Comparing positive and negative patterning in human learning.

Area of Science:

• Cognitive Psychology
• Animal Behavior
• Machine Learning

Background:

• Human causal learning often shows similar acquisition rates for positive patterning (PP) and negative patterning (NP).
• Nonhuman animal studies typically find PP is learned faster than NP.
• This discrepancy may stem from how learning tasks present outcome significance.

Purpose of the Study:

• To investigate how outcome salience influences the differential learning rates of PP and NP.
• To test predictions derived from associative learning models.

Main Methods:

• A predictive learning paradigm using a mock gambling task was employed.
• Experiments manipulated the salience of outcomes (winning, losing, no change) by varying their frequency.
• Participants completed PP and NP discrimination tasks.

Main Results:

• In Experiment 1, PP was learned faster than NP when one outcome (winning) was less frequent and thus more salient.
• In Experiment 2, PP was learned faster than NP specifically when discriminating between winning and no-change outcomes.
• Learning differences between PP and NP were contingent on outcome salience.

Conclusions:

• The relative difficulty of PP and NP discriminations is influenced by the salience of the outcomes involved.
• Findings support associative learning models where outcome salience plays a critical role in associative strength.
• Differential salience, rather than inherent task structure, may explain learning rate differences in causal learning.