Temporal contiguity determines overshadowing and potentiation of human Action-Outcome performance

Abstract

Three experiments (n = 81, n = 81, n = 82, respectively) explored how temporal contiguity influences Action-Outcome learning, assessing whether an intervening signal competed, facilitated, or had no effect on performance and causal attribution in undergraduate participants. Across experiments, we observed competition and facilitation as a function of the temporal contiguity between Action and Outcome. When there was a strong temporal relationship between Action and Outcome, the signal competed with the action, hindering instrumental performance but not causal attribution (Experiments 1 and 3). However, with weak temporal contiguity, the same signal facilitated both instrumental performance and causal attribution (Experiments 1 and 2). Finally, the physical intensity of the signal determined the magnitude of competition. As anticipated by associative learning models, a more salient signal attenuated to a greater extent instrumental performance (Experiment 3). These results are discussed by reference to a recent adaptation of the configural theory of learning.

Keywords: Action-outcome; Cue competition; Overshadowing; Potentiation; Temporal contiguity.

Fig. 1

Snapshots of the different images used in the game. Panel (a) displays the default scenario that participants experienced in the absence of any other events. Panel (b) shows aSnapshot of the delivery of the outcome (0.1-s length). Panel (c) represents the grey sky used as Low Signal in Experiments 1–3 (0.5-s length). Panel (d) characterizes the white sky used as High Signal in Experiment 3 (0.5-s length)

Fig. 2

Design and results of Experiment 1. Panel (a) depicts the schematic timeline of each experimental condition for Experiment 1. D2s refers to the condition in which the outcome was delayed 2 s and D6s refers to 6-s delay. The hand symbolizes the action starting the timeline, the grey square represents the signal, and the explosion the outcome. Panel (b) represents the mean square root of the number of presses; Panel (c) depicts the mean of causal judgments. Error bars are SEM applying the within-subjects correction suggested by O’Brien and Cousineau (2014)

Fig. 3

Design and results of Experiment 2. Panel (a) depicts the schematic timeline of each experimental condition for Experiment 2. D6s means that the outcome was delayed for 6 s. The hand symbolizes the action starting the timeline, the grey square represents the signal and the explosion the outcome. Panel (b) represents the mean square root of the number of presses. Panel (c) depicts the mean of causal judgments. Error bars are SEM applying the within-subjects correction suggested by O’Brien and Cousineau (2014)

Fig. 4

Design and results of Experiment 3. Panel (a) depicts the schematic timeline of each experimental condition for Experiment 3. D2s means that the outcome was delayed for 2 s. The hand symbolizes the action, the light-grey square represents the Low-Signal condition (similar to Experiments 1 and 2), and the dark-grey represents the High-Signal condition. Panel (b) represents the mean square root of the number of presses. Panel (c) depicts the mean of causal judgments. Error bars are SEM applying the within-subjects correction suggested by O’Brien and Cousineau (2014)