Predicting the unknown: Novelty processing depends on expectations

Abstract

Fulfilled predictions lead to neural suppression akin to repetition suppression, but it is currently unclear if such effects generalize to broader stimulus categories in the absence of exact expectations. In particular, does expecting novelty alter the way novel stimuli are processed? In the present study, the effects of expectations on novelty processing were investigated using event-related potentials, while controlling for the effect of repetition. Sequences of five stimuli were presented in a continuous way, such that the last stimulus of a 5-stimulus sequence was followed by the first stimulus of a new 5-stimulus sequence without interruption. The 5-stimulus sequence was predictable: the first three stimuli were preceded by a cue indicating that the next stimulus was likely to be a standard stimulus, and the last two by a cue indicating that the next stimulus was likely to be novel. On some trials a cue typically predicting a standard was in fact followed by an unexpected novel stimulus. This design allowed to investigate the independent effects of (violated) expectations and repetition on novelty processing. The initial detection of expected novels was enhanced compared to unexpected novels, as indexed by a larger anterior N2. In contrast, the orienting response, as reflected by a novelty P3, was reduced for expected compared to unexpected novels. Although the novel stimuli were never repeated themselves, they could be presented after one another in the sequence. Such a category repetition affected the processing of novelty, as evidenced by an enhanced anterior N2, and a reduced novelty P3 for novels preceded by other novels. Taken together, the current study shows that novelty processing is influenced by expectations.