degree of lag between the time-referenced input signal and the oscillatory brain response
is measured by the phase of the aSSR.
Results I: phase tracking Our first finding was that the phase of the measured brain
signal varied with the pitch of the tone sequence. As pitch increased, phase advanced
(corresponding to a decreasedlag between stimulus and brain response), and vice-versa.
This relationship between aSSR phase and carrier frequency was suggested by early work,^53
and has been confirmed by recent studies,55,59but had not previously been studied in a
dynamic fashion. The underlying mechanism for this phenomenon likely involves the
tonotopic layout of the basilar membrane in the cochlea of the human ear, where higher
frequencies are closer to the oval window and hence are stimulated earlier than lower
frequencies. However, the degree of phase change observed in the aSSR cannot be solely
explained by peripheral neural mechanisms (Patel and Balaban, in preparation).
In our study, variation of phase with stimulus carrier frequency manifested itself in
‘phase tracking’, that is, in a correlation between the shape of the phase-time contour and
the stimulus carrier frequency-time contour (Figure 21.8). As just mentioned, this tracking
may be largely driven by peripheral neural mechanisms, and we make no claim that phase
tracking has a causal relationship to tone sequence processing (indeed, it is only present
because of the AM we have imposed on the tone sequences). Rather, the interesting ques-
tion is whether phase tracking is modulated by perceptual aspects of the stimulus. We
found evidence that this may be the case, in that phase tracking improved as sequences
became more predictable in structure. Examples of phase-time contours (solid lines) overlaid
on their corresponding pitch-time contours (dashed lines) are shown in Figure 21.8A–D,
showing how tracking improves across the stimulus conditions. The best tracking occurred
for musical scales, which have a completely predictable structure.
Each subject showed a number of sensor locations where this ‘phase tracking’of pitch
was observed. Across participants, these locations tended to be in fronto-temporal regions,
with a slight right-hemisphere bias. A similar set of locations was identified when we
looked for sensors where the amplitude of the aSSR was strong. However, we found no evid-
ence that the amplitude time series of the aSSR correlated with the heard pitch contour
(see Patel and Balaban^44 for details).
Results II: phase coherence Knowing that the phase of the brain response reflected stimu-
lus properties (i.e. pitch contour), we then turned to looking at patterns of phase coherence
between different brain regions. Phase coherence does not measure the lag between an
oscillatory signal and brain response but rather the stability of the phase differencebetween
oscillatory activity in different brain areas. Thus phase coherence measures the degree of
temporally correlated activity in distinct brain regions. If two brain areas show greater
phase coherence during certain conditions, this is suggestive of a greater degree of
functional coupling between regions under those conditions.60, l338 lIn studying phase coherence at the aSSR frequency, we are not assuming that the aSSR is itself part of the
causal chain of melodic processing, but rather that its dynamics are influenced by ongoing brain processes which
areinvolved in melody perception. Thus aSSR coherence serves as a proxy for coherent brain activity, which
is related to melodic processing. (It should be noted that the idea that phase coherence between brain regions
has some functional significance is a topic of active research in modern neuroscience, which enjoys increasing
evidence but which has yet to be conclusively proved.)