for robots, thus
thus impacting academic achievement [4].
Information and communication technologies a
Service de psychiatrie de l’enfant et de l’adolescent, APHP, GHU
(ICTs) have opened new ways to help people with Pitié-Salpêtrière Charles-Foix, 47, boulevard de l’Hôpital and bCNRS
ASD. These technologies allow creation of real-life UMR 7222, Institut des systèmes intelligents et robotiques (ISIR),
Sorbonne Université, 75005 Paris, France
situations in a controlled area and offer clinicians
different supports to work with [5]. Moreover, the Correspondence to David Cohen, Service de psychiatrie de l’enfant et de
l’adolescent, APHP, GHU Pitié-Salpêtrière Charles-Foix, 47, boulevard
presentation of information seems to be particularly de l’Hôpital, 75013 Paris, France; CNRS UMR 7222, Institut des
adapted to individuals with ASD, and some studies systèmes intelligents et robotiques (ISIR), Sorbonne Université,
have found that persons with ASD are particularly 75005 Paris, France. Tel: +33 142162351;
interested by ICT [6]. e-mail: david.cohen@aphp.fr
Information and communication technology- Curr Opin Psychiatry 2018, 31:474–483
based interventions can be classified into three main DOI:10.1097/YCO.0000000000000455

www.co-psychiatry.com Volume 31  Number 6  November 2018

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ICT and autism care Grossard et al.

Regarding the efficacy of serious games in both

KEY POINTS domains, Grynszpan et al. [15] recently reviewed the
 ICTs are particularly attractive to children with ASD and literature and performed a meta-analysis to measure
are adapted to their needs. clinical impact through a nonspecific variable (stan-
dardized effect size) in individuals with ASD. The
 Serious games and robots are two support mechanisms results showed that the use of ICT has a significant
that are particularly used for the treatment of social
impact on the targeted skills. However, no study has
skills in children with ASD.
evaluated the maintenance of treatment effects and
 Previous studies showed benefits in using ICT for the the generalization to real life. Equally, the research-
treatment of social skills. ers could not contend that the use of ICT is more
 The field needs more robust studies to conclude that ICT adapted to a subgroup of children with ASD. Finally,
is effective and that the skills learned during treatment the authors reported that the use of ICT seems more
can be generalized to real life. efficient when a professional is involved during the
session than when the child has to play alone
at home.
Since this review, an interesting study described
the use of the game SEMA-TIC [16], which aims to
narrowing down the expectation of social training teach literacy skills. SEMA-TIC was developed by the
[10]. Cobtek laboratory and is novel in the literature
The present study aims to briefly review the because it aims to teach literacy to children with
recent literature about the use of ICT in the form poor functional language skills and ASD. The design
of serious games and robotics applied to individuals of the game is thought to specifically increase the
with ASD. ICT can be used to not only train abilities motivation of these individuals. The game includes
but also to compensate difficulties. We do not feedback and reinforcements that align with the
describe assistive technologies (see [12] and [13] interests of the children, such as bright, moving
for a review), but rather focus on serious games objects and illustrations from mechanics. The player
and robotics studies with a treatment agenda. Then, can follow his/her progression with a colored gauge,
to illustrate how ICT research can integrate clinician which gives him/her visual information. In addi-
insights into multidisciplinary research, we describe tion, each game is thought to train specific literacy
two large projects, namely, JEMImE and Michelan- skills, such as word recognition, alphabet knowl-
gelo, as examples of current studies that seek to edge, syntactic rule, and so on. A 23-week study
enhance social skills of children with ASD by includ- enrolled 25 children and showed good game usabil-
ing novel machine learning algorithms based on ity and adaptability in the targeted population. The
clinical expectations. children exposed to SEMA-TIC (experimental
group) improved their literacy skills on the targeted
exercise significantly better than children in the
SERIOUS GAMES control group. Moreover, three of the trained chil-
dren were able to decode words after training.
Generality
Serious games can be described as a combination of
education and entertainment. At present, many Serious games to teach social skills
applications exist on portable devices that aim to A large part of the recent research on serious games
be playful and to teach a targeted skill. However, in ASD has dealt with games that try to teach social
most of these applications have not been the object skills to patients. This interest is supported by two
of specific research in a clinical population and facts: the development of social skills represents a
cannot be evaluated (see [14] for a list). Here, we challenge for the integration of individuals with
focus on serious games created and evaluated on ASD [1]; and ICTs provide very good support in
other platforms such as computers, multitouch tab- creating social situations in an immersive way [5],
letops, and screens. Generally, the object of their contrary to the classic therapy in an office.
training can be separated into two distinct areas. On &
In a recent review, Grossard et al. [17 ] listed 31
the contrary, we found serious games teaching a serious games that aim to teach social interaction.
specific cognitive domain (e.g. executive function) We separated those games into two groups: the first
and academic skills (e.g. reading). On the contrary, group was composed of 15 games that teach social
there are several serious games that teach the core skills, such as joint attention, collaboration or adap-
difficulties of ASD, such as social skills and emotion tation to social context; and the other group
recognition. included 16 games that teach emotion recognition

0951-7367 Copyright ß 2018 Wolters Kluwer Health, Inc. All rights reserved. www.co-psychiatry.com 475

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

Child and adolescent psychiatry

and production. Regarding the games to teach social (Table 1). It is designed to be not only fun and
skills, half of them try to teach collaboration due to engaging but also to provide the player with several
ICTs that force collaboration. The other half uses visual indications to understand and progress in the
virtual reality to work on how a player must adapt to game, such as in the training part, different gauges
a given context (e.g. how to sit in a restaurant) [18]. of emotion that fill with pieces and when they are
In the second group, a large part of the serious games full, the game stops; and a wallet in the game part;
focus only on emotion recognition. Although emo- when it’s full, the game stops. The specificity of the
tion recognition is multimodal in nature [19], most game in terms of ICT research is the computation
of these games only propose static visual stimuli. and implementation of an algorithm (Fig. 1a) that
Four games try to teach emotion production, but provides feedback in real time regarding the quality
only one of them aims to train this skill in a social of the child’s production. This feedback allows the
situation [20]. Of the 31 games reviewed, we found children to correct their productions instantly in
that only a minority included a study trying to assess response to the colored gauges that represent a
the clinical interest of the game with evidence-base specific emotion (Fig. 1b). The performance of the
&
methodology [17 ]. In Table 1, we summarize, for algorithm has been evaluated recently and reached
both domains (social skills training and emotion excellent performance, with an average accuracy of
recognition), the games that also reported a clinical 82% [34]. A study evaluating the usability of JEMImE
study. We also added Emotiplay, which is a serious by children with ASD is currently underway. Ini-
game developed within a European project that was tially, the results seem to indicate that children have
&
reported recently [21 ]. fun playing JEMImE and that they quickly learn how
Most of the games present limitations regarding to adapt their productions to meet the expectations
the evidence of their clinical benefits. Studies are of the algorithm. However, we must now develop
often conducted with individuals with ASD and more social scenarios to significantly increase train-
high IQs, omitting a large population of individuals ing duration and settle a clinical study.
with ASD. The studies do not include large samples
of patients nor do they have control groups. How-
ever, in addition, most of the games are not attrac- AUTISM SPECTRUM DISORDER AND
tive in terms of design and do not use the large SOCIAL ROBOTS
possibilities offered by ICT regarding playability.
Most of the games only use static images, do not Generality
include virtual environments and do not involve The use of robots for people with ASD began 20 years
the player as a regular game does. This is particularly ago, and has specifically increased over the past
the case for the games that target emotion recogni- 10 years [35]. Different robots exist, and their char-
tion and production, whereas games targeting social acteristics depend on the research interaction. The
skills are often more attractive. However, the games robots can take humanoid, nonhumanoid, animal,
targeting social skills were evaluated with weaker or biomechanical forms. The possibility of control-
methodologies, contrary to the games targeting ling them, their capacities to move, to offer feedback
emotion recognition and expression. Despite these and rewards, and to socially interact do not allow
limitations, we believe that the use of serious games the same type scenario when one wants to settle an
to train people with ASD is promising. Accessibility experiment or a therapy session with a child–robot
has not been a major focus but may also impact the interaction [10]. However, what is the purpose of
way ICT can be used. The GOLIAH team (Table 1) is researching the use of a robot with a child with ASD?
currently computing a novel version of the game e- This is not an easy problem to grasp. However, the
GOLIAH that will be able, via a web platform, to literature can be classified as follows: research aim-
provide easy use for families at home in combina- ing to induce a targeted behavior, such as joint
tion with automatic gaming information that will attention, imitation, emotional expression, or spon-
be sent to therapists to follow a child’s progress [32]. taneous interaction; research used to teach special
tasks to the children; and less research attempting to
The JEMImE project study the quality and quantity of interactions
As said previously regarding serious games that between a child and a robot [10], which could help
teach emotion, these games usually lack education clinicians in their diagnosis [11].
of emotion production and social context. In this
context, we are developing a novel serious game
called JEMImE [33], which aims to train facial Social robots and autism spectrum disorder
expression production in social contexts in children As for serious games, children with autism seem to
with ASD. The game is the next step of JeStimule have greater interest in robots than typical children

476 www.co-psychiatry.com Volume 31  Number 6  November 2018

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 Table 1. Description of the main serious games targeting social interaction
Games targeting emotion recognition and production
Name Population Targeted skills Design Playability Results

Junior detective program Children with AS Emotion recognition and social RCT (49 patients) Game in 3D allowing a good Improvement in social skills after
[22] skills in context immersive experience training and at 5 month follow-up
FaceSay2 [23] Children with ASD Facial expression recognition RCT (49 patients) Static photos not allowing a Significant improvement in emotion
and joint attention real game experience recognition for the individuals with
HF-ASD after training
JeStimule [24] Children with ASD Emotion recognition in context Open trial (33 patients) Game in 3D allowing a good Significant effect of
immersive experience session  task  emotion interaction
for avatars and near significance for
pictures of real-life characters
Let’s Face it [25] Children with ASD and AS Emotion recognition RCT (42 patients) Static photographs, not Significant improvement in attention to
allowing a real game eyes and mouth in the treatment
experience group
LifeIsGame [20] Individuals with ASD Emotion recognition and Qualitative study Use of an avatar in 3D Participants enjoyed playing the game.
production in context No assessment
MindReading [26] Individuals with HF ASD Emotion recognition in context 2 non-RCTs (experiment 1 : 19 children Films and written examples, Improvement in emotion recognition
with ASD and 2 control groups with 22 not allowing a real game only in close but not distant
children with ASD and 24 typical experience generalization tasks
children; experiment 2: 13 children in
each group – ASD expe, ASD control
and typical control
The transporters [27] Children with ASD Emotion comprehension in non-RCT (56 children) Series of videos with QCM at Improvement in emotion recognition
context the end
&
Emotiplay [21 ] Children with HF ASD Emotion recognition non-RCT (89 children) 3D environment and short Improvement in all emotion recognition
games with photos measures and parents reported
reductions in autistic symptoms

0951-7367 Copyright ß 2018 Wolters Kluwer Health, Inc. All rights reserved.
Games targeting social skills

Cooperative Puzzle Game Children with HF ASD Collaborative skills 2 open group studies (22 boys with ASD Puzzle game with enforced Higher number of ‘negotiation moves’
[28,29] in experiment 1 and 16 boys with ASD collaboration on a
in experiment 2) multitouch tabletop
ECHOES [30] Children with ASD Joint attention and symbol use Open group study (19 children) Interaction with an avatar who The children responded significantly
can adapt his responses to more often to the practitioner after
the child’s behavior the sessions
Teachtown [31] Children with ASD Social, adaptative and non-RCT study (40 children) Software educational program The treatment group performed better
emotional skills, language with photos and video. Not than the control group on the
an immersive game Brigance Inventory of Early
Development, but these differences
were not significant
Goliah [32] Children with ASD Imitation and joint attention Study 1: Non-RCT (24 patients) Short tasks with enforced Improvement in both groups with no
(IQ > 60) Study 2: Open 6-month study with EEG collaboration on 2 tabletops significant superiority of the training
correlates connected to each other group

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

AS, Asperger syndrome; ASD, autism spectrum disorders; HF, high functioning.

www.co-psychiatry.com
ICT and autism care Grossard et al.

477
Child and adolescent psychiatry

FIGURE 1. JEMImE game. Due to the creation of an algorithm of facial expression recognition (a), the player can receive real-
time feedback on his/her production (b).

[36]. Robots present some characteristics that make attention, imitation, and social behavior, and less
them less anxiety-inducing for people with ASD. often language. Very few studies have focused on
They interact in a simple manner and make social the reduction in stereotypes, but the results are
&
situations less complex [37]. This is why researchers actually promising [38 ]. Given the number of
focus their experiments on robots and children with experiments and robotics platforms, it is not within
ASD in order to work on social and communication the scope of this brief review to summarize all the
skills [35]. These researchers generally target joint literature. At a glance, Table 2 presents the main

478 www.co-psychiatry.com Volume 31  Number 6  November 2018

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ICT and autism care Grossard et al.

Table 2. Main robots used with individuals with ASD during the years of 2017 and 2018

Robots Description Targeted skills

Nao [39–53] 50 cm tall Joint attention

25 degrees of freedom Imitation
2 cameras Turn taking
Microphones Eye contact
Speakers Pointing
Touch sensors Basic academic skills
LEDS Facial expressions
Sonars Verbal communication
Wi-Fi and ethernet connection Improve spontaneous social interaction
Software allows personalization Improve robot responses to children’s
Can be used in classroom, hospital, etc. affective state and engagement
Kaspar [54 ,55,56,57] Semiautonomous Support interaction
&

Software allows creation of new Support simple facial and body

scenarios expressions, gestures and speech
Child sized Learn about socially acceptable tactile
Touch sensors interaction
Head, arm and hand movements Imitation
Customizable Turn taking
Speaker Collaboration skills
Hygiene or food learning

Charlie [58] Low cost (200 USD) Joint attention

2 degrees of freedom in arms Imitation
2 degrees of freedom in the head Turn taking
1 webcam
1 speaker
Can be connected to a PC via USB
Can detect hands and head

TEO4 [59] 80 cm tall Teo was used in the krog project [60] as
2.5 h autonomy a companion for the child during
Possibility to stick different faces game sessions on a large screen. It
magnetically reacts when it is touched.
Dedicated to children with ASD
Distance sensors
Touch sensors
1 camera
Autonomous reactions but can also be
driven by an operator
Can move and speak

Riby [59] The new version of the TEO robot Engaging interaction
Dedicated to adults with ASD
130 cm tall
Sonar sensors

0951-7367 Copyright ß 2018 Wolters Kluwer Health, Inc. All rights reserved. www.co-psychiatry.com 479

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

Child and adolescent psychiatry

Table 2 (Continued)

Robots Description Targeted skills

R50-Alice « Mina »
[41,61] 69 cm tall Facial expression production
32 degrees of freedom with 11 degrees Facial expression recognition
of freedom in the head of which 8 are Imitation
for facial expression and 3 are for
neck movements
Speaker

QT Robot [62] Screen as his face Training emotional abilities

14 degrees of freedom for upper-body Body language
gestures Increasing the efficiency of therapy by
3d camera encouraging an active and engaged
1 microphone interaction
Connection by Wi-Fi

IRobiQ [63] 45 cm tall IrobiQ showed the following four

Speakers emotions with the different shapes of
1 touch screen its mouth: happiness, sadness,
Touch sensors surprise, and shyness. It is a
Sonars sensors commercial robot that had access to
IR sensors many children’s songs that were
1 RGB camera in the head accompanied by adjustable arm
LEDS movements

CARO [63] 93 cm tall Caro is specially designed to engage in

1 touch screen emotional interplay that is focused on
Touch sensors emotions that are expressed through
Depth camera the eyes of children with ASD
RGB camera
LEDS

KiliRo [64,65] Semiautonomous Improves learning and social interaction

2 degrees of freedom in each leg abilities
Head can move down, up, right and left
Tail can move right and left
One speaker attached on the robot

480 www.co-psychiatry.com Volume 31  Number 6  November 2018

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ICT and autism care Grossard et al.

robots used in relation to autism in 2017 and 2018. and France. One task was dedicated to the design of
&
Previous studies were reviewed in [9–11,38 ]. GOLIAH – a serious game that aims to improve joint
Some studies reported that the maximum child attention and imitation that was based on the Early
engagement appears with nonhumanoid robots. Start Denver Model agenda [67]. Two experiments
However, humanoid robots elicit a better generali- were also performed to investigate scenarios for
zation of the skills learned by the child during the controlled social interactions between children with
child–robot interaction [66]. A key limitation in the ASD and the Nao robot. The first experiment was
use of robotics is the fact that most of the interac- designed to induce the behavior of joint attention
tions provided thus far are similar to the wizard of with the child. Nao proposed different behaviors to
Oz, meaning that an operator is teleoperating the solicit a response from the child: the robot could
robot during the child–robot interaction [10]. In look at a picture, look at it and point it, or point to it
comparison with serious games, the clinical appli- and ask the child to look at it. During the experi-
cation of robotics in ASD is still limited, although ment, automatic measures were collected with
progress has been great in recent years. An interest- three-dimensional (3D) cameras, and specific algo-
ing achievement is the formulation of a deep learn- rithms were used to assess the child’s movements.
ing algorithm for the automatic perception of the The results showed that, contrary to some expecta-
affective state and engagement of children during tions, children with ASD more easily produced joint
robot-assisted autism therapy that can offer a better attention behavior with a human partner than with
tailored robot response to tackle the heterogeneity Nao. However, the experiment showed us that chil-
among individuals with ASD [53]. dren with autism produce more micromovements
and are more unstable in their posture than typical
children [51]. The second experiment targeted
The Michelangelo project motor imitation and imitation learning. Children
The FP7 Michelangelo project is a European project had to imitate movements that the robot produced
involving partners in Malta, the UK, Ireland, Italy, randomly. The robot was able to recognize when

FIGURE 2. Nao imitating a child’s movement. Reproduce from [68].

0951-7367 Copyright ß 2018 Wolters Kluwer Health, Inc. All rights reserved. www.co-psychiatry.com 481

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

Child and adolescent psychiatry

he/she was imitating. After 2 min of interaction, the Financial support and sponsorship
robot learned to imitate, and the two roles could be The JEMImE project was supported by the Agence Natio-
exchanged: the child proposed the movement and nale pour la Recherche (ANR) in his program CONTINT
the robot imitated him (Fig. 2). When we compared (JEMImE, ANR-13-CORD-0004).
how the robot was able to learn with adults, with The Michelangelo project was supported by the Euro-
typically developing children, and with children pean Commission (FP7: MICHELANGELO under Grant
with ASD, we found that, despite adequate learning Agreement No. 288241) and the endowment fund
in the three groups of partners, the number of ‘Entreprendre pour aider’.
computational neurons recruited to allow learning
increased significantly during interaction involving
children with ASD [52]. In other words, it seems that Conflicts of interest
the robot was able to identify a form of social There are no conflicts of interest.
signature in ASD children motricity.
To conclude, regarding robotics, the use of
robots with ASD individuals seems to be relevant REFERENCES AND RECOMMENDED
to research on social and communication skills. A READING
Papers of particular interest, published within the annual period of review, have
recent review found that participants with ASD been highlighted as:
often show better performance when interacting & of special interest
&& of outstanding interest
with a robot than when interacting with a human
&
[38 ]. Individuals with ASD demonstrate some social 1. Sasson NJ, Faso DJ, Nugent J, et al. Neurotypical peers are less willing to
interact with those with autism based on thin slice judgments. Sci Rep 2017;
skills with robots that typical children only use with 7:740700.
humans. However, even if some researchers are 2. Craig F, Margari F, Legrottaglie AR. A review of executive function deficits in
autism spectrum disorder and attention-deficit/hyperactivity disorder. Neu-
optimistic, the studies published so far indicate that ropsychiatr Dis Treatment 2016; 12:1191–1202.
autonomous interactions between robots and chil- 3. Demouy J, Plaza M, Xavier J, et al. Differential language markers of pathology
in autism, pervasive developmental disorder not otherwise specified and
dren remain limited [10]. Moreover, most studies specific language impairment research. Autism Spectr Disord 2011;
included very few children and were not adequately 5:1402–1412.
4. Kim SH, Bal VH, Lord C. Longitudinal follow-up of academic achievement in
powered to show a possible generalization of the children with autism from age 2 to 18. J Child Psychol Psychiatry 2018;
targeted skills. 59:258–267.
5. Josman N, Ben-Chaim H, Friedrich H, Weiss P. Effectiveness of virtual reality
for teaching street-crossing skills to children and adolescents with autism. Int
J Disabil Hum Dev 2008; 7:49–56.
6. Bernard-Opitz V, Sriram N, Nakhoda-Sapuan S. Enhancing social problem
CONCLUSION solving in children with autism and normal children through computer assisted
instruction. J Autism Dev Disord 2001; 31:377–384.
The use of ICT for individuals with ASD has 7. Boucenna S, Narzisi A, Tilmont E, et al. Interactive technologies for autistic
increased over the past 10 years. ICTs present some children: a review. Cogn Comput 2014; 6:722–740.
8. Park JH, Abirached B, Zhang Y. A framework for designing assistive tech-
characteristics, such as predictability, visual sup- nologies for teaching children with ASDs emotions. In: CHI’12 Extended
port, and a sequential presentation of information, Abstracts on Human Factors in Computing Systems. ACM: New York; 2012.
pp. 2423–2428.
which particularly align with the interest and needs 9. Diehl JJ, Schmitt LM, Villano M, Crowell CR. The clinical use of robots for
of individuals with ASD. Moreover, children with individuals with autism spectrum disorders: a critical review. Res Autism
Spectr Disord 2012; 6:249–262.
ASD have a specific need for these supports. Equally, 10. Huijnen CA, Lexis MA, Jansens R, Witte LP. Mapping robots to therapy and
ICTs are particularly adapted to work on social and educational objectives for children with autism spectrum disorder. J Autism
Dev Disord 2016; 46:2100–2114.
communication skills, such as imitation, joint 11. Scassellati B. How social robots will help us to diagnose, treat, and under-
attention, emotion production, and recognition, stand autism. In: Robotics Research. Springer; 2007. pp. 552–563.
12. Stasolla F, Boccasini A, Perilli V. Assistive technology-based programs to
which are impaired in ASD. ICTs offer clinicians support adaptive behaviors by children with autism spectrum disorders: a
new ways to interact and work with people with literature overview. In: Supporting the Education of Children with Autism
Spectrum Disorders. IGI Global; 2017. pp. 140–159.
ASD. Serious games and robots are two different 13. Benssassi EM, Gomez JC, Boyd LE, et al. Wearable assistive technologies
support mechanisms that are particularly promising for autism: opportunities and challenges. IEEE Pervasive Comput 2018;
17:11–21.
in terms of research, although the former domain 14. Shoaib M, Hussain I, Mirza HT, Tayyab M. The role of information and
has reached better clinical achievements than the innovative technology for rehabilitation of children with autism: a systematic
literature review. In: Computational Science and its Applications (ICCSA),
latter one. However, the lack of robust studies with 17th International Conference on IEE; 2017. pp. 1–10.
strong methodologies and the lack of proof for 15. Grynszpan O, Weiss PLT, Perez-Diaz F, Gal E. Innovative technology-based
interventions for autism spectrum disorders: a meta-analysis. Autism 2014;
generalization do not allow evaluation of the bene- 18:346–361.
fits of ICT interventions in individuals with ASD. 16. Serret S, Hun S, Th€ ummler S, et al. Teaching literacy skills to French minimally
verbal school-aged children with autism spectrum disorders with the serious
game SEMA-TIC: an exploratory study. Front Psychol 2017; 8:1523.
17. Grossard C, Grynspan O, Serret S, et al. Serious games to teach social
Acknowledgements & interactions and emotions to individuals with autism spectrum disorders
(ASD). Comput Educ 2017; 113:195–211.
The authors thank the families and children that par- This review lists 31 serious games to teach social skills and proposes two different
ticipated in the studies. scales to evaluate serious game on methodology and on playability

482 www.co-psychiatry.com Volume 31  Number 6  November 2018

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

 ICT and autism care Grossard et al.

18. Mitchell P, Parsons S, Leonard A. Using virtual environments for teaching 44. Conti D, Di Nuovo A, Trubia G, et al. Adapting robot-assisted therapy of
social understanding to 6 adolescents with autistic spectrum disorders. children with autism and different levels of intellectual disability: a preliminary
J Autism Dev Disord 2007; 37:589–600. study. In: Companion of the 2018 ACM/IEEE International Conference on
19. Vannetzel L, Chaby L, Cautru F, et al. Neutral versus emotional human stimuli Human-Robot Interaction. ACM; 2018. pp. 91–92.
processing in children with pervasive developmental disorders not otherwise 45. So WC, Wong MKY, Lam WY, et al. Robot-based intervention may reduce
specified. Res Autism Spectr Disord 2011; 5:775–783. delay in the production of intransitive gestures in Chinese-speaking pre-
20. Fernandes T, Alves S, Miranda J, et al. LIFEisGAME: a facial character schoolers with autism spectrum disorder. Mol Autism 2018; 9:34.
animation system to help recognize facial expressions. In: International Con- 46. Desideri L, Negrini M, Malavasi M, et al. Using a humanoid robot as a
ference on ENTERprise Information Systems. Berlin: Springer, Heidelberg; complement to interventions for children with autism spectrum disorder: a
2011. pp. 423–432. pilot study. Adv Neurodevelop Disord 2018; 1–13.
21. Fridenson-Hayo S, Berggren S, Lassalle A, et al. ‘Emotiplay’: a serious game 47. Talaei-Khoei A, Lewis L, Kaul M, et al. Use of Lean Robotic Communication
& for learning about emotions in children with autism: results of a cross-cultural to Improve Social Response of Children with Autism; 2017. https://www.
evaluation. Eur Child Adolesc Psychiatry 2017; 26:979–992. researchgate.net/profile/Jay_Daniel3/publication/318129979_Use_of_Lean_
The only studies in serious games proposing a cross-cultural evaluation. Robotic_Communication_to_Improve_Social_Response_of_Children_with_
22. Beaumont R, Sofronoff K. A multicomponent social skills intervention for Autism/links/595b77010f7e9bf415b48660/Use-of-Lean-Robotic-Communi-
children with Asperger syndrome: the Junior Detective Training Program. cation-to-Improve-Social-Response-of-Children-with-Autism.pdf.
J Child Psychol Psychiatry 2008; 49:743–753. 48. González JC, Pulido JC, Fernández F. A three-layer planning architecture for
23. Hopkins IM, Gower MW, Perez TA, et al. Avatar assistant: improving social the autonomous control of rehabilitation therapies based on social robots.
skills in students with an ASD through a computer-based intervention. Cogn Syst Res 2017; 43:232–249.
J Autism Dev Disord 2011; 41:1543–1555. 49. Palestra G, Cazzato D, Adamo F, et al. Assistive robot, RGB-D sensor and
24. Serret S, Hun S, Iakimova G, et al. Facing the challenge of teaching emotions graphical user interface to encourage communication skills in ASD popula-
to individuals with low-and high-functioning autism using a new serious game: tion. J Med Robot Res 2017; 2:1740002.
a pilot study. Mol Autism 2014; 5:37. 50. Rudovic O, Lee J, Mascarell-Maricic L, et al. Measuring engagement in robot-
25. Tanaka JW, Wolf JM, Klaiman C, et al. Using computerized games to teach assisted autism therapy: a cross-cultural study. Front Robot AI 2017; 4:36.
face recognition skills to children with autism spectrum disorder: the Let’s 51. Anzalone SM, Tilmont E, Boucenna S, et al. How children with autism
Face It! program. J Child Psychol Psychiatry 2010; 51:944–952. spectrum disorder explore the 4-dimension (spatial 3D þ time) environment
26. Golan O, Baron-Cohen S. Systemizing empathy: teaching adults with during a joint attention induction task. Res Autism Spectr Disord 2014;
Asperger syndrome or high-functioning autism to recognize complex emo- 8:814–826.
tions using interactive multimedia. Dev Psychopathol 2006; 18:591. 52. Boucenna S, Anzalone S, Tilmont E, et al. Extraction of social signatures
27. Young RL, Posselt M. Using the transporters DVD as a learning tool for through imitation learning between a robot and a human partner. IEEE Trans
children with autism spectrum disorders (ASD). J Autism Dev Disord 2012; Autonom Mental Dev 2014; 6:213–225.
42:984–991. 53. Rudovic O, Lee J, Dai M, et al. Personalized machine learning for robot
28. Battocchi A, Gal E, Sasson AB, et al. Collaborative puzzle game-an interface perception of affect and engagement in autism therapy. Sci Robot 2018;
for studying collaboration and social interaction for children who are typically 3:eaao6760.
developed or who have Autism Spectrum Disorder. 7th ICDVRAT with 54. Robins B, Dautenhahn K, Nadel J. Kaspar, the social robot and ways it may
ArtAbilitation. ICDVRAT/University of Reading, UK; 2008. pp. 127–134. & help children with autism: an overview. Enfance 2018; 1:91–102.
29. Battocchi A, Pianesi F, Tomasini D, et al. Collaborative puzzle game: a A good recapitulation of the different uses of Kaspar.
tabletop interactive game for fostering collaboration in children with autism 55. Mengoni SE, Irvine K, Thakur D, et al. Feasibility study of a randomised
spectrum disorders (ASD). J Assist Technol 2010; 4:4–13. controlled trial to investigate the effectiveness of using a humanoid robot to
30. Bernardini S, Porayska-Pomsta K, Smith TJ. ECHOES: an intelligent serious improve the social skills of children with autism spectrum disorder (Kaspar
game for fostering social communication in children with autism. Inform Sci RCT): a study protocol. BMJ Open 2017; 7:e017376.
2014; 264:41–60. 56. Robins B, Dautenhahn K, Wood L, Zaraki A. Developing interaction scenarios
31. Whalen C, Moss D, Ilan AB, et al. Efficacy of TeachTown: basics computer- with a humanoid robot to encourage visual perspective taking skills in children
assisted intervention for the intensive comprehensive autism program in Los with autism: preliminary proof of concept tests. In: International Conference
Angeles unified school district. Autism 2010; 14:179–197. on Social Robotics. Springer: Cham; 2017. pp. 147–155.
32. Jouen AL, Narzisi A, Xavier J, et al. GOLIAH (Gaming Open Library for 57. Wood LJ, Zaraki A, Walters ML, et al. The Iterative Development of
Intervention in Autism at Home): a 6-month single blind matched controlled the Humanoid Robot Kaspar: An Assistive Robot for Children with Autism.
exploratory study. Child Adolesc Psychiatry Mental Health 2017; 11:17. In: International Conference on Social Robotics. Springer, Cham; 2017;
33. Grossard C, Hun S, Serret S, et al. The reeducation of emotional expressions pp. 53–63.
for children with autism spectrum disorders thanks to information commu- 58. Boccanfuso L, Scarborough S, Abramson RK, et al. A low-cost socially
nication technologies: JEMImE project. Neuropsychiatrie de l’Enfance et de assistive robot and robot-assisted intervention for children with autism spec-
l’Adolescence 2017; 65:21–32. trum disorder: field trials and lessons learned. Autonomous Robots 2017;
34. Dapogny A, Grossard C, Hun S. JEMImE: a serious game to teach children 41:637–655.
with ASD how to adequately produce facial expressions. In: Automatic Face 59. Romero M, Bonarini A, Brivio A, Rogacheva K. Incremental and radical
and Gesture Recognition (FG 2018). 2018 13th IEEE International Con- innovation: design in robotics for autism. Teo and Riby robots. Evolutionary
ference on IEE; 2018. pp. 723–730. development. Design J 2017; 20(Suppl 1):S2375–S2388.
35. Aresti-Bartolome N, Garcia-Zapirain B. Technologies as support tools for 60. Gelsomini M, Bonarini A, Clasadonte F. Blending robots and full-body
persons with autistic spectrum disorder: a systematic review. Int J Environ Res interaction with large screens for children with intellectual disability. IDC
Public Health 2014; 11:7767–7802. 2015; 315–354.
36. Boucenna S, Narzisi A, Tilmont E, et al. Information communication technol- 61. Pour AG, Taheri A, Alemi M, Meghdari A. Human–robot facial expression
ogy (ICT) and autism: overview and focus on early developmental issues and reciprocal interaction platform: case studies on children with autism. Int J Soc
social robotics. Cogn Comput 2014. Robot 2018; 10:179–198.
37. Goodrich MA, Colton M, Brinton B, et al. Incorporating a robot into an autism 62. Costa AP, Steffgen G, Lera FR, et al. Socially assistive robots for teaching
therapy team. IEEE Intell Syst 2012; 27:52–59. emotional abilities to children with autism spectrum disorder. In: 3rd Work-
38. Pennisi P, Tonacci A, Tartarisco G, et al. Autism and social robotics: a shop on Child-Robot Interaction at HRI; 2017.
& systematic review. Autism Res 2016; 9:165–183. 63. Yun SS, Choi J, Park SK, et al. Social skills training for children with autism
This review resumed the different social robots use in ASD, their application, and spectrum disorder using a robotic behavioral intervention system. Autism Res
the remaining questions. 2017; 10:1306–1323.
39. David DO, Costescu CA, Matu S, et al. Developing joint attention for children 64. Bharatharaj J, Huang L, Al-Jumaily A, et al. Investigating the effects of robot-
with autism in robot-enhanced therapy. Int J Soc Robot 2018; 6:1–11. assisted therapy among children with autism spectrum disorder using bio-
40. Anzalone SM, Xavier J, Boucenna S, et al. Quantifying patterns of joint markers. In: IOP Conference Series: Materials Science and Engineering, Vol.
attention during human-robot interactions: an application for autism spectrum 234. IOP Publishing; 2017. p. 012017.
disorder assessment. Pattern Recogn Lett 2018. 65. Bharatharaj J, Huang L, Mohan RE, et al. Robot-assisted therapy for learning
41. Taheri A, Meghdari A, Alemi M, Pouretemad H. Human–robot interaction in and social interaction of children with autism spectrum disorder. Robotics
autism treatment: a case study on three pairs of autistic children as twins, 2017; 6:4.
siblings, and classmates. Int J Soc Robot 2018; 10:93–113. 66. Ricks DJ, Colton MB. Trends and considerations in robot-assisted autism
42. Saadatzi MN, Pennington RC, Welch KC, Graham JH. Small-group technol- therapy. In: Robotics and Automation (ICRA), IEEE (Institute of Electrical and
ogy-assisted instruction: virtual teacher and robot peer for individuals with Electronics Engineers) International Conference; 2010. pp. 4354–4359.
autism spectrum disorder. J Autism Dev Disord 2018; doi: 10.1007/s10803- 67. Dawson G, Rogers S, Munson J, et al. Controlled trial of an intervention for
018-3654-2 [Epub ahead of print] toddlers with autism: the early start Denver model. Pediatrics 2009;
43. Zheng Z, Zhao H, Swanson AR, et al. Design, development, and evaluation of 125:17–23.
a noninvasive autonomous robot-mediated joint attention intervention system 68. Boucenna S, Cohen D, Meltzoff AN, et al. Robots learn to recognize
for young children with ASD. IEEE Trans Hum Machine Syst 2018; 48: individuals from imitative encounters with people and avatars. Sci Rep
125–135. 2016; 6:; 19908. doi:10.1038/srep19908.

0951-7367 Copyright ß 2018 Wolters Kluwer Health, Inc. All rights reserved. www.co-psychiatry.com 483

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.