dominated by two main theories: one that postulates an exclusively tonotopic coding and
another that is in favour of an exclusively ‘temporal’ coding.1–3 Recently, however,
Langner4,5has suggested that temporal information, coded at the subcortical level as spikes
synchronized to the periodicity of the acoustic signal, is subsequently converted to spatial
information coded in the form of tonotopic maps in the auditory cortex. This hypothesis
is consistent with the finding that the temporal resolution of the auditory nerve^6 and inferior
colliculus7,8may be higher than at the cortical level, where the temporal resolution of
auditory information is possible only at frequencies below 100 Hz for the majority of
neurons.9,10At the cortical level, pitch appears to be coded according to a frequency
spectrum (i.e. tonotopically).
Experimental studies using single-unit recording techniques indicate that for most aud-
itory neurons, a ‘best frequency’ (BF) can be determined at which a low-intensity auditory
stimulus evokes the greatest electrophysiological response in a given region. The auditory
cortex is said to be tonotopically organized because an ordered change in BFs occurs as one
moves along the cortex. The auditory cortex is organized according to a frequency gradient,
with low frequencies represented laterally and high frequencies represented medially.11–18
The tonotopic organization of the auditory cortex is best understood in nonhuman
animals. The cytoarchitecture of the auditory cortex is not always uniform,^14 however,
rendering the mapping of different tonotopic regions difficult. Merzenich and Brugge^12
have found that the auditory cortex in the superior temporal plane of the macaque can be
subdivided into at least six distinct fields on the basis of cytoarchitectonic characteristics
and BFs. Likewise, Reale and Imig^16 observed a tonotopic organization in six different
auditory areas of the cat, and a similar tonotopic organization has been described in the
albino rat.^18
In humans, the use of noninvasive techniques such as magnetoencephalography (MEG)
and positron emission tomography (PET) has shed some light on the functional organ-
ization of the auditory cortex with respect to frequency. In their PET study, Lauter et al.^19
observed that changes in regional cerebral blood flow induced by a high-frequency tone
(4 kHz) were located deeper and more posteriorly in the temporal lobe than those induced
by a low-frequency tone (500 Hz). In a similar vein, MEG studies on the tonotopic organiza-
tion of N1 generators have shown that their estimated depth can vary as a function of fre-
quency: higher frequencies involve more medial regions of the Heschl’s gyrus (HGs),
whereas lower frequencies involve more lateral areas.20–24
In a previous study,^25 using depth electrodes implanted in the HG of human subjects with
medically intractable epilepsy, we identified four separable components of middle-latency
auditory evoked potentials (MLAEPs), which peaked at approximately 30, 50, 60, and 80 ms.
The sources of these components were definable, although some overlap was observed. In a
recent MEG study, Pantev et al.^23 explored the sources of different middle-latency com-
ponents as well as their distribution. In all subjects and for all the frequencies studied, only two
components were consistently observed: the 30-ms ‘Pam’ component and the N1m. Also, a
tonotopic organization was observed in the primary auditory cortex that mirrored that
observed in the secondary cortex. The localization of Pam’s source moved more superficially
with higher frequencies, while the localization of the N1m’s source moved progressively
deeper with higher frequencies. The authors, however, did not consistently observe the
153