It is well-documented that feature integration across perception and action creates a retrievable episodic representation, known as a stimulus-response episode or an event file. Previous studies have demonstrated that a task-irrelevant stimulus, which functions as contextual information, can be integrated in various ways. In some cases, the context modulated the binding between a stimulus and a response, resulting in a configural binding structure. In other cases, the context was found to be directly bound with the response in a binary fashion. The current study examined the integration of context within a stimulus-response (S-R) episode, with a focus on the role of inter-trial variability. Specifically, the context variability was manipulated across five experimental groups, ranging from the minimum to the maximum level. The minimum-variability group maintained a consistent pattern of two context tones per block, while the maximum-variability group used a uniformly random order of eight different context tones. Intermediate groups progressively employed greater degrees of variability in the presentation of contextual stimuli. Results showed that the integration of context changed as a function of its variability level: The contextual stimuli with minimal to low level of variability did not exhibit a pattern of integration, while those with moderate to high variability were involved in a configural binding with another stimulus and the response. Only when the context exhibited maximal variability did it become directly bound with the response in a binary fashion. The current findings extend previous assumptions about saliency thresholds for stimulus integration into the realm of inter-trial variability and underscore the role of stimulus uncertainty in shaping context integration.

In the zipped file, there are: 1. datasets for each experimental group 2. data and model for MPT model analysis 3. codebook