The Effect of Music Experience on Auditory Sequential Learning: An ERP Study

The Effect of Music Experience on Auditory Sequential Learning: An ERP Study Samantha N. Emerson (semerson2@student.gsu.edu) Department of Psychology, P.O. Box 5010, Atlanta, GA 30302 USA Jerome Daltrozzo (jdaltrozzo@gsu.edu) Department of Psychology, P.O. Box 5010, Atlanta, GA 30302 USA Christopher M. Conway (cconway@gsu.edu) Department of Psychology, P.O. Box 5010, Atlanta, GA 30302 USA Abstract The existence of an advantage in sequential learning for musicians over nonmusicans is highly debated. The current study used an auditory sequential learning task to investigate the neurophysiological correlates of sequential learning in adults with either high or low music aptitudes. While behavioral results alone revealed no difference between the reaction times of the two groups, event-related potential data showed that higher music aptitude was associated with decreased amplitudes of the P300 and Contingent Negative Variation effect between two conditions with different transitional probabilities relative to a target stimulus. These data suggest that increased music training and skill leads to more efficient processing of (i.e., reduced attentional demands for) auditory sequential patterns. Keywords: Sequential learning; potentials (ERP); P300; CNV music; event-related Introduction Sequential learning (SL) is the ability to either implicitly or explicitly extract statistical probabilities from series of discrete elements and to form expectations based on that probabilistic information (Conway & Christiansen, 2001; Conway & Pisoni, 2008). This skill is particularly important to the development of language and has been implicated in the acquisition of word boundaries (Saffran, Aslin, & Newport, 1996), syntax (Ullman, 2004), and word order (Conway, Bauernschmidt, Huang, & Pisoni, 2010). The role of experience in shaping SL mechanisms is still relatively underspecified. Conway, Pisioni, Anaya, Karpicke, and Henning (2011) found that an early period of sound deprivation (i.e., in children with cochlear implants) led to deficits in SL abilities. On the other hand, it is possible that increased experience or skill with sound—for example music—might lead to an advantage in SL. Sequential Learning in Musicians The existence of an advantage in SL for musicians over nonmusicians is highly debated. For example, Rohrmeier, Rebuschat, and Cross (2011) found that musicians did not show an advantage over nonmusicians in their familiarity with musical sequences produced from an artificial grammar. Similarly, Bigand’s (2003) review of the literature strengthens this view by demonstrating similarities in performance between musicians and nonmusicians in the processing of melodic and harmonic structures, in the processing of large-scale structures, and in implicit learning for musical structures. Bigand argues that nonmusicians’ every day exposure to music makes them “expert listeners” and therefore are as competent as musicians with respect to the implicit understanding of the complex structures—or grammars—that underlie music. This behavioral (as well as some neural) evidence appears to suggest that musical expertise does not improve SL abilities in the auditory domain. However, recent neural studies support the view of an advantage in SL with increased musical expertise. For example, two Magnetoencephalographic (MEG) studies (Herholz, Boh, & Pantev, 2011; Paraskevopoulos, Kuchenbuch, Herholz, & Panteva, 2012) examined exposure to deviant sequences of tones embedded within more standard sequences. Both studies found that musicians and nonmusicians responded similarly to the deviant sequences. However, Herhloz et al. (2011) found that only musicians exhibited an increased mismatch negativity (MMN; Naatanen & Alho, 1995) within 10 minutes of exposure, and Paraskevopoulos et al. (2012) found a significantly larger amplitude of the P50 in comparison to nonmusicians. These results demonstrate an effect of musical expertise on pre-attentive auditory abilities and short-term auditory learning of statistical regularities. An electrophysiological study examined musicians and nonmusicians’ learning of statistical regularities in a sung “language” in which each syllable was associated with a particular note (Francois & Schon, 2011). Behavioral analysis revealed that both musicians and nonmusicians were able to segment the syllables and notes based on the musical structure. Electroencephalographic (EEG) analysis, however, revealed that, compared to nonmusicians, musicians exhibited a larger N1 component and a larger negativity in the 750 to 850 ms latency band in response to untrained linguistic segments. For untrained musical segments, compared with nonmusicians, musicians exhibited larger N1 and P2 components which were larger over the left hemisphere than the right, a negativity in the 350 to 500 ms latency band which was largest over the central and left frontal regions (i.e., an N400-like effect), and a negativity in the 700 to 800 ms latency band which