Autism Spectrum Disorder (ASD) is a highly heterogeneous neurodevelopmental disorders with multiple causes, courses, and a wide range in symptom severity. Although the etiology of the disorder is generally considered multifactorial, high heritability estimates suggest a critical role for genetic factors. However, the notable clinical heterogeneity within the broad behavioral phenotype has been a major obstacle to gene identification. Furthermore, the hallmark heterogeneity of ASD makes the quest for personalized treatment and potential precision medicine inherently difficult. Starting from these considerations, in the last years there has been an increasing need for developing a reliable marker for ASD, currently diagnosed on the basis of the clinical judgment of symptoms. The purpose of the present project was to provide further evidence supporting the use of motor impairments as a bio-behavioral marker of ASD. Indeed, abnormalities in motor behavior are one of the features most frequently associated to the disorder and can have a significant impact on quality of life and social development. Specifically, in this thesis we investigated the gait pattern and the motor adaptation to discrete gait perturbations in school-aged children with ASD using an innovative multi-sensor platform based on immersive virtual reality (Chapter 1). Further, we developed a supervised machine-learning method to identify and correctly discriminate preschool children with ASD from typically developing children by means of kinematic analysis of a simple reach, grasp and drop task (Chapter 2). Finally, using diffusion tensor imaging (DTI), we explored the hypothesis of reduced long-range connectivity between frontal lobes and posterior brain regions in ASD given the key role of these pathways for language, praxis, imitation, and basic motor coordination (Chapter 3). Results highlighted an altered gait pattern in children with ASD and slower rates of adaptation to the perturbation. Diminished learning adaptation was also significantly related with more severe autistic traits. With respect to classification based on kinematics analysis, our machine-learning method reached a good mean individual classification in the comparisons between children with ASD and healthy controls (overall mean accuracy = 84.9%). Thus, we demonstrated that machine-learning classification approach might be helpful for supporting the clinical practice of diagnosing ASD, even fostering a computer-aided diagnosis perspective. Last, the DTI study provided evidence for alterations in white matter diffusivity of the left superior longitudinal fasciculus in a well-characterized group of high-functioning children with ASD. All in all, our findings offer insight on a possible, multi-domain (i.e., behavioral, computational, and imaging) motor signature of autism that is potentially useful to identify a well-defined subset of patients, thus reducing the clinical heterogeneity within the broad behavioral phenotype. This may guide further exploration of neuropathology of the disorder adding power to genetic analysis.

I Disturbi dello Spettro Autistico (in inglese Autism Spectrum Disorders, ASD) rappresentano un insieme di disturbi del neurosviluppo fortemente eterogenei nella loro manifestazione, con una molteplicità di cause, differenti percorsi evolutivi e livelli di compromissione funzionale. L’alta ereditarietà degli ASD porta a considerare una forte implicazione di fattori genetici nella sua eziologia, ma l’ampia eterogeneità clinica all’interno del fenotipo comportamentale risulta essere un importante ostacolo nell’identificazione dei geni coinvolti. Inoltre, la eterogeneità tipica degli ASD rende estremamente difficoltosa la ricerca di trattamenti personalizzati e di una medicina di precisione. A partire da queste considerazioni, negli ultimi anni si è sviluppata una crescente richiesta di sviluppare marcatori affidabili per un disturbo come ASD, attualmente diagnosticato sulla base di una valutazione clinica dei sintomi. Lo scopo del presente progetto è stato quello di fornire ulteriori evidenze scientifiche in supporto dell’utilizzo di deficit motori come marcatore comportamentale e biologico degli ASD. Le difficoltà motorie sono infatti una delle caratteristiche più frequentemente associate al disturbo, con un’importante prevalenza ed un impatto significativo sulla qualità della vita e sullo sviluppo sociale del bambino. Precisamente, in questa tesi abbiamo indagato il pattern di cammino e l’adattamento motorio ad una perturbazione discreta in bambini con ASD usando una innovativa piattaforma integrata basata su realtà virtuale immersiva (Cap. 1). Abbiamo inoltre sviluppato un metodo di classificazione basato su un approccio machine learning per confrontare la cinematica del movimento di raggiungimento ed inserimento di una pallina in bambini con e senza ASD (Cap. 2). Infine, usando l’imaging di diffusione del tensore (DTI), abbiamo esplorato l’ipotesi di una ridotta conettività a lungo raggio tra le regioni cerebrali frontali e posteriori, dato il ruolo cruciale di tali connessioni in funzioni quali il linguaggio, la prassia, l’imitazione e la coordinazione motoria (Cap. 3). I risultati hanno mostrato un pattern di cammino alterato in bambini con ASD e minore adattamento alla perturbazione. Tale ridotto livello di adattamento è risultato relato in modo significativo a una maggior gravità della sintomatologia autistica. Relativamente alla classificazione basata su cinematica, il nostro metodo machine-learning si è dimostrato efficace nel discriminare bambini con ASD da bambini con sviluppo tipico con un’accuratezza media pari al 84.9%. L’identificazione di queste caratteristiche potrebbe essere potenzialmente utile nel supportare e facilitare le diagnosi clinica dei Disturbi dello Spettro Autistico. Infine lo studio DTI ha rivelato alterazioni nella diffusività della materia bianca a livello del fascicolo longitudinale superiore sinistro in un gruppo di bambini con ASD ad alto funzionamento. Nel complesso, i risultati della presente tesi supportano l’ipotesi di un possibile marcatore motorio di ASD multi-dominio (i.e., comportamentale, computazionale e di neuroimmagine), potenzialmente utile per identificare specifico endofenotipo all’interno della sindrome autistica. Tale caratterizzazione può aumentare sensibilmente il potere predittivo di indagini rispetto modelli eziopatogenetici del disturbo, mediante studi di neuroimmagine e analisi genetica.

(2018). MOTOR IMPAIRMENTS AS A BIO-BEHAVIORAL MARKER OF AUTISM SPECTRUM DISORDER. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2018).

MOTOR IMPAIRMENTS AS A BIO-BEHAVIORAL MARKER OF AUTISM SPECTRUM DISORDER

CRIPPA, ALESSANDRO
2018

Abstract

Autism Spectrum Disorder (ASD) is a highly heterogeneous neurodevelopmental disorders with multiple causes, courses, and a wide range in symptom severity. Although the etiology of the disorder is generally considered multifactorial, high heritability estimates suggest a critical role for genetic factors. However, the notable clinical heterogeneity within the broad behavioral phenotype has been a major obstacle to gene identification. Furthermore, the hallmark heterogeneity of ASD makes the quest for personalized treatment and potential precision medicine inherently difficult. Starting from these considerations, in the last years there has been an increasing need for developing a reliable marker for ASD, currently diagnosed on the basis of the clinical judgment of symptoms. The purpose of the present project was to provide further evidence supporting the use of motor impairments as a bio-behavioral marker of ASD. Indeed, abnormalities in motor behavior are one of the features most frequently associated to the disorder and can have a significant impact on quality of life and social development. Specifically, in this thesis we investigated the gait pattern and the motor adaptation to discrete gait perturbations in school-aged children with ASD using an innovative multi-sensor platform based on immersive virtual reality (Chapter 1). Further, we developed a supervised machine-learning method to identify and correctly discriminate preschool children with ASD from typically developing children by means of kinematic analysis of a simple reach, grasp and drop task (Chapter 2). Finally, using diffusion tensor imaging (DTI), we explored the hypothesis of reduced long-range connectivity between frontal lobes and posterior brain regions in ASD given the key role of these pathways for language, praxis, imitation, and basic motor coordination (Chapter 3). Results highlighted an altered gait pattern in children with ASD and slower rates of adaptation to the perturbation. Diminished learning adaptation was also significantly related with more severe autistic traits. With respect to classification based on kinematics analysis, our machine-learning method reached a good mean individual classification in the comparisons between children with ASD and healthy controls (overall mean accuracy = 84.9%). Thus, we demonstrated that machine-learning classification approach might be helpful for supporting the clinical practice of diagnosing ASD, even fostering a computer-aided diagnosis perspective. Last, the DTI study provided evidence for alterations in white matter diffusivity of the left superior longitudinal fasciculus in a well-characterized group of high-functioning children with ASD. All in all, our findings offer insight on a possible, multi-domain (i.e., behavioral, computational, and imaging) motor signature of autism that is potentially useful to identify a well-defined subset of patients, thus reducing the clinical heterogeneity within the broad behavioral phenotype. This may guide further exploration of neuropathology of the disorder adding power to genetic analysis.
MARZOCCHI, GIAN MARCO
Autism; Spectrum; Disorder; motor; impairments
Autism; Spectrum; Disorder; motor; impairments
M-PSI/04 - PSICOLOGIA DELLO SVILUPPO E PSICOLOGIA DELL'EDUCAZIONE
English
20-feb-2018
PSICOLOGIA, LINGUISTICA E NEUROSCIENZE COGNITIVE - 77R
30
2016/2017
open
(2018). MOTOR IMPAIRMENTS AS A BIO-BEHAVIORAL MARKER OF AUTISM SPECTRUM DISORDER. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2018).
File in questo prodotto:
File Dimensione Formato  
phd_unimib_070619.pdf

Accesso Aperto

Descrizione: tesi di dottorato
Tipologia di allegato: Doctoral thesis
Dimensione 1.73 MB
Formato Adobe PDF
1.73 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/199065
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact