Models for this temporal integration mechanism^31 allow it to occur at an early stage in the
auditory pathway, perhaps as early as the cochlear nucleus, and ongoing work is being
carried out with fMRI to further delimit it to brain stem structures.
The work using IRN therefore also supports the existence of a representation of temporal
regularity in the auditory cortex, with closer control of the auditory spectrum than in the
case of the missing fundamental.
Higher-order temporal structure
In the IRN experiment,^2 higher-order structure was also investigated. The individual notes
with pitches determined by temporal regularity were presented as two pitch patterns: as a
regular pitch ‘staircase’and as novel diatonic melodies. We hypothesized that areas involved
in the analysis of sound sequences would show greater activation as a function of the pitch
strength of the individual notes for the melody condition than for the staircase condition.
Sound sequence analysis represents an emergent property of the detection of the pitch of
individual sounds, and areas involved in sequence analysis will show a more marked
dependence on pitch strength for more complex pitch sequences like the melody condition.
This was demonstrated in four areas, distinct from the primary auditory cortex, in both
anterior temporal lobes near the poles and in the posterior superior temporal lobes. These
areas are shown in Figure 11.3. A scheme is hypothesized, consistent with the data, where the
temporal properties of the individual notes are extracted in the region of the primary audit-
ory cortex before higher-order temporal properties of the stimulus sequence are analysed in
the distinct network shown in Figure 11.3.
A subsequent study used PET to demonstrate areas involved in the processing of simple
sequences that are not musical.^32 Here, simple six-element atonal sequences were presented
to nonmusical subjects in order to demonstrate areas involved in the analysis of both pitch
sequences and sequences of notes of different duration. Pairs of sequences were presented to
subjects where either the pitch sequence, the duration sequence, neither, or both were var-
ied between the two. During any scan, subjects were required to detect either changes in
pitch sequence or duration sequence. The results showed, again, a network of activation dis-
tinct from the primary auditory cortex (in addition to activation in the primary auditory
cortex in the principal comparison between the sound conditions and rest). Bilateral activa-
tion was demonstrated in the posterior superior temporal planum in the region of the
planum temporale, and in both frontal opercula, with greater spatial extent of the clusters
of activity demonstrated on the right. Highly significant activation was also demonstrated
in both lobes of the cerebellum. The activation during the pitch sequence and duration
sequence tasks was strikingly similar, and direct comparison of the two conditions showed
no significant differences in activation. This would be consistent with a common mechan-
ism for the two tasks at some level, consistent with some work in the musical domain.
These studies using ‘submusical’stimuli, therefore, suggest the existence of networks for
processing of higher-order structure in both temporal lobes that are distinct from the prim-
ary auditory cortices. The planum temporale is consistently activated in a number of
studies, with less consistency in activation anterior to the primary auditory cortex.
Activation anterior to the auditory cortex occurs in the anterior temporal lobe in the IRN
174