Task-relevant chunking in sequence learning.

Abstract

In the present study, we investigated possible influences on the unitization of responses. In Experiments 1, 2, 3, and 6, we found that when the same small fragment (i.e., a few consecutive responses in a sequence) was presented as part of two larger sequences, participants responded to it faster when it was part of the sequence that was presented more often. This indicates that chunking can be driven by task-relevant considerations, as opposed to co-occurrence. The results are discussed in the context of chunking theories and the relevant motor learning literature.

Goal Commitments and the content of thoughts and dreams: basic principles

“…EEG evidence with event-related potentials indicates that infants as young as nine months react with N400 deflections (negative deflections after 400 ms post-stimulus) when sequences they observe end unexpectedly (Reid et al, 2009). Perceptual decoupling may, accordingly, be part of a more extensive process that protects ongoing behavior (see also Klinger, 1971, 2011, in regard to a meaning-complex theory of response organization; also behavioral chunking, e.g., Perlman et al, 2010). …”

Chunking improves symbolic sequence processing and relies on working memory gating mechanisms

“…This finding is consequential because, although evidence for chunking has been reported repeatedly during sequence execution (Gobet et al 2001;Sakai et al 2003;Boyd et al 2009;Perlman et al 2010;Seidler et al 2012;Lungu et al 2014), previous attempts to determine its causal role in sequence processing have failed. For instance, a recent study investigating the neural correlates of chunking was unsuccessful in establishing a relationship between learning rate and chunking, as quantified by computing the network modularity (Wymbs et al 2012).…”

The notion of contextual locking: Previously learnt items are not accessible as such when appearing in a less common context

“…Namely, if participants were behaving in this way, performance would be at chance; i.e., error rates would be high, (as it turns out such an option is also inconsistent with results). A third possibility is that participants encode both spatial orientation and item identity, we would expect shorter response times to non-overlapping items in the more frequent array versus the less frequent array a finding typically observed in such paradigms (Perlman, et al, 2010;and observed in all current experiments).…”

“…In sequence learning (e.g., Cleeremans & McClelland, 1991), for example, the notion of chunking is central and refers to a situation where adjacent stimuli in a fixed sequence (e.g., A and B) may eventually be chunked (i.e., eventually the response to A may automatically generate the B response). Perlman, et al (2010) showed that, as chunking knowledge develops, participants respond in a manner suggesting that the smaller units of a chunked sequence disappear or decay, as larger units of representation are developed (see also e.g., Perruchet, Vinter, Pacteau & Gallego, 2002;Pothos & Wolff, 2006). In essence, while chunking is conceived as a hierarchical process by which items are bound to each other to form sub-units, which eventually will be bound to form a unitized presentation comprised of the entire set, the notion of contextual locking is a lateral form of unitization, whereby different items are unitized by being bound to a common context.…”