Baby Driver: Determine how the child drives the development of social interaction 

Zsofia is currently finalizing a review that explores the feedforward effects of infants’ face perception on vocabulary growth. An abstract of the review was submitted to the special edition of the Infant Behaviour and Development Journal at the end of February. Zsofia is also organizing a collaborative project with the EU-AIMS, a network that compares the  social and cognitive development of infants with either low or high likelihood of autism diagnosis. Zsofia’s project is looking at the differences in content of the vocabularies of high vs low likelihood infants at two timepoints (14 and 24 months of age). Zsofia is interested in whether differences are  observable in the words that high vs low likelihood infants can comprehend and produce prior to  diagnosis. Zsofia has presented preliminary data findings of this project at the NVP: The Dutch society for brain and cognition conference and has prepared a manuscript for submission.  

This project aims to study how the integration of vocal and motor production within the first year of life is related to later social functioning of a child. The project is focused on identifying motor and vocal patterns crucial for later social interaction outcomes. In order to detect early signs of Autism Spectrum Disorder (ASD), motor and vocal developmental trajectories were analysed using longitudinal data already collected by the Italian Network for early detection of ASD (NIDA Network) established in 2012 and involving all the pediatric hospitals of the National Health Service. 

The analysis conducted by Marzena were performed on 226 videos of 87 infants including five age groups. The analysis involved training a classifier to distinguish between typically developing (TD) children and those diagnosed with a Neurodevelopmental Disorders (NDD) at the age of 24/36 months, and to further determine the best time in development for conducting an early motor analysis. Results have showed that the motor features computed at the age of 10 days and 6 weeks were useful in terms of an early identification of NDD and might provide a better distinction between TD and NDD infants, than the same motor features analyzed for later stages of development (12, 18, and 24 weeks). The current work suggested that the RUSBoosted Trees classifier trained for the two youngest groups of infants (10 days and 6 weeks post-term) could be included in a system for early detection of NDD. In the future, other key points could be considered for tracking with Deeplabcut (DLC) to find out whether additional information could further improve the classification results. 

This study provides an example of DLC’s application, which holds significant promise for early assessment of neurodevelopmental delays in research and clinical settings, combined with gold standard tools. Infant motor development can be objectively quantified and can predict neurodevelopmental outcomes. This work is currently under review (Marzena Oliveira Ribas (previously Marzena Szkodo) et al., 2023). Future research should seek to promote the development of automatized tools that permit to detect possible neurodevelopmental deficits early enough to intervene promptly and improve future outcomes. Automatized tools may be extremely useful in clinical settings often lacking in human and economic resources. Marzena also published a systematic review that explored and summarized findings from 221 studies published between August 2019 and February 2022, covering a wide range of technologies used for diagnosis and/or treatment of NDD, with the biggest focus on ASD. The most popular technologies included machine learning, functional magnetic resonance imaging, electroencephalogram, magnetic resonance imaging, and neurofeedback. The results of the review indicate that technology-based diagnosis and intervention for NDD population is promising, but high-quality research is needed.

As part of her doctoral research, Kloe explored the neurophysiological and behavioural aspects associated with Autism Spectrum Disorder (ASD) traits in a diverse group of children from the local community. Four pivotal studies shaped her research journey. The structure of the thesis is as follows: two studies are based on backup-data and two are based on newly collected data. The rationale behind this approach stems from the unforeseen challenges posed by the COVID-19 pandemic, which necessitated the development of two backup projects that could seamlessly utilize existing data. Both of these projects were thoughtfully aligned with our original research protocol and objectives, focusing on the examination of how individual differences in infancy may exert influence on subsequent social interaction.

In her first study, she conducted a multi-level meta-analysis with the aim of synthesizing the existing knowledge on early face processing in ASD; over 300 children were included. The focus was on the N290/P400 complex, which serves as the developmental precursor of the N170 component observed in later childhood and adulthood. The main finding was that the P400 but not the N290 is smaller in children at risk of developing ASD (EL children). This outcome was unexpected, as it contradicted the hypothesis that the N290 component would be smaller in EL children. Notably, the low statistical power and evidence of publication bias might have contributed to the lack of significant findings for the N290 component. 

In her second study she examined the relationship between atypical sensory scores from the ITSP questionnaire (Infant/Toddler Sensory Profile) and later social behaviours in children, with a specific focus on individuals with high or low ASD traits. Specifically, she aimed to investigate whether sensory trajectory parameters at 10, 14, 24 and 36 months mediated the association between alpha asymmetry at 5, 10, 14 and 24 months and ASD outcome at the endpoint of those trajectories (i.e., 36 months). Longitudinal data analysis, such as mixed modelling and structural equation modelling were employed to explore the relationship between EEG power, sensory scores and ASD outcome. The repeated measures across months were nested under participants using a random intercept. The main finding was that ITSP intercept and slope measures did not mediate relationships between EEG from 5 to 24 months and outcome at 36 months. In addition, she explored the developmental trajectories of EEG measures in infants at low and elevated likelihood for ASD. The main finding was that there was an increase in power across all bands between 10 and 14 months of age for all infants, regardless of familiar status. Transitioning to the third and fourth chapters, the research shifted its focus to the data collected in accordance with the original research protocol. 

In the fourth chapter, she investigated to what extent do individual differences in EEG spectral power within specific frequency bands explain variations in social communication abilities and repetitive behaviours among children with varying levels of ASD traits. For each participant, she computed the variability in EEG spectral power by computing the standard deviation (SDs). Similarly, for ASD trait scores, she used SDs to assess how much individual trait scores deviate from the group mean. She didn’t find any significant associations. For the fourth chapter she investigated whether expressive language plays a moderating role in the relationship between peak gamma activity and ASD traits, providing insights into how cognitive development influences this association. However, the findings from this analysis did not reveal expressive language proficiency as a mediating factor.

In summary, these four pivotal studies collectively contribute to the ever-evolving understanding of the relationship between neurophysiological markers, sensory responsiveness, and ASD traits in children. The amalgamation of diverse methodologies, from meta-analysis to longitudinal data modelling, allows for a multifaceted exploration of these intricate associations, providing valuable insights into the complex developmental trajectories of ASD traits and their potential predictive factors.