BackHome
Refresh: OffView Personal Schedule
Individual Submission Summary
Share...

Direct link:

Poster #45 - Automated video tracking biomarkers of parent-child interaction correlate with children’s autism symptom severity

Sat, March 23, 9:45 to 11:00am, Baltimore Convention Center, Floor: Level 1, Exhibit Hall B

Integrative Statement

Autism Spectrum Disorders (ASD) are characterized by deficits in social communication and restricted and repetitive behaviors. Current measures of diagnosis and response to intervention include clinical assessments and coding of parent-child interactions (PCIs) (e.g., Kasari et al., 2010), which require trained clinicians and raters, and are subjective, time-consuming, and difficult to implement for large multi-site trials. Automated, objective measures that quantify behaviors during PCIs will advance early identification and intervention research.
Recently, one study explored automated tracking of children with ASD during a PCI in which the parent remained seated while the child played (Cohen et al., 2014). Associations between parent- and clinician-rated measures of ASD and tracking were found. Additional research is needed to assess the validity of video tracking in more naturalistic contexts (e.g., where the parent moves freely). The purpose of this exploratory study was to: 1) identify and compare tracking metrics for different PCI tasks; and 2) explore associations of these tracking metrics with standard measures of ASD symptoms.
We tracked parents and children with ASD (N=78, Mean Age=64.4 months, Range =29-96 months) during a two-part, 12-minute, standardized PCI. For Task 1, parents remained stationary while children played. For Task 2, both were free to move and play together. Parents and clinicians completed measures of child symptoms and skills (ADOS-2; ADI-R; Clinical Global Impressions-Severity; Vineland; Pervasive Developmental Disorder Behavior Inventory; Differential Ability Scales/Mullen Scales of Early Learning).
Location and activity data were generated with Noldus EthoVision XT 11.5. Associations between tracking outcomes (e.g., proportion of time moving and in activity zones; average velocity; proximity to parent) and clinical measures of child symptoms and skills were tested with Spearman correlations. Data analyses are ongoing and behavioral coding of social engagement is underway; final results will include additional data.
Movement outcomes (Mean Velocity, Moving Duration Percent) correlated with child characteristics for both PCI tasks (Tables 1 and 2). Time spent in activity regions (Center and Table) was significantly associated with child characteristics during Task 1 (Table 1, Duration Percent: Center, Table). Measures of proximity between parent and child were associated with child characteristics in Task 2 (Table 2, Between Subjects Mean Distance; In-Proximity to Parent Duration %). The size of correlations was small, but statistically significant.
Tracking and clinical correlations across PCI tasks reveal a pattern that more severely affected children (poorer cognitive, social-communication, or adaptive skills) may exhibit greater movement during both tasks, spend less time playing in a central activity region in Task 1, and spend more time in closer proximity to their parent when the parent is free to move. Measures of time in activity regions may be indicators of child functioning during independent play, whereas parent proximity measures may be appropriate when the parent and child are free to interact and move. Findings support the need for future research to study the use of automated tracking variables from PCIs as measures of autism and functional severity, and to explore this method as a scalable, quantitative, and reliable measure for diagnosis and intervention.

Authors