Support of neurodevelopmental difficulties in school-age children remains a critical challenge. In Italy, despite a clear legislative framework (Law 170/2010), a persistent gap exists between the intent of early identification and training and the tools available to schools to make it happen: teachers observe but lack structured instruments, clinical services assess but have limited access to ecological behavioral data, and families are often caught between disconnected systems. In this context, this PhD research aimed to design, develop, and evaluate an integrated technological ecosystem to support early identification and training of neurodevelopmental difficulties in primary schools, with a particular focus on handwriting development and the early detection of graphomotor weaknesses. The work spans multiple interconnected contributions: a suite of serious games for developmental monitoring, deployed with 66 children in ecological classroom settings; a unified platform architecture integrating tablet games, sensorized pen data, and teacher observations; a collaborative multiplayer exergame for inclusive physical activity, tested with both typically developing children and adults with intellectual disabilities; and four targeted handwriting training games, evaluated through a three-arm longitudinal controlled trial with 210 first-grade children. All solutions were co-designed with the relevant stakeholders -- teachers, clinicians, children, and adults with disabilities -- following an adaptation of the CeHRes Roadmap 2.0, and iteratively refined through pilot and field testing. The design approach was deliberately functional rather than diagnostic: accessibility was achieved by addressing ability dimensions, not diagnostic labels, and the same transdiagnostic principle informed the handwriting games and the AI models built on their data. A large longitudinal dataset of gameplay-derived behavioural features was collected, which served as the basis for leveraging machine learning techniques to predict clinical risk for handwriting difficulties and to characterise developmental profiles through clustering and explainability analyses. The results suggest that these are promising solutions to support the early screening of neurodevelopmental difficulties -- handwriting weaknesses and dysgraphia risk, in particular -- and to provide schools with a complementary, data-driven layer of information that enriches professional judgment rather than replacing it, in a technological ecosystem at the service of a pedagogical vision of learning.
Il supporto alle difficoltà del neurosviluppo nei bambini in età scolare rimane una sfida critica. In Italia, nonostante un quadro legislativo chiaro (Legge 170/2010), persiste un divario significativo tra le intenzioni di identificazione precoce e potenziamento e gli strumenti realmente disponibili nelle scuole per realizzarle: gli insegnanti osservano ma mancano di strumenti strutturati, i servizi clinici valutano ma hanno accesso limitato a dati comportamentali ecologici, e le famiglie si trovano spesso intrappolate tra sistemi disconnessi. In questo contesto, la presente ricerca di dottorato ha avuto l'obiettivo di progettare, sviluppare e valutare un ecosistema tecnologico integrato per supportare l'identificazione precoce e il potenziamento delle difficoltà del neurosviluppo nella scuola primaria, con particolare attenzione allo sviluppo della scrittura manuale e all'individuazione precoce di debolezze grafomotorie. Il lavoro comprende molteplici contributi interconnessi: una suite di serious game per il monitoraggio dello sviluppo, condotta con 66 bambini in contesti ecologici di classe; un'architettura di piattaforma unificata che integra giochi su tablet, dati provenienti da una penna sensorizzata e osservazioni degli insegnanti; un exergame collaborativo multiplayer per l'attività fisica inclusiva, testato sia con bambini a sviluppo tipico che con adulti con disabilità intellettiva; e quattro giochi mirati per il potenziamento della scrittura, valutati attraverso uno studio controllato longitudinale a tre bracci con 210 bambini di prima elementare. Tutte le soluzioni sono state co-progettate con gli attori coinvolti — insegnanti, clinici, bambini e adulti con disabilità — seguendo un adattamento della CeHRes Roadmap 2.0, e affinate iterativamente attraverso test pilota e sul campo. L'approccio progettuale è stato deliberatamente funzionale piuttosto che diagnostico: l'accessibilità è stata perseguita affrontando le dimensioni delle abilità, non le etichette diagnostiche, e lo stesso principio transdiagnostico ha guidato i giochi di scrittura e i modelli di intelligenza artificiale costruiti sui loro dati. È stato raccolto un ampio dataset longitudinale di feature comportamentali derivate dal gameplay, che ha costituito la base per l'applicazione di tecniche di machine learning volte a predire il rischio clinico per le difficoltà di scrittura e a caratterizzare i profili di sviluppo attraverso analisi di clustering e di spiegabilità. I risultati suggeriscono che queste rappresentano soluzioni promettenti per supportare lo screening precoce delle difficoltà del neurosviluppo — in particolare delle debolezze nella scrittura manuale e del rischio di disgrafia — e per fornire alle scuole un livello complementare di informazione data-driven che arricchisce il giudizio professionale senza sostituirlo, in un ecosistema tecnologico al servizio di una visione pedagogica dell'apprendimento.
An integrated co-designed ecosystem for preclinical screening and personalized training in primary schools: supporting a pedagogical vision of child development
PIAZZALUNGA, CHIARA
2026
Abstract
Support of neurodevelopmental difficulties in school-age children remains a critical challenge. In Italy, despite a clear legislative framework (Law 170/2010), a persistent gap exists between the intent of early identification and training and the tools available to schools to make it happen: teachers observe but lack structured instruments, clinical services assess but have limited access to ecological behavioral data, and families are often caught between disconnected systems. In this context, this PhD research aimed to design, develop, and evaluate an integrated technological ecosystem to support early identification and training of neurodevelopmental difficulties in primary schools, with a particular focus on handwriting development and the early detection of graphomotor weaknesses. The work spans multiple interconnected contributions: a suite of serious games for developmental monitoring, deployed with 66 children in ecological classroom settings; a unified platform architecture integrating tablet games, sensorized pen data, and teacher observations; a collaborative multiplayer exergame for inclusive physical activity, tested with both typically developing children and adults with intellectual disabilities; and four targeted handwriting training games, evaluated through a three-arm longitudinal controlled trial with 210 first-grade children. All solutions were co-designed with the relevant stakeholders -- teachers, clinicians, children, and adults with disabilities -- following an adaptation of the CeHRes Roadmap 2.0, and iteratively refined through pilot and field testing. The design approach was deliberately functional rather than diagnostic: accessibility was achieved by addressing ability dimensions, not diagnostic labels, and the same transdiagnostic principle informed the handwriting games and the AI models built on their data. A large longitudinal dataset of gameplay-derived behavioural features was collected, which served as the basis for leveraging machine learning techniques to predict clinical risk for handwriting difficulties and to characterise developmental profiles through clustering and explainability analyses. The results suggest that these are promising solutions to support the early screening of neurodevelopmental difficulties -- handwriting weaknesses and dysgraphia risk, in particular -- and to provide schools with a complementary, data-driven layer of information that enriches professional judgment rather than replacing it, in a technological ecosystem at the service of a pedagogical vision of learning.| File | Dimensione | Formato | |
|---|---|---|---|
|
Piazzalunga_PhD_revised.pdf
embargo fino al 15/05/2029
Licenza:
Tutti i diritti riservati
Dimensione
75.46 MB
Formato
Adobe PDF
|
75.46 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/376640
URN:NBN:IT:POLIMI-376640