different leads were studied for each patient. Tonotopic maps were based on a comparison
of the results obtained from patients with different anatomical recording sites.
The acquisition system was a Nicolet Pathfinder II with 16 acquisition channels and an
8-bit-resolution digital converter. Negative potentials were displayed as up and positive as
down. Recordings of brainstem evoked potentials carried out before SEEG recordings
confirmed that all the patients studied had normal afferent pathway conduction.
Temporospectral mapping
Data were stored on the hard drive of the computer and transferred via an asynchronous
line to an IBM PC/AT. Off-line processing consisted of the construction of temporospec-
tral maps, using a technique adapted from Rémond,^34 to depth recordings, particularly
appropriate to the depth electrode geometry.35,36
Results
Primary and secondary areas were identified on the basis of component latency (for a
description, see Ref. 25). The selectivity of an auditory area’s response (and hence its BF
was determined from frequency-related fluctuations in a component’s amplitude and/or
polarity inversions.
Recordings from the primary auditory cortex
Right hemisphere Figure 10.2 shows AEPs recorded from 10 contiguous leads of an elec-
trode implanted in the right auditory cortex of one patient. The first four leads recorded
activity from the primary auditory cortex; the other six recorded activity from secondary
areas. Tone bursts at different frequencies were presented to the patient’s left ear. A tripha-
sic N30/P50/N80 complex was recorded from the first four leads for which a polarity inver-
sion was observed between leads 4 and 5, as well as a shift in latency between leads 6 and
- AEPs recorded from lead 4 were highly sensitive to tonal frequency. For tones higher
than 1 kHz, the amplitude of the 50- and 80-ms components progressively decreased until
an isopotential response (note the polarity inversion between leads 3 and 5) was observed
for tones at 4 kHz, suggesting that the cortical region at lead 4 has a high sensitivity for
frequencies in the 4 kHz range. AEPs did not differ in terms of tonal frequency on any of
the other leads.
It should be noted, of course, that the amplitude changes observed were never an all-
or-none phenomenon, but were observed as increasing or decreasing from lead to lead in
progressive steps. Also, unlike the 50- and 80-ms components, the amplitude of the 30-ms
component did not appear to be frequency dependent.
Figure 10.3 is a temporospectral representation of AEPs recorded for different tonal fre-
quencies from a second patient. In this figure, the amplitudes of the same three com-
ponents (30, 50, and 80 ms) as for the first patient are shown for the three most responsive
leads (2, 3, and 6 for different tonal frequencies). In each case, frequency is represented
along the x-axis and latency along the y-axis. On lead 2, amplitude was maximal for tone
bursts with the highest frequencies (3–4 kHz), although frequency-dependent fluctuations
156