When we process stimulus characteristics and our motoric response, we integrate them to have a coherent episodic representation of the event – an event file. In this study, we tested a prediction derived from the event file hypothesis that learning contents can be transferred from one visual feature to another if the features have been bound together in an event file. We did so in an implicit learning paradigm, because there are conflicting theoretical assumptions on whether such integrative processes should be possible in the absence of awareness.
We used a visual search task in which three out of six distractor shapes predicted the target position in a learning phase. In the transfer phase, three out of six distractor colors predicted the target position. In a pre-learning phase, in the experimental condition, each of the six shapes had been bound to a color. In the control condition, the shapes had been associated with meaningless spatial positions.
We found that in both conditions, response times to the target in a predictable position was faster than to an unpredictable position. But only the experimental condition showed a transfer effect when the distractor shapes were replaced with colors. The associations of shapes and colors with target positions remained implicit.
We discuss these findings against the background of consciousness theories and the event file hypothesis.