Spark — A Game to foster Social Initiation in Children with Autism
2019 - 2020
type of work
Research, Game Design
Imagine you’re regularly commuting to your workplace using the local bus. One day—as you are waiting at the bus stop—you look around. You recognize another person who seems to be waiting for the same bus as you. You come to realize that the same person has already been there the day before. The next day—as you’re approaching the bus stop—the same person is standing there again. This time the person notices you, makes eye contact and greets you. How would you respond? Most people will start exchanging non-verbal or verbal greetings with someone they see repeatedly, as it might be considered rude to not at least acknowledge the other person. Individuals affected by an autism spectrum condition (ASC) however might find it more difficult to engage in such social interactions.
research context and objective
The difficulties with social interaction and social communication experienced by ASC individuals can already be observed at a young age. In the long run these difficulties may lead to social fragmentation, mitigating the individual’s overall quality of life. To support ASC children in learning about social initiation, researchers have been dedicating efforts to develop various intervention techniques. A promising line of research was found in the usage of game-based technology-enhanced interventions facilitating physical embodiment. Lands of Fogs is one exemplary intervention using mixed reality face-to-face full-body interaction to foster social initiation in ASC children. My thesis presents the initial exploration of a possible Lands of Fog successor, which I designed, developed and evaluated with the support of a PhD student from Italy.
Issues identified during Lands of Fog's research
The epic adventure format of the game is more attractive to male than to female users;
The limited amount of possible activities and outcomes restricts the collaboration possibilities between players and limits the overall playtime;
The feedback to users’ actions is non-contingent; and
The game’s reward structure is focused on the pursuit of a single large reward.
A preliminary analysis—contrasting preceding projects with relevant research literature—provided deeper knowledge on successful design features and possible design limitations. Appropriate design principles were derived from the so gained knowledge and guided the development of Spark. The scope of this thesis made it difficult to involve children, especially ASC children, as co-designers of the experience. Considering that this approach can result in the creation of experiences which do not satisfy users’ expectations and needs, the design process was informed by non-ASC children during two testing sessions of provisional prototypes. The design was iteratively improved and extended based on the feedback received.
What children want and expect is likely to be different from what adults think children want and expect.
— Good, J. and Robertson, J., 2006
In total 44 non-ASC children from local school classes provided insights for the evaluation of Spark’s game mechanics and usability during two testing sessions. After a short briefing about the general play rules, the children were asked to explore Spark during a ten-minute play session. Following, the children reported their experiences to a classmate (KidReporter) or through a digital questionnaire. Afterwards each child was asked to prepare three drawings displaying their experiences during the play session.
Appeal to all sexes equally
Build upon abstract game elements
Use tangibles to interact with the virtual environment
Consistency and predictability
All game mechanics have to follow a consistent and clear set of rules, the consequences of interactions need to be foreseeable
Avoid fast motion of virtual elements at all times
Keep primal interactions readily acceptable and gradually introduce novel, more complex interactions
Foster social initiation through the implementation of collaborative tasks
Intermediate goals and rewards
Provide multiple, stimulating rewards on an intermediate basis for the completion of sub-tasks
Provide sufficient, understandable feedback for all actions taken
Offer a potentially infinite number of situations and activities
Diversify players’ behaviors and avoid encouraging repetitive behaviors
React with intelligence
Automatically adjust game mechanics according to the current game state and shift the players’ foci to novel tasks
Hardware setup (environment)
The project was installed at the premises of the Full-Body Interaction Laboratory at the Universitat Pompeu Fabra, Barcelona. Two high-quality projectors were mounted onto scaffolds and positioned on opposite ends of the room. The projectors were oriented towards the floor, allowing the projection of a six by six meters play area. The floor was equipped with special, white tiles providing a high-quality projection surface. A speaker was positioned right next to each projector scaffold, allowing the reproduction of panned audio. Both speakers were set to a third of their maximum volume to not discourage users from speaking. Four cameras were mounted on scaffolds and positioned beneath the ceiling, centered above the play area. The four cameras together covered the whole play area and allowed the detection of RGB-light emissions within the play area. Asides the play area, chairs provided seating options for spectators. The rest of the play area remained empty without any further physical objects.
Hardware setup (controller)
Each child was equipped with a custom-made handheld device used to approximate the respective child’s position. The design of the device was inspired by tools used to manipulate objects in deep space, whilst maintaining a rather abstract appearance. To avoid a rejection caused by discomfort of wearables, the tool was designed to be used by holding on to a handle. An adjustable pole allowed accounting for the children’s different body heights. The foot of the device was meant to be hovered slightly above the ground to improve the accuracy of the position tracking. A LED strip, installed in a transparent tube at the foot of the device, was powered up by eight AAA batteries, stored in a box attached to the lower end of the pole. The two devices could be lit up in either a blue or a red color. A switch at the top of the box thereby allowed an easy turning-on and -off of the device.
The virtual environment (VE) used in Spark was created using Unity. The Unity environment was connected to a custom-built position tracking software using an open sound control protocol. The position tracking software was originally created for Lands of Fog. It used the input from four ceiling mounted cameras to track a set of four customizable RGB light emitters within the play area.
Game design (content)
Spark was developed to be used by ASC and non-ASC children aged between 7 and 14 years. The system requires two users playing simultaneously to engage in all possible interactions. Ideally Spark is used by pairs of children with one ASC child and one non-ASC child. The VE used in Spark builds upon a deep space theme. The interactive area is limited by a circular, fading white border. The achieved circular shape is meant to avoid the isolation of players in corners and to redirect players’ movements into the center of the area. Within this interactive area each child can interact with algorithmically generated elements by hovering their handheld controller close to an element. Through the combination of different elements or through performing tasks in synchrony with their co-player, the children proceed within the game. The completion of a subtask is accompanied by a visual effect and respective audio.
Spark engages children from both sexes equally;
Spark leads to a higher number of social initiations by ASC players than Lands of Fog; and
Spark leads to stronger intention of its players to engage in further play sessions.
Evaluation during COVID-19
Due to restrictions enacted in response to the COVID-19 pandemic, the final evaluation of Spark was based on a study with 18 non-ASC children. Pairs of non-ASC children thereby tested Spark in a ten-minute play session before independently reporting their engagement and intention to join future sessions through a digital post-play questionnaire. Additionally, semi-structured interviews with three psychologists yielded an assessment of Spark’s potential efficacy when being used by pairs formed by one ASC and one non-ASC child.
The results of the study show no significant difference in engagement between male and female non-ASC children while playing Spark. The psychologists attest Spark a comparable or even greater efficacy in fostering social initiations in ASC children than Lands of Fog—primarily attributable to Spark’s greater visual appeal, diversity of interactions and richness of stimuli. Furthermore, the study evinces a strong intention of non-ASC children to engage in future play-sessions of Spark.
On this basis, it is concluded that the alternative game design implemented in Spark allows overcoming the current limitations seen in Lands of Fog. In particular the usage of an abstract game theme and algorithmically generated stimuli support the intervention’s sex-neutrality, efficacy as well as its potential to be used in multiple consecutive sessions. Even though the study’s validity was diminished by testing with non-ASC children exclusively, the results indicate Spark’s positive potential as an intervention fostering social initiation in ASC children. Consequently, this thesis provides an affirmation for the expansion of research efforts to enhance the current design of Spark.