The emerging dynamics of parent-infant interaction and its role in typical and atypical development: a dual eye tracking study
Early parent-infant interaction is a foundation for infant development. Further, transactional processes involving parent infant interaction may play a role in the development of clinical conditions such as Autism Spectrum Disorder (ASD). Traditional approaches for studying parent child interaction involve manual coding of video recordings, which is extremely time consuming and has limited precision in terms of fine-grained analysis of a key aspect of the interaction: looking patterns. Here, I propose a novel approach to studying early dyadic interaction, by the use of dual live eye tracking. This non-invasive method can be used to record eye movements of both infant and parent simultaneously, as they interact with each other. I will thus be able to, for the first time, provide a detailed characterization of gaze patterns during parent infant interaction, as well as to determine how these patterns relate to characteristics of both infant and parent. By taking a longitudinal approach, I will also be able to investigate how interactional gaze patterns evolve over time. Specifically, I will evaluate the recent hypothesis that infants actively shape their own social environment (parental behaviors) through their early looking patterns. Finally, by including both typically developing infants and infants at high risk for ASD, putative associations with later atypical development will be directly assessed.
Final report
Aim and development of the project
Autism spectrum condition (henceforth autism) is a neurodevelopmental condition that entails deficits in social interaction and the presence of restricted and repetitive behaviors. Autism has a global prevalence of 1-2 % and is rarely diagnosed before the age of 2-3 years, which entails that knowledge about early development is restricted. The main aim of this project was to investigate gaze patterns during interaction with adults in infants with and without elevated likelihood of autism. We aimed to study how infants and adults influence each other through eye gaze, as well as examine the potential role of early gaze behaviors in the development of autism. While the original aim was to focus on infant gaze during interaction with parents, this has been expanded to encompass interaction with unknown adults as well. The project has thus far resulted in four scientific publications, of which two in the leading journal Biological Psychiatry. The project is expected to generate more publications in the near future.
Execution
The project plan has largely been followed, but some changes have been made – both due to necessity and to strengthen the study design. The Covid-19 pandemic conveyed substantial delay in terms of data collection, and the longitudinal study of parent-infant interaction with dual gaze measures is therefore not yet finished. All data from the early time points (10, 14 and 18 months) is collected, but data on diagnostic outcome will not be available for all participants until the later part of 2024. In the early phases of the project, collaboration with the EASE project (Sw. Projekt Småsyskon), a larger project investigating the development of infants at elevated likelihood of autism and other developmental conditions, was established. The collaboration entails that we are able to follow children until 24 months, rather than until 18 months as originally planned. While this delays analysis slightly more, it is a great advantage as the 24-month measurement includes a full clinical assessment. We will thus be able to relate the early gaze measures to later diagnostic status, rather than to scores on a diagnostic test. The number of early measurement points has been increased from two to three (at 10, 14 and 18 months) which allows us to secure more data from each participating dyad and allows for better precision in detecting time effects. The collaboration with the EASE project has also allowed me to analyze existing data related to the main aim of the project. This has been a great advantage since the pandemic has limited the opportunities for novel data collection. The dual eye tracking technology that was initially planned did not generate data with sufficient reliability and precision. Therefore, the project has utilized a combination of ‘live eye tracking’, where the gaze of one individual is measured during interaction, and video coding of the gaze of two interacting individuals. This methodological change does not affect statistical analysis or the possibility to address the initial study questions.
Results and conclusions
As data on diagnostic outcome is not yet available for all participants of the longitudinal study with dual gaze measures, some analyses are not yet finished. This entails that some of the project’s hypotheses, including whether infants’ early gaze behaviors affect later parental gaze behaviors, have not yet been tested. A number of studies on infant gaze during interaction in relation to later autism have however been conducted, and I will here focus on the published results.
Social attention, i.e. attention to faces and other social stimuli, is typically measured in response to video-recorded stimuli of unknown people. The ecological validity of this has been questioned, and an aim for the current project was thus to investigate how much 10-month-olds attend to their parents’ faces during free play (Thorup et. al., 2024). The results showed that infants spent only ~3 % of their time attending to parents’ faces, which is much less compared to what screen-based studies typically report. In addition, infants with an older sibling with an autism diagnosis – and hence an elevated likelihood of receiving a later diagnosis as well – made fewer gaze shifts toward their parents’ faces than infants with typical likelihood of autism. This could not be explained by differences in terms of autistic traits in the participating parents. There were also no differences between those infants at elevated likelihood who were later diagnosed with autism and those infants at elevated likelihood who were not. The study contributes to the field by providing insight into how infants look at their parents during natural interaction.
A different study assessed gaze following between 10 and 18 months in relation to later autism diagnosis (Thorup et. al., 2021). The focus of this study was on gaze following in the absence of salient target objects. Performance was compared between a condition in which an adult looked towards an empty hole and a condition in which a puppet emerged in the hole. The results showed that whereas infants with later autism followed gaze to the same degree as other infants in all conditions, they had shorter latencies to look back at the face of the adult when the puppet was missing. This may indicate a different appreciation of the communicative function of gaze in infants with later autism.
In a related study, the development of initiation as well as response to joint attention was assessed in relation to later autism (Nyström, Thorup, 2019). The results showed no differences in terms of early response to joint attention (measured as gaze following) and later autism. However, infants with later autism initiated joint attention (measured as alternating gaze between an adult and an object) to a lower degree than other infants at 10 months, as well as displayed a different developmental trajectory of the behavior between 10 and 18 months. These results suggest that initiation of joint attention might be an important target for early intervention.
In addition to the empirical studies, the project has enabled the writing of a review article of social attention (largely based on studies of the type of gaze behaviors that the project concerns) in relation to autism and early development (Falck-Ytter et al., 2023).
Novel research questions
Future studies on parent-infant interaction in relation to later autism should aim to investigate different explanatory models. We have shown that infants at elevated likelihood of autism make fewer gaze shifts towards parents’ faces than infants at typical likelihood of autism (Thorup et. al., 2024) but the question of why remains open. Is the less frequent face looking in infants at elevated likelihood of autism caused by autistic traits – which may be present also in undiagnosed family members of autistic individuals (Broader Autism Phenotype) – in the infants? Or are infant gaze behaviors a consequence of differences in parental behaviors? Parents in the elevated likelihood group may for example have adjusted their interaction styles to their older autistic child, which may spill over on interaction with the younger child as well. Another important focus for future research will be to find clinical applications for findings such as those generated by the present project. May such findings be used to develop programs for early intervention, and may participation in such intervention programs alter infant development?
Communication of results and collaboration
The project has been conducted in collaboration with the EASE project, led by Professor Terje Falck-Ytter, and with Karolinska Institutet Center of Neurodevelopmental Disorders (KIND), led by Professor Sven Bölte. For methods development, we have collaborated with Roy Hessels, Ignace Hooge, and Niilo Valtakari at the University of Utrechts, the Netherlands. Results have been published in scientific journals with Open Access, as well as shared on social media. Results have been presented at the international conference International Society for Autism Research (INSAR) in Austin, Texas, 2022, and Stockholm, Sweden, 2023. Results have also been presented at seminars at the department of Psychology at Lund University.
Autism spectrum condition (henceforth autism) is a neurodevelopmental condition that entails deficits in social interaction and the presence of restricted and repetitive behaviors. Autism has a global prevalence of 1-2 % and is rarely diagnosed before the age of 2-3 years, which entails that knowledge about early development is restricted. The main aim of this project was to investigate gaze patterns during interaction with adults in infants with and without elevated likelihood of autism. We aimed to study how infants and adults influence each other through eye gaze, as well as examine the potential role of early gaze behaviors in the development of autism. While the original aim was to focus on infant gaze during interaction with parents, this has been expanded to encompass interaction with unknown adults as well. The project has thus far resulted in four scientific publications, of which two in the leading journal Biological Psychiatry. The project is expected to generate more publications in the near future.
Execution
The project plan has largely been followed, but some changes have been made – both due to necessity and to strengthen the study design. The Covid-19 pandemic conveyed substantial delay in terms of data collection, and the longitudinal study of parent-infant interaction with dual gaze measures is therefore not yet finished. All data from the early time points (10, 14 and 18 months) is collected, but data on diagnostic outcome will not be available for all participants until the later part of 2024. In the early phases of the project, collaboration with the EASE project (Sw. Projekt Småsyskon), a larger project investigating the development of infants at elevated likelihood of autism and other developmental conditions, was established. The collaboration entails that we are able to follow children until 24 months, rather than until 18 months as originally planned. While this delays analysis slightly more, it is a great advantage as the 24-month measurement includes a full clinical assessment. We will thus be able to relate the early gaze measures to later diagnostic status, rather than to scores on a diagnostic test. The number of early measurement points has been increased from two to three (at 10, 14 and 18 months) which allows us to secure more data from each participating dyad and allows for better precision in detecting time effects. The collaboration with the EASE project has also allowed me to analyze existing data related to the main aim of the project. This has been a great advantage since the pandemic has limited the opportunities for novel data collection. The dual eye tracking technology that was initially planned did not generate data with sufficient reliability and precision. Therefore, the project has utilized a combination of ‘live eye tracking’, where the gaze of one individual is measured during interaction, and video coding of the gaze of two interacting individuals. This methodological change does not affect statistical analysis or the possibility to address the initial study questions.
Results and conclusions
As data on diagnostic outcome is not yet available for all participants of the longitudinal study with dual gaze measures, some analyses are not yet finished. This entails that some of the project’s hypotheses, including whether infants’ early gaze behaviors affect later parental gaze behaviors, have not yet been tested. A number of studies on infant gaze during interaction in relation to later autism have however been conducted, and I will here focus on the published results.
Social attention, i.e. attention to faces and other social stimuli, is typically measured in response to video-recorded stimuli of unknown people. The ecological validity of this has been questioned, and an aim for the current project was thus to investigate how much 10-month-olds attend to their parents’ faces during free play (Thorup et. al., 2024). The results showed that infants spent only ~3 % of their time attending to parents’ faces, which is much less compared to what screen-based studies typically report. In addition, infants with an older sibling with an autism diagnosis – and hence an elevated likelihood of receiving a later diagnosis as well – made fewer gaze shifts toward their parents’ faces than infants with typical likelihood of autism. This could not be explained by differences in terms of autistic traits in the participating parents. There were also no differences between those infants at elevated likelihood who were later diagnosed with autism and those infants at elevated likelihood who were not. The study contributes to the field by providing insight into how infants look at their parents during natural interaction.
A different study assessed gaze following between 10 and 18 months in relation to later autism diagnosis (Thorup et. al., 2021). The focus of this study was on gaze following in the absence of salient target objects. Performance was compared between a condition in which an adult looked towards an empty hole and a condition in which a puppet emerged in the hole. The results showed that whereas infants with later autism followed gaze to the same degree as other infants in all conditions, they had shorter latencies to look back at the face of the adult when the puppet was missing. This may indicate a different appreciation of the communicative function of gaze in infants with later autism.
In a related study, the development of initiation as well as response to joint attention was assessed in relation to later autism (Nyström, Thorup, 2019). The results showed no differences in terms of early response to joint attention (measured as gaze following) and later autism. However, infants with later autism initiated joint attention (measured as alternating gaze between an adult and an object) to a lower degree than other infants at 10 months, as well as displayed a different developmental trajectory of the behavior between 10 and 18 months. These results suggest that initiation of joint attention might be an important target for early intervention.
In addition to the empirical studies, the project has enabled the writing of a review article of social attention (largely based on studies of the type of gaze behaviors that the project concerns) in relation to autism and early development (Falck-Ytter et al., 2023).
Novel research questions
Future studies on parent-infant interaction in relation to later autism should aim to investigate different explanatory models. We have shown that infants at elevated likelihood of autism make fewer gaze shifts towards parents’ faces than infants at typical likelihood of autism (Thorup et. al., 2024) but the question of why remains open. Is the less frequent face looking in infants at elevated likelihood of autism caused by autistic traits – which may be present also in undiagnosed family members of autistic individuals (Broader Autism Phenotype) – in the infants? Or are infant gaze behaviors a consequence of differences in parental behaviors? Parents in the elevated likelihood group may for example have adjusted their interaction styles to their older autistic child, which may spill over on interaction with the younger child as well. Another important focus for future research will be to find clinical applications for findings such as those generated by the present project. May such findings be used to develop programs for early intervention, and may participation in such intervention programs alter infant development?
Communication of results and collaboration
The project has been conducted in collaboration with the EASE project, led by Professor Terje Falck-Ytter, and with Karolinska Institutet Center of Neurodevelopmental Disorders (KIND), led by Professor Sven Bölte. For methods development, we have collaborated with Roy Hessels, Ignace Hooge, and Niilo Valtakari at the University of Utrechts, the Netherlands. Results have been published in scientific journals with Open Access, as well as shared on social media. Results have been presented at the international conference International Society for Autism Research (INSAR) in Austin, Texas, 2022, and Stockholm, Sweden, 2023. Results have also been presented at seminars at the department of Psychology at Lund University.