Summation effects in human learning: evidence from patterning discriminations in goal-tracking

ABSTRACT

Participants in two human goal-tracking experiments were simultaneously trained with negative patterning (NP) and positive patterning (PP) discriminations (A+, B+, AB–, C–, D–, CD+). Both elemental and configural models of associative learning predict a PP advantage, such that NP is solved less readily than PP. However, elemental models like the unique cue approach additionally predict responding in AB– trials to be initially stronger than that in A+ and B+ trials due to summation of associative strength. Both experiments revealed a PP advantage and a strong summation effect in AB– trials in the first half of the experiments, irrespective of whether the same US was used for both discriminations (Experiment 1) or two different USs (Experiment 2). We discuss that the correct predictions of the unique cue approach are based on its assumptions of non-normalized and context-independent stimulus processing rather than elemental processing per se.

KEYWORDS:

• Patterning
• Goal tracking
• Discrimination
• Configural and elemental processing
• Normalization

Acknowledgements

We would like to thank the research assistants of the Marburg lab for help in collecting data and preparing this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 Summation can alternatively be assessed by comparing the maximal response in A+ and B+ trials with the maximal response in the corresponding AB– trials in each block. The pattern of results and thereby the conclusion of the current paper are not affected by this.

2 Alternatively, one can describe it as the model of Pearce (1994) lacking its configural features, as the only difference is that the elements that represent the single CSs do not converge onto configural units but are directly associated with US representation. For a more detailed discussion on the equivalence of elemental and configural models see, for example, Ghirlanda (2015; also Thorwart, 2010).

3 One might be surprised by the small “summation” effect in the right panel of the IEM and the Pearce model. This is due to a constraint that necessarily applies to behavioural experiments and that these simulations mimic. It is not possible to measure the current associative strength, or US prediction, of all stimuli at the same time. Instead, one can only measure the response to one stimulus or stimulus compound in each trial and only after learning occurred in this trial, one can measure the response to another stimulus in the next trial. When A+ and B+ trials are presented, the associations of A and B will be strengthened in these trials. The response measured in these trials therefore underestimates the associative strength that is activated in a following AB– trial. When response in AB– is measured first, some associations will become more inhibitory in this trial and counter responding in the following A+ and B+ trials more strongly than expected based on their measurement.

Additional information

Funding

The research reported in this article was supported by a grant from the German Research Foundation (DFG) [grant number TH 1923/1–1] to Anna Thorwart. The research is partly based on work done while Anna Thorwart was a postdoctoral researcher at the University of Sydney, financed by grants from the German Academic Exchange Service (DAAD) and the Australian Research Council (ARC).