---

PATHWAYS EARLY INTERVENTION PROJECTS 

The Pathways projects focus on evaluating the efficacy of the first seven units of the Pathways parent-mediated intervention (Pathways). Pathways is a naturalistic developmental behavioral intervention (NDBI) that aligns with the service delivery model for early intervention programs in Texas (IDEA Part C). The first seven units of Pathways are designed to coach parents on how to facilitate early reciprocal social interaction and mutual gaze, which are critical components of early dyadic interactions in infants with typical development. These interactions may activate the social brain network (Johnson et al., 2015; Jones & Klin, 2013; Mundy & Bullen, 2022; Senju & Johnson, 2009).

Unlike typically developing infants, autistic infants often experience a decline in eye gaze between two and six months, which suggests they may miss opportunities to engage in mutual gaze and foundational social interactions. Interventions that specifically target mutual gaze, such as Pathways, may shift brain activity patterns toward a more social trajectory (Johnson et al., 2015; Jones & Klin, 2013; Mundy & Bullen, 2022; Senju & Johnson, 2009).

The Social Communication Lab has been studying the efficacy of these first seven units from the original Pathways manual. Our findings indicate that, compared to children receiving community early intervention services, Pathways enhances social skills, social communication, language, and adaptive functioning. Although the effects of Pathways on social communication are more pronounced in children under three years than in those over three (Rollins & De Froy, 2022), there was no age-related effect on social attention (Patel & Rollins, under review).

Current Studies

ROLE OF MUTUAL GAZE

Funded by the Texas Higher Education Coordinating Board, Autism Grants Program
This project seeks to expand our previous work by investigating whether the mutual gaze procedure in the Pathways Early Autism Intervention is the key factor for improving core social challenges and adaptive functioning in very young, culturally and linguistically diverse (CLD) autistic children. Specifically, we will evaluate whether mutual gaze serves as the “active ingredient” for enhancing social development and adaptive functioning in CLD children with social interaction and social communication challenges who are at risk for autism. 

A novel aspect of our study is the use of EarliPoint, a newly FDA-authorized portable eye-tracking tool that provides a social disability index for autism. Eighty children with social interaction and social communication challenges, aged 16–30 months, and their families will be randomly assigned to either the experimental group (Pathways with mutual gaze) or the control group (Pathways without mutual gaze). All families will receive 12 Pathways sessions over a 15-week period. Assessments will be conducted at baseline, post-intervention, and three months post-intervention to measure change over time.

Need: This project aims to accelerate intervention for very young autistic children and address a critical mechanistic question: What is the impact of mutual gaze on social development? The results will guide clinical decision-making for Early Childhood Intervention (ECI) providers implementing Pathways and reduce barriers to autism intervention for young CLD children.

CULTURALLY AND LINGUISTICALLY RESPONSIVE EARLY AUTISM INTERVENTION

Funded by a US Department of Defense Congressionally Mandated Medical Research Clinical Trial Award.  

Hispanic autistic children, particularly those from low-income families or whose primary language is Spanish, face significant contextual, linguistic, and cultural barriers to accessing early intervention services (Funk & Lopez, 2022). For Aim 1, we will develop a culturally and linguistically responsive adaptation of the Pathways parent-mediated early autism intervention (CLR-Pathways) tailored for young Hispanic autistic children from low-income households. To create this adaptation, we are collaborating with an expert stakeholder group composed of Hispanic and bilingual (Spanish/English) community partners and parents of autistic children.

Similar to the original Pathways, CLR-Pathways will consist of a low-dose (90-minute weekly visits) manualized parent-mediated naturalistic developmental behavioral intervention (NDBI) that aligns with the service delivery model outlined by the Individuals with Disabilities Education Act (IDEA) and Early Childhood Intervention (ECI). For Aim 2, we will assess whether the mutual gaze component of CLR-Pathways is the active ingredient responsible for improving social attention, social communication, language, and adaptive functioning. Two hundred Hispanic autistic children from low-income families and their caregivers will be randomly assigned—like flipping a coin—into one of two groups: CLR-Pathways with the mutual gaze component (experimental condition) or without it (control condition).

This project aims to enhance early autism intervention for young Hispanic autistic children by addressing critical barriers and identifying the key elements that drive social and communication improvements.

AI-ASSISTED GAZE TARGET DETECTION FOR AUTISTIC CHILDREN

Funded by a UT Dallas Seed Program for Interdisciplinary Research (SPIRe) award

This project aims to develop an advanced AI system that automatically detects gaze targets in videos of autistic children, enabling a precise, objective, and scalable assessment of social attention behaviors. The videos will be sourced from the Social Communication lab. For model training, we will use CVAT, an efficient video annotation tool, to annotate bounding boxes around the children’s faces and their gaze targets. While existing object detection datasets are vast, they lack specialized annotations for children’s faces. To address this gap, we will fine-tune RTMDet, a cutting-edge object detection model, and implement Dr. Yapeng Tian’s approach to cross-modal object localization. By using the child’s face as a query, we will encode its features to dynamically search across the entire frame, mimicking the human ability to interpret where a child is looking. This AI system has the potential to revolutionize the measurement of early social attention behaviors.