Marcus T. Pearce

Unsupervised statistical learning underpins computational, behavioural, and neural manifestations of musical expectation

The ability to anticipate forthcoming events has clear evolutionary advantages, and predictive successes or failures often entail significant psychological and physiological consequences. In music perception, the confirmation and violation of expectations are critical to the communication of emotion and aesthetic effects of a composition. Neuroscientific research on musical expectations has focused on harmony. Although harmony is important in Western tonal styles, other musical traditions, emphasizing pitch and melody, have been rather neglected. In this study, we investigated melodic pitch expectations elicited by ecologically valid musical stimuli by drawing together computational, behavioural, and electrophysiological evidence. Unlike rule-based models, our computational model acquires knowledge through unsupervised statistical learning of sequential structure in music and uses this knowledge to estimate the conditional probability (and information content) of musical notes. Unlike previous behavioural paradigms that interrupt a stimulus, we devised a new paradigm for studying auditory expectation without compromising ecological validity. A strong negative correlation was found between the probability of notes predicted by our model and the subjectively perceived degree of expectedness. Our electrophysiological results showed that low-probability notes, as compared to high-probability notes, elicited a larger (i) negative ERP component at a late time period (400-450 ms), (ii) beta band (14-30 Hz) oscillation over the parietal lobe, and (iii) long-range phase synchronization between multiple brain regions. Altogether, the study demonstrated that statistical learning produces information-theoretic descriptions of musical notes that are proportional to their perceived expectedness and are associated with characteristic patterns of neural activity.

Improved Methods for Statistical Modelling of Monophonic Music

N-Gram based models have been used for a variety of musical tasks including computer-assisted composition, machine improvisation, music information retrieval, stylistic analysis and cognitive modelling. We present an application-independent evaluation of some recent techniques for improving the performance of a subclass of n-gram models on a range of monophonic music data. We have applied these techniques incrementally to eight melodic datasets using cross entropy computed by 10-fold cross-validation on each dataset as our performance metric. The results demonstrate that significant and consistent improvements in performance are afforded by several of the evaluated techniques. We discuss the results in terms of previous research carried out in the field of data compression and with natural language and music corpora and conclude by presenting some important directions for future research.

Auditory Expectation: The Information Dynamics of Music Perception and Cognition

Following in a psychological and musicological tradition beginning with Leonard Meyer, and continuing through David Huron, we present a functional, cognitive account of the phenomenon of expectation in music, grounded in computational, probabilistic modeling. We summarize a range of evidence for this approach, from psychology, neuroscience, musicology, linguistics, and creativity studies, and argue that simulating expectation is an important part of understanding a broad range of human faculties, in music and beyond.

The Role of Expectation and Probabilistic Learning in Auditory Boundary Perception: A Model Comparison

Grouping and boundary perception are central to many aspects of sensory processing in cognition. We present a comparative study of recently published computational models of boundary perception in music. In doing so, we make three contributions. First, we hypothesise a relationship between expectation and grouping in auditory perception, and introduce a novel information-theoretic model of perceptual segmentation to test the hypothesis. Although we apply the model to musical melody, it is applicable in principle to sequential grouping in other areas of cognition. Second, we address a methodological consideration in the analysis of ambiguous stimuli that produce different percepts between individuals. We propose and demonstrate a solution to this problem, based on clustering of participants prior to analysis. Third, we conduct the first comparative analysis of probabilistic-learning and rule-based models of perceptual grouping in music. In spite of having only unsupervised exposure to music, the model performs comparably to rule-based models based on expert musical knowledge, supporting a role for probabilistic learning in perceptual segmentation of music.

Methods for combining statistical models of music

The paper concerns the use of multiple viewpoint representation schemes for prediction with statistical models of monophonic music. We present an experimental comparison of the performance of two techniques for combining predictions within the multiple viewpoint framework. The results demonstrate that a new technique based on a weighted geometric mean outperforms existing techniques. This finding is discussed in terms of previous research in machine learning.

The Copenhagen Neuroaesthetics conference: Prospects and pitfalls for an emerging field

Neuroaesthetics is a young field of research concerned primarily with the neural basis of cognitive and affective processes engaged when an individual takes an aesthetic or artistic approach towards a work of art, a non-artistic object or a natural phenomenon. In September 2009, the Copenhagen Neuroaesthetics Conference brought together leading researchers in the field to present and discuss current advances. We summarize some of the principal themes of the conference, placing neuroaesthetics in a historical context and discussing its scope and relation to other disciplines. We also identify what we believe to be the key outstanding questions, the main pitfalls and challenges faced by the field, and some promising avenues for future research.

Motivations and Methodologies for Automation of the Compositional Process

Our aim in this paper is to clarify the range of motivations that have inspired the development of computer programs for the composition of music. We consider this to be important since different methodologies are appropriate for different motivations and goals. We argue that a widespread failure to specify the motivations and goals involved has lead to a methodological malaise in music related research. A brief consideration of some of the earliest attempts to produce computational systems for the composition of music leads us to identify four activities involving the development of computer programs which compose music each of which is inspired by different practical or theoretical motivations. These activities are algorithmic composition, the design of compositional tools, the computational modelling of musical styles and the computational modelling of music cognition. We consider these four motivations in turn, illustrating the problems that have arisen from failing to distinguish between them. We propose a terminology that clearly differentiates the activities defined by the four motivations and present methodological suggestions for research in each domain. While it is dearly important for researchers to embrace developments in related disciplines, we argue that research in the four domains will continueto stagnate unless the motivations and aims of research projects are clearly stated and appropriate methodologies are adopted for developing and evaluating systems that compose music.

Probabilistic models of expectation violation predict psychophysiological emotional responses to live concert music

We present the results of a study testing the often-theorized role of musical expectations in inducing listeners’ emotions in a live flute concert experiment with 50 participants. Using an audience response system developed for this purpose, we measured subjective experience and peripheral psychophysiological changes continuously. To confirm the existence of the link between expectation and emotion, we used a threefold approach. (1) On the basis of an information-theoretic cognitive model, melodic pitch expectations were predicted by analyzing the musical stimuli used (six pieces of solo flute music). (2) A continuous rating scale was used by half of the audience to measure their experience of unexpectedness toward the music heard. (3) Emotional reactions were measured using a multicomponent approach: subjective feeling (valence and arousal rated continuously by the other half of the audience members), expressive behavior (facial EMG), and peripheral arousal (the latter two being measured in all 50 participants). Results confirmed the predicted relationship between high-information-content musical events, the violation of musical expectations (in corresponding ratings), and emotional reactions (psychologically and physiologically). Musical structures leading to expectation reactions were manifested in emotional reactions at different emotion component levels (increases in subjective arousal and autonomic nervous system activations). These results emphasize the role of musical structure in emotion induction, leading to a further understanding of the frequently experienced emotional effects of music.

Brain responses in humans reveal ideal observer-like sensitivity to complex acoustic patterns

Significance We reveal the temporal dynamics and underlying neural sources of the process by which the brain discovers complex temporal patterns in rapidly unfolding sound sequences. We demonstrate that the auditory system, supported by a network of auditory cortical, hippocampal, and frontal sources, continually scans the environment, efficiently represents complex stimulus statistics, and rapidly (close to the bounds implied by an ideal observer model) responds to emergence of regular patterns, even when these are not behaviorally relevant. Neuronal activity correlated with the predictability of ongoing auditory input, both in terms of deterministic structure and the entropy of random sequences, providing clear neurophysiological evidence of the brains capacity to automatically encode high-order statistics in sensory input. We use behavioral methods, magnetoencephalography, and functional MRI to investigate how human listeners discover temporal patterns and statistical regularities in complex sound sequences. Sensitivity to patterns is fundamental to sensory processing, in particular in the auditory system, because most auditory signals only have meaning as successions over time. Previous evidence suggests that the brain is tuned to the statistics of sensory stimulation. However, the process through which this arises has been elusive. We demonstrate that listeners are remarkably sensitive to the emergence of complex patterns within rapidly evolving sound sequences, performing on par with an ideal observer model. Brain responses reveal online processes of evidence accumulation—dynamic changes in tonic activity precisely correlate with the expected precision or predictability of ongoing auditory input—both in terms of deterministic (first-order) structure and the entropy of random sequences. Source analysis demonstrates an interaction between primary auditory cortex, hippocampus, and inferior frontal gyrus in the process of discovering the regularity within the ongoing sound sequence. The results are consistent with precision based predictive coding accounts of perceptual inference and provide compelling neurophysiological evidence of the brains capacity to encode high-order temporal structure in sensory signals.

Neuroaesthetics: The Cognitive Neuroscience of Aesthetic Experience.

The field of neuroaesthetics has gained in popularity in recent years but also attracted criticism from the perspectives both of the humanities and the sciences. In an effort to consolidate research in the field, we characterize neuroaesthetics as the cognitive neuroscience of aesthetic experience, drawing on long traditions of research in empirical aesthetics on the one hand and cognitive neuroscience on the other. We clarify the aims and scope of the field, identifying relations among neuroscientific investigations of aesthetics, beauty, and art. The approach we advocate takes as its object of study a wide spectrum of aesthetic experiences, resulting from interactions of individuals, sensory stimuli, and context. Drawing on its parent fields, a cognitive neuroscience of aesthetics would investigate the complex cognitive processes and functional networks of brain regions involved in those experiences without placing a value on them. Thus, the cognitive neuroscientific approach may develop in a way that is mutually complementary to approaches in the humanities.The field of neuroaesthetics has gained in popularity in recent years but also attracted criticism from the perspectives both of the humanities and the sciences. In an effort to consolidate research in the field, we characterize neuroaesthetics as the cognitive neuroscience of aesthetic experience, drawing on long traditions of research in empirical aesthetics on the one hand and cognitive neuroscience on the other. We clarify the aims and scope of the field, identifying relations among neuroscientific investigations of aesthetics, beauty, and art. The approach we advocate takes as its object of study a wide spectrum of aesthetic experiences, resulting from interactions of individuals, sensory stimuli, and context. Drawing on its parent fields, a cognitive neuroscience of aesthetics would investigate the complex cognitive processes and functional networks of brain regions involved in those experiences without placing a value on them. Thus, the cognitive neuroscientific approach may develop in a way that is mutually complementary to approaches in the humanities.