Taken together these two studies are consistent with the pathway up to and including the
primary auditory cortex being a necessary and sufficient substrate for the perception of
complex pitch. I have discussed these studies under the heading fine temporal structure;
both studies can be interpreted in terms of a need for primary cortex (particularly on the
right) to perceive fine temporal structure, but the missing fundamental work is equally
consistent with a spectral template model. These data would be consistent with activity of
the human primary cortex being a neural correlate of the conscious perception of complex
pitch.^25 Such processing is important for the perception of the pitch of the individual notes
produced by most musical instruments.
Higher-order temporal structure
Lesion studies can also allow inference about the detection of higher-order temporal pat-
terns of segmented sounds, including music. A number of studies have looked at very simple
tests of segmented sound processing in patients with auditory agnosias including
amusia (see Ref. 19 for review). The tests used include gap detection and click-fusion
threshold. Many studies have shown deficits, and been interpreted in terms of a temporal
processing deficit, but conventional gap detection paradigms are confounded by the pre-
sence of spectral artifact in the stimulus (this is overcome in more refined paradigms such
as those developed by Phillips^26 ). Many of the patients in these studies had large bilateral
lesions involving the superior temporal cortices, consistent with a role for the secondary
and association areas of auditory cortex in this region being needed for the perception of
this type of temporal pattern.
A deficiency in the evaluation of patients with auditory agnosia is the lack of a system-
atic battery of tests of sound sequence processing above the level of gap detection but below
the level of actual musical sequences. Such a battery would have the potential to demon-
strate ‘generic’sound sequence mechanisms that might be relevant to a number of differ-
ent sound processing disorders without higher-level confounds. A crude test of sequence
discrimination in a patient, H.V., with amusia^27 suggested a loss of the perception ofrapid
but not slow sequences. This paralleled the phenomenology in a patient who had lost the
ability to recognize and enjoy music with a high tempo. That patient had a unilateral lesion
affecting the posterior superior temporal lobe and inferior parietal lobe, and the deficit in
his case was less profound than in other cases reported following bilateral lesions.
Auditory agnosia can also be studied by a parallel approach based on the modulation
transfer function as discussed earlier. In terms of the temporal ‘window’, the level of higher-
order structure for patterns of notes (hundreds of milliseconds or seconds) parallels the
level of low modulation rates (of tens of Hz or less). The patient with mild amusia dis-
cussed above (H.V.^27 ) had a mild deficit in the detection of sinusoidal frequency modula-
tion (FM) and amplitude modulation (AM) at a rate of 2 Hz. The dissociation between this
and his profound deficit in the discrimination of rapid sequences of segmented sound was
striking, and cautions against any universal approach in these patients based on modula-
tion characterization. In contrast, the patient I.R., with a more marked amusia due to bilat-
eral lesions, had a striking deficit in the detection of both AM and FM, particularly at low
modulation rates (her threshold for 2-Hz sinusoidal AM was 11 times normal). Apart from
demonstrating auditory deficits linked to the phenomenology, her case allows the inference
172