Regular articleSpectrotemporal features of the auditory cortex: the activation in response to dynamic ripples
Introduction
In this study we focused on sound properties related to speech, the processing of which obviously is an important function of the human brain Belin et al., 2000, Belin et al., 2002, Jäncke et al., 2002, Wong et al., 2002, Wise et al., 2001, Scott et al., 2000. In systematic studies, the structural complexity of vocal sounds is a significant problem when defining their exact properties (Hickok and Poeppel, 2000). Therefore, simpler sound signals have been derived that can be defined with mathematical exactitude, but that still contain some essential characteristics of vocal sounds. An important category of such signals comprises modulated noise. The most basic of these modulated noise signals contain either temporal amplitude modulations Giraud et al., 2000, Wakefield and Viemeister, 1990 or static spectral modulations, also known as rippled noise Supin et al., 1999, Calhoun and Schreiner, 1998, Versnel and Shamma, 1998. However, modulations in the spectrum of speech are dynamic. Similar dynamic behavior can be incorporated into modulated noise by combining temporal and spectral modulations, resulting in a spectral ripple drifting in time along the frequency axis (Fig. 1). Such signals have a similar spectrotemporal envelope as speech and can model formant transitions between vowels and changes in pitch due to intonation. However, the resemblances between dynamic ripples and vocal sounds are limited to acoustic and phonetic properties. Dynamic ripples are not associated with speech on a phonological or linguistical level, involving the recognition of syllables or the comprehension of words.
Dynamic ripples can be characterized by four parameters (Fig. 2): the spectral modulation density Ω (in cycles per octave, c/o), the temporal modulation frequency ω (in cycles per second, c/s), the modulation amplitude A (%), and the drift direction d (downward or upward). Any sound pattern can be regarded as a superposition of such dynamic ripples, since they form a complete orthonormal basis of functions. In electrophysiological animal studies, dynamic ripples play a useful role in characterizing neuronal spectrotemporal response fields Miller et al., 2002, Depireux et al., 2001, Klein et al., 2000, Kowalski et al., 1996. Selective neuronal responses to the speed and direction of spectral changes have been reported using a range of stimuli Liang et al., 2002, deCharms et al., 1998, Tian and Rauschecker, 1998, as well as corresponding topographic mappings on the brain's cortical surface Versnel et al., 1995, Mendelson et al., 1993, Shamma et al., 1993. However, in humans electrophysiological experiments are unfeasible due to the invasive nature of this technique. Psychoacoustical experiments have been performed to measure thresholds for the perceptual detection of spectrotemporal modulations; near an optimum at Ω = 0.5 c/o and ω = 4 c/s, dynamic ripples with amplitudes A as low as 5% can be discerned from noise without modulations (Chi et al., 1999).
We used functional MRI (fMRI), which is a volume imaging technique that is suitable for humans, with the additional advantage of a high spatial resolution. Although the poor time resolution prevented us from directly detecting any synchronization of neuronal responses with the temporal sound modulations, sustained changes in overall activity due to selective neural activation were measurable (Seifritz et al., 2002). Given that the presence of tonotopy in the auditory cortex has previously been demonstrated using fMRI Engelien et al., 2002, Le et al., 2001, Talavage et al., 2000, Bilecen et al., 1998, Wessinger et al., 1997, it may also be able to reveal any topographic organization associated with dynamic ripples. In contrast to animal studies, the activation of the human auditory cortex in response to dynamic ripples has not yet been studied systematically; in this paper we investigate the magnitude and spatial distribution of the cortical activation as a function of the previously mentioned dynamic ripple characteristics.
Section snippets
Subjects
Nine healthy subjects with normal hearing were recruited (eight male, one female; aged 23–59 years, mean 37 years) and gave written informed consent to participate in this study. Of all subjects, four were right-handed, three left-handed, and two ambidextrous (Oldfield, 1971). Subjects received an explanation and a short demonstration beforehand, but were not trained to the task.
Stimulus paradigm
A sound generation setup was used to present the auditory stimuli. Each stimulus presentation consisted of a series
Activation extent and activation level
The cerebral activation caused by the spectrotemporal modulations in dynamic ripples depended on the parameters of the stimulus. Fig. 4 shows the results with respect to the ae, averaged over all subjects. We found that ae is maximal for low modulation densities ( ) and small modulation frequencies (ω = 2 c/s) and decreases monotonically for both Ω and ω. With regard to the modulation amplitude we found that ae increased with A. In contrast, we did not find any significant changes in ae
Activation dependencies
We considered two quantitative measures, the extent (ae) and the level (al) of activation. In principle it is possible for these two measures to show different dependencies on the dynamic ripple parameter values. After all, large volumes that show marginally significant activation might well be imagined, as well as confined regions with high activation levels. Nevertheless, our data show an almost perfect correspondence between both measures (R2 = 0.97); ae shows the same general trends as al.
References (62)
- et al.
Human temporal-lobe response to vocal sounds
Brain Res. Cogn. Brain Res.
(2002) - et al.
Event-related fMRI of the auditory cortex
Neuroimage
(1999) - et al.
Tonotopic organization of the human auditory cortex as detected by BOLD-FMRI
Hearing Res.
(1998) - et al.
Functional fields in human auditory cortex revealed by time-resolved fMRI without interference of EPI noise
Neuroimage
(2001) - et al.
Physiological Mapping of Human Auditory Cortices with a Silent Event Related fMRI Technique
Neuroimage
(2002) - et al.
Hemispheric asymmetries in auditory evoked neuromagnetic fields in response to place of articulation contrasts
Brain Res. Cogn. Brain Res.
(2002) - et al.
Heschl's gyrus is more sensitive to tone level than non-primary auditory cortex
Hearing Res.
(2002) - et al.
Towards a functional neuroanatomy of speech perception
Trends Cogn. Sci.
(2000) - et al.
Phonetic perception and the temporal cortex
Neuroimage
(2002) - et al.
Intensity coding of auditory stimulian fMRI study
Neuropsychologia
(1998)