Measuring Auditory Discrimination Thresholds in Preschool Children: An Empirically Based Analysis

ABSTRACT

Auditory discrimination is an important perceptual skill that seems to develop substantially during early childhood and is predictive of key developmental outcomes like language ability. However, the estimation of reliable auditory discrimination thresholds is impeded by non-sensory limitations in young children that impact task performance. Here we used computerized simulations of child-like and adult-like performance, as well as novel behavioral task modifications with 3- and 4-year old preschool children and adults, to investigate key parameters in the successful estimation of auditory discrimination thresholds in preschool-aged children. The results indicated the most suitable adaptive procedure that is not widely used with young children (75% weighted 1-up 1-down procedure), an appropriate number of trials (20) and step size (relatively small). In addition, using a novel manipulation of level of physical engagement, it was found that threshold estimates are more reliable when children were given a task that involved greater physical engagement (i.e., 3D objects, rather than 2D images on a screen). The overall results provide recommendations for designing procedures to estimate thresholds in preschool or developmentally delayed children.

KEYWORDS:

• Auditory discrimination
• adaptive procedures
• frequency discrimination
• preschool auditory development

Acknowledgments

NG and TF co-wrote the paper. NG and TA designed the modeling procedures. NG and TF designed the experimental procedures. All authors contributed independently to different programming and analyses, and all authors discussed the results and reviewed the final manuscript. This project was funded by a Department for Education and Learning studentship grant awarded to TF and TA for the supervision of NG.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1. The percentage sign could represent a wide range of different stimulus parameters that can be incremented and decremented by addition and subtraction, such as decibels or log-frequency. In the model, the “10%” step is the parameter difference that would lead to a 10% improvement in performance, correcting for chance, at the steepest part of the psychometric function. Relating the model to Experiment 3, the model steps are equivalent to $\mathrm{m}\mathrm{l}\mathrm{o}{\mathrm{g}}_2\left(\mathrm{\Delta }\mathrm{F}/{\mathrm{F}}_{\mathrm{r}\mathrm{e}\mathrm{f}}\right)$, where $\mathrm{\Delta }\mathrm{F}$ is the frequency difference between the pure tones in the first and second interval, $F_{ref}$ is an arbitrary reference frequency, and $m$ is an arbitrary constant related to the gradient of the psychometric function. Setting $m=20\mathrm{\%}$ and $F_{ref}=$30.25 Hz would correspond to an initial difference of 1000 Hz, a 50% threshold (corrected for chance) at a frequency difference of 125 Hz, and a psychometric function whose peak gradient is equivalent to a 20% improvement in performance (corrected for chance) for each doubling of frequency.

Additional information

Funding

This work was supported by the Department for Employment and Learning, Northern Ireland.