Tracking the implicit acquisition of nonadjacent transitional probabilities by ERPs

Abstract

The implicit acquisition of complex probabilistic regularities has been found to be crucial in numerous automatized cognitive abilities, including language processing and associative learning. However, it has not been completely elucidated how the implicit extraction of second-order nonadjacent transitional probabilities is reflected by neurophysiological processes. Therefore, this study investigated the sensitivity of event-related brain potentials (ERPs) to these probabilistic regularities embedded in a sequence of visual stimuli without providing explicit information on the structure of the stimulus stream. Healthy young adults (N = 32) performed a four-choice RT task that included a sequential regularity between nonadjacent trials yielding a complex transitional probability structure. ERPs were measured relative to both stimulus and response onset. RTs indicated the rapid acquisition of the sequential regularity and the transitional probabilities. The acquisition process was also tracked by the stimulus-locked and response-locked P3 component: The P3 peak was larger for the sequence than for the random stimuli, while the late P3 was larger for less probable than for more probable short-range relations among the random stimuli. According to the RT and P3 effects, sensitivity to the sequential regularity is assumed to be supported by the initial sensitivity to the transitional probabilities. These results suggest that stimulus–response contingencies on the probabilistic regularities of the ongoing stimulus context are implicitly mapped and constantly revised. Overall, this study (1) highlights the role of predictive processes during implicit memory formation, and (2) delineates a potential to gain further insight into the dynamics of implicit acquisition processes.

Not All Words Are Equally Acquired: Transitional Probabilities and Instructions Affect the Electrophysiological Correlates of Statistical Learning

“…Thus, it is not surprising that recent studies have been using online (neural) and not only post-learning offline measures of SL (e.g., 2-AFC) to study the processes and mechanisms that underlie the extraction of the statistical regularities embedded in the input and also to shed light on other controversial issues largely unexplored in the SL literature, as the nature of representations that arise from SL tasks (e.g., Batterink et al, 2015a , 2019 ; Batterink and Paller, 2017 ; Batterink et al, 2019 ; Kóbor et al, 2018 , 2019 ; Batterink, 2020 ; Horváth et al, 2020 ; see Batterink et al, 2019 ; and Daltrozzo and Conway, 2014 for recent reviews). This strongly contrasts with what has been investigated in the related implicit learning field (see Perruchet and Pacton, 2006 , and also Christiansen, 2019 ) where a significant amount of research has been devoted to examining the type of representations and the (implicit vs. explicit) nature of the knowledge emerging from tasks such as the artificial grammar learning (AGL) task (Reber, 1967 ) or the serial reaction time (SRT) task (Nissen and Bullemer, 1987 ) or any version of it, either using subjective confidence scales [see for instance Jiménez et al, 2020 or Soares (under review) for recent examples] or dissociating these two types of knowledge through the manipulation of the instructions.…”

“…In the SL literature, the few studies conducted so far on the type of representations emerging from sequential learning tasks either in the auditory or visuomotor modality suggest that both implicit and explicit representations might arise from SL tasks (e.g., Turk-Browne et al, 2009 , 2010 ; Franco et al, 2011 ; Bertels et al, 2012 , 2015 ; Batterink et al, 2015a , b ; Kóbor et al, 2018 , 2019 ; Horváth et al, 2020 ). For instance, Batterink et al ( 2015a ) recorded behavioral (RTs/accuracy) and ERP responses while participants performed two post-learning tasks: a speeded target detection task, aimed to assess SL indirectly as it asks participants to detect as fast and accurately as possible a specific syllable within a continuous speech stream, and the abovementioned 2-AFC task combined with a remember/know procedure to assess SL directly under either implicit or explicit conditions.…”

Intention to learn differentially affects subprocesses of procedural learning and consolidation: Evidence from a probabilistic sequence learning task

“…Previous studies have also reported such result, even using self-paced timing (Nemeth, Janacsek, & Fiser, 2013;Song et al, 2008); the underlying process, however, is still unclear (cf. Kóbor et al, 2019). Better performance on highfrequency random trials may suggest that statistical learning is primary compared to sequence learning.…”

“…Consolidation refers to the changes in the acquired knowledge during the post-learning offline periods (Krakauer & Shadmehr, 2006;Meier & Cock, 2014;Robertson, 2009;Robertson, Pascual-Leone, & Miall, 2004). Typically, incidentally acquired (statistical 5 or sequence) knowledge is retained during the offline period (Hallgató, Győri-Dani, Pekár, Janacsek, & Nemeth, 2013;Janacsek & Nemeth, 2012;Kim, Seitz, Feenstra, & Shams, 2009;Kóbor et al, 2019;Rickard et al, 2008;Song, Howard, & Howard, 2007b). In contrast, previous studies on intentionally acquired sequence knowledge led to mixed findings: while in some cases successful consolidation of the acquired knowledge means lower degree of forgetting (Mednick, Cai, Kanady, & Drummond, 2008), in other cases retention or even improvement in performance can occur after the delay period (Pan & Rickard, 2015;Robertson, Pascual-Leone, & Miall, 2004).…”

Neurophysiological coding of statistical and deterministic rule information

“…While statistical learning occurs in a rapid manner and reaches its plateau quickly, learning of sequential rules is characterized by a gradual change (Kóbor et al, 2018;Simor et al, 2019). Despite the growing interest in these simultaneous learning mechanisms (Conway, 2020), the neurophysiological processes behind them remain largely unknown (Kóbor et al, 2019(Kóbor et al, , 2018 and it is an open question how these parallel learning mechanisms are coded at the neurophysiological level. However, solving this puzzle represents a methodological challenge.…”