Information Technology Reference
In-Depth Information
5
Machine Learning to Identify Neural
Correlates of Music and Emotions
Ian Daly, Etienne B. Roesch, James Weaver and Slawomir J. Nasuto
Abstract
While music is widely understood to induce an emotional response in the
listener, the exact nature of that response and its neural correlates are not yet
fully explored. Furthermore, the large number of features which may be
extracted from, and used to describe, neurological data, music stimuli, and
emotional responses, means that the relationships between these datasets
produced during music listening tasks or the operation of a brain
computer
music interface (BCMI) are likely to be complex and multidimensional. As such,
they may not be apparent from simple visual inspection of the data alone.
Machine learning, which is a
-
field of computer science that aims at extracting
information from data, provides an attractive framework for uncovering stable
relationships between datasets and has been suggested as a tool by which neural
correlates of music and emotion may be revealed. In this chapter, we provide an
introduction to the use of machine learning methods for identifying neural
correlates of musical perception and emotion. We then provide examples of
machine learning methods used to study the complex relationships between
neurological activity, musical stimuli, and/or emotional responses.
5.1
Introduction
It is widely understood that music is able to induce a wide range of emotions in the
listener. What is not so well understood is the speci
c neurological mechanism by
which this process takes place.
