How people learn about causal influence when there are many possible causes: A model based on informative transitions

Area of Science:

• Cognitive Psychology
• Machine Learning
• Causal Inference

Background:

• Human causal learning is complex, especially with multiple potential causes for a single effect.
• Existing theories often focus on cue-effect co-occurrence, potentially overlooking temporal dynamics.

Purpose of the Study:

• To investigate how temporal changes (transitions) in cues and effects influence human causal learning.
• To test the hypothesis that 'one-change transitions' are more informative for learning than simultaneous changes.

Main Methods:

• Four experiments presented participants with scenarios involving multiple cues and a single effect.
• Manipulated the number of cues changing simultaneously between trials and within learning sequences.
• Assessed participants' ability to identify cue-effect relationships and causal strength.

Main Results:

• Causal learning was significantly better when only one cue changed per trial compared to multiple simultaneous changes.
• Learning was impaired when trial presentation minimized 'one-change transitions' (e.g., grouping by effect or cue state).
• A modified Rescorla-Wagner model incorporating 'Informative Transitions' was developed.

Conclusions:

• Temporal transition information, particularly 'one-change transitions,' plays a crucial role in human causal learning.
• The way information is presented significantly affects the ability to infer cause-effect relationships.
• The 'Informative Transitions' model offers a more nuanced account of learning with multiple cues.