Functional differences between statistical learning with and without explicit training

Abstract

Humans are capable of rapidly extracting regularities from environmental input, a process known as statistical learning. This type of learning typically occurs automatically, through passive exposure to environmental input. The presumed function of statistical learning is to optimize processing, allowing the brain to more accurately predict and prepare for incoming input. In this study, we ask whether the function of statistical learning may be enhanced through supplementary explicit training, in which underlying regularities are explicitly taught rather than simply abstracted through exposure. Learners were randomly assigned either to an explicit group or an implicit group. All learners were exposed to a continuous stream of repeating nonsense words. Prior to this implicit training, learners in the explicit group received supplementary explicit training on the nonsense words. Statistical learning was assessed through a speeded reaction-time (RT) task, which measured the extent to which learners used acquired statistical knowledge to optimize online processing. Both RTs and brain potentials revealed significant differences in online processing as a function of training condition. RTs showed a crossover interaction; responses in the explicit group were faster to predictable targets and marginally slower to less predictable targets relative to responses in the implicit group. P300 potentials to predictable targets were larger in the explicit group than in the implicit group, suggesting greater recruitment of controlled, effortful processes. Taken together, these results suggest that information abstracted through passive exposure during statistical learning may be processed more automatically and with less effort than information that is acquired explicitly.

Online neural monitoring of statistical learning

“…Stimulus timing parameters were identical to those in the Exposure task. Based on our previous findings (Batterink et al, 2015a, 2015b), we expect graded reaction time (RT) effects as a function of syllable position. Syllable targets that occur in the final position of a word should elicit faster RTs, indexing facilitation due to statistical learning.…”

“…Finally, as in our previous statistical learning studies (Batterink et al, 2015a, 2015b), participants completed a speeded target detection task as a final measure of statistical learning. On each trial, participants were presented with a speech stream containing the four words from the structured language repeated four times each, which was shorter but otherwise similar to the speech stream presented during the Exposure task.…”

“…Knowledge of words in the structured stream was assessed directly using a rating task, in which participants provided familiarity ratings to words from the stream and foil items made up of recombined syllables. Statistical learning was also assessed indirectly using a target detection task, an online, speeded, performance-based measure (Batterink et al, 2015a, 2015b, Franco et al, 2015). This task requires participants to respond to target syllables occurring in a continuous syllable stream and assesses the extent to which learners use their acquired statistical knowledge to optimize online processing.…”

“…This task requires participants to respond to target syllables occurring in a continuous syllable stream and assesses the extent to which learners use their acquired statistical knowledge to optimize online processing. We expected to observe faster reaction times (RTs) to relatively predictable targets (i.e., those occurring in later syllable positions compared to the first position), indexing more efficient processing (e.g., Turk-Browne et al 2005; Kim et al 2009; Batterink et al 2015a, 2015b; Franco et al 2015). Whereas the rating task is presumably most sensitive to explicit memory, the target detection task has the potential to capture knowledge above and beyond what is reflected by explicit memory and may partially reflect contributions from implicit memory (Batterink et al, 2015a).…”

Rapid Statistical Learning Supporting Word Extraction From Continuous Speech

“…However, in contrast to most other studies, we used an online, reaction-time (RT) based measure of statistical learning, which requires participants to respond to target syllables occurring in a continuous syllable stream made up of repeating trisyllabic words. This target detection task has been previously shown to be sensitive to statistical learning, as reflected by faster RTs to predictable as compared to word-initial, non-predictable syllables (Batterink, Reber, Neville, & Paller, 2015; Batterink, Reber & Paller, 2015; Franco, Eberlen, Destrebecqz, Cleeremans, & Bertels, 2015). To evaluate whether statistical learning occurs within a very short timescale, each syllable stream in the present study was comprised of a novel set of repeating nonsense words, such that statistical learning began from square one for each stream.…”