Under the hood of statistical learning: A statistical MMN reflects the magnitude of transitional probabilities in auditory sequences

Abstract

Within the framework of statistical learning, many behavioural studies investigated the processing of unpredicted events. However, surprisingly few neurophysiological studies are available on this topic, and no statistical learning experiment has investigated electroencephalographic (EEG) correlates of processing events with different transition probabilities. We carried out an EEG study with a novel variant of the established statistical learning paradigm. Timbres were presented in isochronous sequences of triplets. The first two sounds of all triplets were equiprobable, while the third sound occurred with either low (10%), intermediate (30%), or high (60%) probability. Thus, the occurrence probability of the third item of each triplet (given the first two items) was varied. Compared to high-probability triplet endings, endings with low and intermediate probability elicited an early anterior negativity that had an onset around 100 ms and was maximal at around 180 ms. This effect was larger for events with low than for events with intermediate probability. Our results reveal that, when predictions are based on statistical learning, events that do not match a prediction evoke an early anterior negativity, with the amplitude of this mismatch response being inversely related to the probability of such events. Thus, we report a statistical mismatch negativity (sMMN) that reflects statistical learning of transitional probability distributions that go beyond auditory sensory memory capabilities.

Pitch-class distribution modulates the statistical learning of atonal chord sequences

“…The present study did not detect a statistical learning effect on the N1m and P2m responses, although previous research detected an effect on the N1m and P2m as well as P1m (Daikoku et al, 2014(Daikoku et al, , 2015Furl et al, 2011;Koelsch et al, 2016;Paraskevopoulos et al, 2012). From the viewpoint of signal processing, infinite averaging of continuous data at every SOA of 0.5 s eliminates signals below 2 Hz in frequency.…”

“…The reduction was reflected not only in the N1 (N1m) response (Abla et al, 2008;Daikoku et al, 2014Daikoku et al, , 2015Francois & Schön, 2010;Koelsch et al, 2016;Sanders, Ameral, & Sayles, 2009) but also in the P1m responses (Paraskevopoulos et al, 2012). Thus, the statistical learning of monophonic auditory sequences has been indexed by the evoked responses.…”

“…In previous studies, learners could acquire the statistical regularities that are involved in transitional probabilities and are embedded in speech tone (François & Schön, 2014;Saffran, Aslin, & Newport, 1996;Saffran, Johnson, Aslin, & Newport, 1999;Saffran, Loman, & Robertson, 2001) or chord sequences (Jonaitis & Saffran, 2009;Smith & Melara, 1990), with no specific knowledge of language and music domains: phonemes, syllables, syntax, and pitch classes (Hauser et al, 2002;Jackendoff & Lerdahl, 2006). Statistical learning has been demonstrated in not only behavioural but also neurophysiological results (Abla, Katahira, & Okanoya, 2008;Daikoku, Yatomi, & Yumoto, 2014François & Schön, 2014;Furl et al, 2011;Koelsch, Busch, Jentschke, & Rohrmeier, 2016;Paraskevopoulos, Kuchenbuch, Herholz, & Pantev, 2012;Schön & François, 2011 According to previous research on music perception, pitch perception may partially depend on the domain-specific processing component called pitch class, which is a set of pitches that are separated by one or more octaves (Drobisch, 1855;Fig. 1).…”

“…According to recent studies, the statistical learning of single tone sequences with higher-order structure introduced by a Markov stochastic model or a word segmentation paradigm is reflected in neuromagnetic responses, such as P1m, N1m, and P2m (the magnetic counterpart of the P1, N1, and P2 potentials) (Daikoku et al, 2014(Daikoku et al, , 2015Furl et al, 2011;Koelsch et al, 2016;Paraskevopoulos et al, 2012). In the studies using a word segmentation paradigm in which sequences were comprised of tone words, neural responses to the final tones within the words were significantly reduced compared to the initial tones, due to the contrast between high across-word variability and low within-word variability.…”

Statistical learning of an auditory sequence and reorganization of acquired knowledge: A time course of word segmentation and ordering

“…In addition, these statistical learning effects reflected in neural responses could be observed in both the ERPs and the magnetic counterparts of ERPs. This neural effect could be observed in the responses at approximately 50 ms (i.e., P1/P1m) (Paraskevopoulos et al, 2012;Daikoku et al, 2016), 100 ms (i.e., N1/N1m) (Abla et al, 2008;Furl et al, 2011;Daikoku et al, 2014Daikoku et al, , 2015Koelsch et al, 2016), and 200 ms (i.e., P2/P2m) (Furl et al, 2011) after stimulus onset.…”