When I began working with infants as an undergraduate research assistant, I quickly realized the value and importance of developing sound methods to characterize infant development. Below are some of the tools I have evaulated efficacy of:

1. Automated eye blink detection algorithms:

   Spontaneous eye blink rate (EBR) is an indirect measure of dopamine, one of the chemicals in our brains that is related to reward processing. Although there is increasing interest and use of EBR data in developmental research, there are no standardized methods to do this. With my graduate advisor, Dr. Julie Markant, and our team of collaborators, Drs. Leigh Bacher (Professor, SUNY Oswego), Jed Elison (Associate Professor, University of Minnesota), and Robin Sifre (Data Scientist at EarliTec Diagnostics), I compared two blink detection methods: hand coding and an eye tracking-based automated algorithm. Results of this investigation revealed vast inconsistencies across methods with higher accuracy among hand coding methods. This work is currently being prepared for publication (Hunter et al., in prep).

2. Online methods to evaluate attention capture by faces:

   In response to the global COVID-19 pandemic, I pivoted to collect data online. I first evaluated the effectiveness of online methods by developing an attention capture task based on a design of a previously published study using traditional in-person methods. This successful replication suggested we can meaningfully measure children’s attention online (Adab, Hunter, & Markant, 2021; Hunter & Markant, 2022). I was then able to modify this task to confidently collect all data for my dissertation online.

3. Image saliency models:

   As an infant attention researcher, I am interested in infants’ increasing ability to look towards meaningful parts of their environments, rather than the bright and shiny physically salient parts. One way to measure attention to physical salience is by comparing infant fixation data to image saliency models, such as Graph-Based Visual Saliency. However, these models were designed with the mature visual system in mind. My work has revealed less accurate model performance using infant vs. adult fixation data (Hunter et al., VSS 2023). I am currently evaluating why this may be the case and am interested in modifying existing models to better approximate infant saliency processing.