Journal of Cognitive Neuroscience
Human visual system automatically encodes sequential regularities of discrete events
Journal of Cognitive Neuroscience
| Hi-index | 0.00 |
In this experiment, we examined the extent to which error-driven learning may operate under implicit learning conditions. We compared error monitoring in a sequence learning task in which stimuli consisted of regular, irregular, or random sequences. Subjects were either informed (explicit condition) or not informed (implicit condition) about the existence of the sequence. For both conditions, reaction times were faster to stimuli from regular sequences than from random sequences, thus supporting the view that sequence learning occurs irrespective of learning condition. Response-locked event-related potentials (ERPs) showed a pronounced ERN/Ne, thereby signaling the detection of committed errors. Deviant stimuli from irregular sequences elicited an N2b component that developed in the course of the experiment, albeit faster for explicit than implicit learners. This observation supports the view that deviant events acquire the status of perceived errors during explicit and implicit learning, and thus, an N2b is generated resembling the ERN/Ne to committed errors. While performing the task, expectations about upcoming events are generated, compared to the actual events, and evaluated on the dimension better or worse than expected. The accuracy of this process improves with learning, as shown by a gradual increase in N2b amplitude as a function of learning. Additionally, a P3b, which is thought to mirror conscious processing of deviant stimuli and is related to updating of working memory representations, was found for explicit learners only.