Automated Study Challenges the Existence of a Foundational Statistical-Learning Ability in Newborn Chicks

Area of Science:

• Developmental Psychology
• Cognitive Neuroscience
• Animal Behavior

Background:

• Previous research suggested newborn chicks learn transitional probabilities (TPs) in sequences, indicating early statistical learning.
• This proposed ability in newborn brains has significant implications for understanding early cognitive development.

Purpose of the Study:

• To rigorously test the hypothesis that newborn chicks can encode transitional probabilities (TPs) between sequentially presented shapes.
• To investigate the presence and mechanisms of statistical learning in the developing brain using an automated, bias-free methodology.

Main Methods:

• Employed a preregistered, automated experimental design to eliminate experimenter bias.
• Collected extensive behavioral data from over 250 times more trials per chick compared to previous studies.
• Utilized precise behavioral measurements for both individual-level and group-level analyses, minimizing error.

Main Results:

• Found no evidence that newborn chicks encode transitional probabilities (TPs) between sequentially presented shapes.
• Demonstrated that newborn chicks robustly encode the shapes of individual objects.
• The automated method proved effective in generating reliable and reproducible results.

Conclusions:

• The findings challenge the assertion that transitional probability-based statistical learning is present in newborn brains.
• Newborn chicks exhibit learning capabilities focused on individual object recognition rather than sequential statistical patterns.
• This study highlights the importance of rigorous, automated methods in developmental learning research.