Stop Frame Coder (SFC): A Tool for Detailed and Reliable Behavior Quantification

Background: Several lines of research suggest that there may be subtle differences in motor and social behavior between children later diagnosed with Autism Spectrum Disorders (ASDs) and unaffected peers. In particular, subtle motor delays seem evident during infancy in high-risk infants (Bhat, Galloway, & Landa, 2012; Flanagan, Landa, Bhat, & Bauman, 2012; Ozonoff, Macari, et al., 2008). However, experimental designs focusing on behavioral observations seem more sensitive than standardized assessments in detecting early signs of ASD in infancy (Ozonoff, Young, et al., 2008; Provost, Lopez, & Heimerl, 2007). This highlights the value of detailed behavior quantification and the need for fast and reliable means to quantify behavior in studies on the earliest signs of ASD. When observing behaviors, information regarding social focus, manual engagement, and emotional status are available simultaneously and need to be recorded and scored at the same time. This is not possible using real-time observation approaches, while video-based coding is a long and error-prone process. In this presentation, we introduce the freely available Stop Frame Coder (SFC) software that addressing these issues by using multiple passes combined with frame-by-frame coding.

Objectives: Provide a fast and reliable tool for behavior quantification in the social, emotional, and motor domains to facilitate research on early ASD markers.

Methods: SFC allows observers to score videos frame-by-frame (10 frames/second). Videos are coded in 4 passes to assess visual attention, manual exploration, social behaviors, and emotional state. At each pass, observers focus on a single behavior (e.g., location of eye gaze) to increase accuracy and reliability. Following coding, behaviors coded in different passes can be combined into complex behavior units (e.g., looking at and grasping an object, smiling and looking at another person). A summary and time course output is provided to allow for in-depth analyses of moment-to-moment changes in behavior. 

Results: SFC is currently being used by eight different research labs to study infant behavior in diverse domains such as cognition, perception, motor, or social development. Published studies using the coding software have investigated looking preferences in typically developing (TD) infants (DeNicola, Holt, Lambert, & Cashon, 2013), infants’ processing and detection of changes in visual displays (Cordes & Brannon, 2011; M. E. Libertus & Brannon, 2010; M. E. Libertus, Starr, & Brannon, 2013; Starr, M. E. Libertus, & Brannon, 2013a, 2013b), and infants’ manual exploration strategies (K. Libertus & Needham, 2010, 2011). More recently, SFC has been used to quantify motor behavior in infants at high genetic risk for ASD or children subsequently diagnosed with ASD. Coding samples and results from select studies will be presented along with a live demonstration of the software.

Conclusions: The SFC software allows extraction of detailed motor, social, and emotional behavior in infants and young children. A number of studies have demonstrated its validity and value in research. Use of SFC in studies with infants and children at risk for ASD may shed light on subtle motor and social delays that emerge in early infancy but cannot be detected using conventional methods or standardized screening measures.