A Computational Model of the Headturn Preference Procedure: Design, Challenges, and Insights

Bergmann, C. 1, 2 , ten Bosch, L. 1 , Versteegh, M. 1, 2 & Boves, L. 1

1 Centre for Language Sciences, Radboud University, Nijmegen, The Netherlands
2 International Max Planck Research School for Language Sciences, Nijmegen, The Netherlands

When investigating linguistic abilities in infants, the Headturn Preference Procedure (HPP) is a frequently used method (e.g., Jusczyk & Aslin, 1995; and subsequent studies). In this paradigm infants are usually first familiarised with words and then tested for listening preference of passages containing those words in comparison to unrelated passages. Listening preference is defined as the time an infant spends attending to those passages with his or her head turned towards a blinking light and the speech stimuli. Turning the head away ends the trial prematurely.
The knowledge and abilities attested during those studies have been used to model early linguistic skills and language acquisition. However, the actual cause of infants’ behaviour has been subject to numerous assumptions as there are no means to directly tap into cognitive processes.
To make these assumptions explicit, and more crucially to understand how infants’ behaviour emerges while general learning mechanisms are assumed, we built a computational model of the HPP. This model relies on a single flow of information, gets real speech as input, builds episodic representations and simulates the actual head turns. The model represents familiarised words in its memory in the form of histogram of occurrences of acoustic patterns. In a separate test phase, the model recognises sentences by reconstructing them using these internal representations in the memory and simulates headturns accordingly.
Simulations with the computational HPP model show that the difference in infant behaviour between familiarised and unfamiliar words in passages can be explained by a basic cognitive recognition mechanism.
Furthermore, the model simulates the procedure on the level of single test items. This means we can derive predictions on how internal factors (e.g., attention) and external influences (e.g., experimental design, number of presented items) interact and affect the overt behaviour of individual children during the HPP.