PORTEFEUILLE

AutoPlay project goal is to give emphasis to the importance of “play” since the very early age. It will establish an innovative and powerful methodology for objective and free-of-context conditioning observation of infants' (9-15 months) ludic development. The long-term objective of AutoPlay is reaching a social systematic change of how ludic behaviour is observed, analysed and considered for evaluating and identifying early signs of neurodevelopmental disorders and social problems.
The project presents a novel and multidisciplinary approach that integrates social, health, technical, informatics, and statistical analysis expertise, aiming at defining a classification methodology for monitoring the toy-usage in infants. To achieve this goal, it exploits a set of augmented toys instrumented with specific sensors, and finally used by infants in a natural ludic environment, in order to characterize playing activities and to identify possible anomalies.
AutoPlay aims at being adopted in the regular practice, by paediatricians and kindergartens, with the goal of detecting children with possible abnormal ludic development at the early stage of life and of facilitating an anticipated intervention on problematic ones. The sooner the better: if a social problem or a neuro-developmental disorder is diagnosed in the very early phases it can be treated properly ensuring a better prognosis and minimizing social negative impact. Autism Spectrum Disorders (ASD) are nowadays one of the main development disabilities, the incidence is very high, 2 - 6 cases every 1000 children and in some references 1 in every 68 children at around 8 years. The public cost for person with disabilities and social integration problems is increasing more and more in the last years: an early intervention can then of course have a social impact but can also lead to minimize the financial cost for the public structures and the society itself (in particular the families). Thanks to this tool-kit we could help many children: helping them means helping their families and finally the society in which we all live.
Additionally, a good ludic culture is often missing within families and among people; the project will help to disseminate a good play culture by means of local and national events and by means of electronic communication (website, forum, a.o.).

To date, many models have been proposed to describe the relationship between play and child development, and psychologists have been extensively interested in children's ludic development, but the knowledge related to the play for infants (less than 2 years old) is still scarce. In general play has been extensively studied with a strong focus on the child's cognitive and socio-affective development. The majority of the studies that analyse the sensory-motor development regard children of more than two years. Ludic sensory-motor exploration development in infants has not been studied yet with the appropriate care. Knowing exactly how an infant approaches the toys, as well as how he/she plays, would be of great interest to the general knowledge of the children development. “How the small child interacts with a toy?” and “How this interaction evolves over time?” are still open questions that cannot be objectively answered, yet.
The information collected with AutoPlay toolkit will increase general knowledge on play and will add complementary criteria to determine a more precise diagnosis.
Nowadays ludic behaviour observation is not regularly adopted in neurodevelopmental disorders diagnosis. The AutoPlay methodology can prevent social exclusion of families with social fragility, thanks to an early intervention, in order to ensure an improved quality of life. We foresee a long-term social and financial impact, as a consequence of a smoother and better integration into the society. In this way, the project supports the concept of prevention of possible social problems, before the real intervention: this approach is twofold beneficial from a social point of view, in fact it helps in reducing the cost of the social intervention, and most important it enhances the quality of life of the involved small children and families.

The innovation of the AutoPlay project lies in the development of algorithms that have the particularity of measuring the movement of the object and not the way the person uses it. Indeed, projects that look at "activity recognition-machine learning" are often oriented towards the movement that the person is doing. In the case of AutoPlay, the information derives from the movement of the object that originates from the manipulation of the child. It is therefore through the object that the developed algorithm recognizes the playful development (manipulation) of the child. This information becomes clinically important. Indeed, the information is virgin of the motor development of the child (very variable in early age) and is interested in the movements that the child causes to the object and therefore to its exploratory and manipulative intention of the game. Its motor skills are not fundamental and do not interfere with the results. The focus of AutoPlay is to understand the playful intent of the child through his manipulation of the object.
To date we have proofed that data collected with the tool-kit can have a clinical interpretation that makes sense and it is relevant from a social and clinical point of view. Currently, we are testing the complete set of toys into the two kindergartens. Next steps are the start of the clinical phases of the project. The next two phases will be the small pilots with children at risk and the clinical trial before the development of the system as a commercial product.
A first clinical phase, funded by SUPSI and the ABREOC fund, a research fund for the Cantonal Hospital of Ticino (Professor Ramelli's neuropediatric department), is in course. This project tests AutoPlay Toy kit with children who have developmental peculiarities to see if the developed algorithms can be relevant to support the neuropediatrician in establishing a differential diagnosis. Two other clinical projects are under development. The first involves a research group from Messina (Italy) and the second involves several experts in autism from Geneva (Swiss). Both projects include two kinds of population: siblings under 12 months (children named at risk) and children with autism (aged between 18 months and three years old).
The research team with the designers will in the near future addresss as well the commercial aspects related to the sale of the prototype in order to raise fundamental data - thanks to the collaboration with other research teams - to actually enter the market. The Messina team has become a fundamental partner since it works with an internationally recognized electronic component developer. Discussions are underway for the development of an electronic component that would be more relevant to the objectives of AutoPlay and other projects that follow the same direction.