flow and metabolism are sluggish signals, which grow and decay over many seconds in
response to the physiological demands of the underlying neural tissue. These techniques
thus do not operate at the rapid time-scale of melody perception.
Each of the approaches outlined above continues to provide valuable information about
melody and the brain, yet none of them fulfils the desideratum mentioned at the beginning
of this section. That is, none of them follows brain activity as perception unfolds in time
over the course of individual melodies. The ERP method, for example,‘zooms in’to exam-
ine the temporal details of neural responses to individual tones, while PET and fMRI ‘zoom
out’to examine the overall response of various brain regions to entire tone sequences.
Melodic processing in the brain, however, is more than the sum of responses to single
tones, and more than an average response to an entire sequence. It is a dynamic process of
building mental relations between tones during the course of individual melodies. For this
reason, techniques are needed that measure patterns of neural activity as perception
unfolds within individual sequences. The study described below is one attempt in this
direction.
A new approach for the neural study of melody: the dynamic aSSR
(auditory steady-state response) method
Background Patel & Balaban^44 set out to examine the temporal evolution of stimulus-
related brain activity during perception of tone sequences. To accomplish this, we used a
brain signal known as the aSSR.^53 The aSSR is a sinusoidal neural oscillation produced in
primary auditory cortex in response to an acoustic stimulus with constant amplitude modu-
lation (AM). aSSR frequency equals the acoustic AM rate, and it is strongest when the AM
is in the 40 Hz range.54,55Figure 21.6A shows an acoustic signal with constant carrier fre-
quency (400 Hz) and constant AM (40 Hz), and Figure 21.6B shows the power spectrum
of a brain signal recorded with an MEG (magnetoencephalography, see later) sensor over
auditory cortex while an individual listened to this tone for 1 min. A clear energy peak is
visible at 40 Hz. When the same tone is presented without AM, no peak at 40 Hz is visible
in the power spectrum (Figure 21.6C). Thus the aSSR is a frequency-specific, stimulus-
related brain response with a high signal-to-noise ratio. Unlike an evoked potential, the
aSSR represents continuous cortical activity, that is, the oscillation is present as long as the
stimulus is on.
How can this seemingly esoteric brain signal be of use to the cognitive neuroscience of
melody? The answer to this question has two parts. The first is an empirical observation,
namely that the frequency specificity of the aSSR is preserved even when the carrier fre-
quency of a tone changes in a step-wise fashion over the course of two octaves, as long as
the AM rate stays constant.^44 Thus one can impose a constant AM on a melody and extract
stimulus-related neural activity from the brain by measuring activity at the AM rate.
Second, while aSSR frequency does not vary in response to changes in carrier frequency, the
amplitude and phase (relative to the stimulus AM) of this oscillation dochange dynamic-
ally over time during the perception of musical sequences.^56 It is these latter variables (the
‘dynamic’aspect of the aSSR) which are of primary interest. In particular, the key question
is whether aSSR amplitude and phase are sensitive to cognitively relevant processes during
melody perception, such as expectancy. If aSSR dynamics are sensitive to cognitive
334