| |
Abstract:
Recent functional neuroimaging data demonstrate selective
activation of the STS when subjects listen to speech compared with
simple, nonspeech sounds. Although these data suggest participation
of the STS in processing phoneme representations, the dependency of
STS activation on purely acoustic variables has not been defined.
Using functional magnetic resonance imaging, we studied effects of
frequency modulation (FM), time-dependent modulation of FM
amplitude (depth modulation, or DM), and presence of formant-like
spectral peaks as factors influencing STS activation. Stimuli
included repeating steady-state noise bursts (SS), FM noise bursts
(FM), DM noise bursts (DM), pure tone sequences (T), and words (W).
Noise bursts contained energy peaks mimicking speech formants, and
were matched for duration and energy across conditions. Stimuli
were presented in a random, intermixed order. The W-T contrast
produced STS activation, replicating previous results. SS-rest
produced focal activation of dorsal superior temporal gyrus (STG)
in and around Heschl's gyrus (HG), but no activation of STS.
FM-rest produced activation of HG and more extensive activation of
dorsal and lateral STG, but no discernible activation of STS.
Despite greater FM "complexity," the DM stimulus produced somewhat
less activation of STG compared to FM, and no activation of STS.
Presence of energy formants and FM alone probably do not account
for selective activation of STS by speech sounds.
|
