A number-line task with a Bayesian active learning algorithm provides insights into the development of non-symbolic...

Area of Science:

• Cognitive Psychology
• Developmental Psychology
• Numerical Cognition

Background:

• The number-line task is a key method for understanding numerical representations.
• A central debate concerns whether symbolic (e.g., numerals) and non-symbolic (e.g., dots) number processing share underlying mechanisms.
• Existing research primarily focuses on symbolic number processing, leaving the non-symbolic domain less explored.

Purpose of the Study:

• To investigate if established findings from symbolic number-line estimation generalize to non-symbolic number-line estimation.
• To determine if numerical representations for symbolic and non-symbolic quantities develop through common pathways.
• To explore numerical cognition across a broad age spectrum (5-73 years).

Main Methods:

• Utilized a number-line task to assess participants' estimation of numerical quantities.
• Applied a novel Bayesian active learning algorithm, Gaussian process active learning (GPAL), for efficient experimental design.
• Collected data from a diverse sample of 238 participants aged 5 to 73 years.

Main Results:

• Non-symbolic number estimation demonstrated three characteristic features previously observed in symbolic number estimation.
• These findings suggest a shared underlying representational system for both symbolic and non-symbolic numbers.
• The observed patterns held across a wide range of ages, indicating developmental continuity.

Conclusions:

• Symbolic and non-symbolic number representations appear to rely on common cognitive processes.
• The developmental trajectory for understanding numbers is likely shared, regardless of the format (symbolic vs. non-symbolic).
• This research provides strong evidence for a unified system of numerical representation.