Cerebral Cortex Advance Access originally published online on November 9, 2005
Cerebral Cortex 2006 16(9):1266-1275; doi:10.1093/cercor/bhj068
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Hemispheric Specialization for Processing Auditory Nonspeech Stimuli
1 Centre for Functional Magnetic Resonance Imaging of the Brain and 2 Department of Experimental Psychology, University of Oxford, Oxford, UK
Address correspondence to Helen L. Jamison, Centre for Functional Magnetic Resonance Imaging of the Brain, John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK. Email: jamison{at}fmrib.ox.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
The left hemisphere specialization for speech perception might arise from asymmetries at more basic levels of auditory processing. In particular, it has been suggested that differences in "temporal" and "spectral" processing exist between the hemispheres. Here we used functional magnetic resonance imaging to test this hypothesis further. Fourteen healthy volunteers listened to sequences of alternating pure tones that varied in the temporal and spectral domains. Increased temporal variation was associated with activation in Heschl's gyrus (HG) bilaterally, whereas increased spectral variation activated the superior temporal gyrus (STG) bilaterally and right posterior superior temporal sulcus (STS). Responses to increased temporal variation were lateralized to the left hemisphere; this left lateralization was greater in posteromedial HG, which is presumed to correspond to the primary auditory cortex. Responses to increased spectral variation were lateralized to the right hemisphere specifically in the anterior STG and posterior STS. These findings are consistent with the notion that the hemispheres are differentially specialized for processing auditory stimuli even in the absence of linguistic information.
Key Words: asymmetry • fMRI • lateralization • spectral • temporal
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
The possibility that the hemispheres are differentially specialized for processing information in the auditory domain has received a great deal of attention. One classical view posits that the left hemisphere is specialized for speech perception (Wernicke 1874
). It is often also suggested that the right hemisphere is specialized for processing certain aspects of music (Bever and Chiarello 1974; Zatorre and others 1994; Ohnishi and others 2001; Tervaniemi and Hugdahl 2003). In the last 30 years, the debate over the nature of hemispheric specialization for auditory processing has grown, and the question as to its precise origin remains unresolved. It has been postulated that the left hemisphere lateralization associated with speech perception, for example, arises as a result of the linguistic relevance and intelligibility of the input (Scott and others 2000; Shtyrov and others 2000; Narain and others 2003; Takegata and others 2004). An alternative hypothesis is that hemispheric specialization at higher levels of auditory cognition exists as a result of asymmetries in more basic auditory processing. Evidence for this hypothesis arose from an original study by Efron (1963), who showed that a group of aphasic patients had difficulty processing rapidly occurring auditory nonspeech stimuli. A number of investigations that followed proposed that rapid temporal processing might be crucial for speech comprehension (Tallal and Piercy 1973, 1974, 1975; Schwartz and Tallal 1980), although this particular viewpoint remains controversial.
The nature of hemispheric specialization for processing acoustic cues that may be relevant for speech processing is a central issue. A number of studies confirm a left hemisphere specialization for rapid temporal processing (Belin and others 1998; Liegeois-Chauvel and others 1999; Nicholls and others 1999; Zaehle and others 2004; Yamasaki and others 2005), whereas others have argued that the evidence is far from compelling (Scott and Wise 2004). One of the more convincing experimental demonstrations that the hemispheres exhibit functional asymmetry, even in the absence of linguistic information, comes from Zatorre and Belin (2001). Using positron emission tomography (PET) they showed that the primary auditory cortex (PAC) (Heschl's gyrus [HG]) responds to increasing temporal variation, whereas nonprimary areas (anterior superior temporal gyrus [STG] in particular) respond to increasing spectral variation. Moreover, they demonstrated that responses to temporal variation were left lateralized, whereas those to spectral variation were right lateralized. From these findings they argued that a trade-off exists between the hemispheres, the left with finer temporal resolution at the expense of spectral resolution and the right with finer spectral resolution at the expense of temporal resolution.
A similar hypothesis suggests that functional asymmetries in the auditory domain might be related simply to the time frames over which auditory information is processed in each of the hemispheres. Poeppel (2003), for example, claims that the left hemisphere extracts information from a short temporal integration window (
20–40 ms), whereas the right operates over a longer time frame (150–250 ms; although the true windows may not be so specific). Recent experimental evidence adds weight to the notion that left–right differences occur as a result of differential feature extraction at early stages of auditory processing (Boemio and others 2005). It suggests that both hemispheres are sensitive to temporal structure and that this response becomes right lateralized as temporal structure increases, rather than as a result of increasing spectral complexity. Data from Patterson and others (2002) also support the hypothesis that the right hemisphere specialization for processing pitch information does not necessarily depend on spectral resolution. However, it is possible that temporal integration windows may be better thought of as relative rather than absolute and might depend on experience. For example, hemispheric specialization for processing particular acoustic cues has been suggested to vary with linguistic experience (Gandour and others 2000, 2003, 2004; Hsieh and others 2001).
Clearly, the precise nature of functional asymmetry and lateralization in auditory cortex is a complex issue that remains to be resolved. We therefore decided to test whether previous demonstrations of hemispheric specialization are replicable and to investigate lateralization effects in greater detail at the individual level. We used a set of stimuli similar to that of Zatorre and Belin (2001) in a functional magnetic resonance imaging (fMRI) experiment to investigate the proposed asymmetries in auditory nonspeech processing. Our specific aims were 1) to test the hypothesis that responses to increased temporal variation are left lateralized, whereas those to increased spectral variation are right lateralized, and 2) to examine the consistency of the proposed hemispheric asymmetries at the individual level as such data were not available in the original PET study. In the original study (Zatorre and Belin 2001), responses to increasing temporal variation were apparently restricted to the "core" or primary auditory areas. As PAC is thought to occupy only the posteromedial two-thirds of HG (Rademacher and others 1993; Penhune and others 1996; Hall and others 2002), we also used a specific region-of-interest (ROI) approach to test the hypothesis that activation in posteromedial HG is left lateralized in response to increased temporal variation, compared with increased spectral variation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Subjects
Fourteen healthy volunteers (7 females, 7 males) with normal hearing (passed hearing screen at 20 dB) and minimal musical experience participated in the study, and all had English as their first language. Their ages ranged from 18 to 34 years (mean = 23.6 years), and all were right handed according to a handedness questionnaire adapted from the Edinburgh Handedness Inventory (Oldfield 1971). Participants gave written consent before taking part. Ethical approval was granted by the Central Oxford Research Ethics Committee.
Auditory Stimulation
Stimuli comprised pure tone patterns, identical to those used by Zatorre and Belin (2001) but segmented at 6-s intervals for the purposes of this experiment. They were taken from 3 conditions in which the frequency and duration of individual tones were varied (see Fig. 1 and Zatorre and Belin 2001 for further details). In the first condition ("standard" condition), stimuli consisted of alternating pure tones at 500 and 1000 Hz (i.e., separated by one octave). Stimuli were generated with a randomly determined duty cycle where the tone with the shortest duration and highest probability of occurrence was 667 ms. They were ramped on and off in counterphase using a 10-ms cosine function in order to avoid acoustic transients. The second condition ("temporal" condition) used stimuli with greater temporal rate of alternation. Stimuli again consisted of alternating pure tones at 500 and 1000 Hz, but in this condition the duration of the shortest possible tone, which had the highest probability of occurrence, was 21 ms (i.e., shorter mean duration). In the third condition ("spectral" condition) the rate of presentation of tones remained the same as in the standard (i.e., longer mean duration), but an increased number of frequencies were sampled in each stimulus. The octave range from 500 to 1000 Hz was divided into 32 logarithmically, and frequencies were selected at random from this distribution. The use of the term "spectral" to denote this third condition refers to the inclusion of an increased number of spectral components, over time, in each 6-s stimulus. It is not meant to imply, however, that the individual component tones of each spectral stimulus were harmonic-complex tones. Stimuli from this condition give a perception of dynamic pitch variation. Also of note is that the temporal and spectral conditions are equivalent to those at the maximal level of variation used by Zatorre and Belin (2001). It was not possible to include every level of variation employed by the previous authors owing to the time constraints imposed by sparse sampling fMRI. A baseline condition of silence was also included.
|
Procedure
Stimuli were presented binaurally at 72–77 dB sound pressure level using the sound system custom made by the Medical Research Council Institute of Hearing Research (Palmer and others 1998) with magnetic resonance–compatible electrostatic headphones (Sennheiser model HE 60) and modified industrial ear protectors (Bilsom model 2452). Headphone orientation was reversed for half of the subjects. Subjects were instructed to lie motionless and listen to the sounds, without performing any explicit task (they were familiarized with the stimuli prior to scanning but were not told to attend to any specific stimulus characteristics). Each subject participated in 2 functional imaging runs, which lasted 10 min and were counterbalanced across subjects. Within each run, 10 "trials" from each condition (standard, temporal, spectral, and silence) were pseudorandomly presented in an event-related manner.
A "sparse sampling" (Hall and others 1999) design was employed, with 1-s image acquisition occurring every 15 s, immediately after the end of each 6-s auditory stimulus. This ensured that the hemodynamic response was sampled at or near its peak, which occurs around 5–6 s after stimulus onset in auditory cortex depending on stimulus duration (Hickok and others 1997; Belin and others 1999; Glover 1999; HL Jamison, unpublished data). A period of 8-s silence followed each acquisition, to allow the hemodynamic response to return to baseline levels prior to presentation of the next stimulus (Belin and others 1999). This design ensured that all stimuli were presented in silence and not masked by the scanner noise. It also served to minimize any overlap between stimulus-related responses and those related to the noise generated by the scanner, which are known to interact in a complex and nonlinear fashion (Talavage and Edmister 2004).
Imaging
All scanning was performed on a Varians-Siemens 3-T scanner at the Centre for Functional Magnetic Resonance Imaging of the Brain. A Magnex head-dedicated gradient insert coil was used in conjunction with a birdcage head radiofrequency coil tuned to 127.4 MHz. Functional imaging consisted of 13 T2*-weighted echo-planar image (EPI) slices (long repetition time [TR] = 15 s, short repetition time [tr] = 1 s, echo time [TE] = 30 ms) 3-mm thick with an in-plane resolution of 4 mm2. Slices were angled obliquely in parallel with the Sylvian fissure and centered on HG viewed on a parasagittal slice. In addition, for registration and anatomical localization purposes, a whole-brain T1-weighted scan with 1-mm2 in-plane resolution and either 1.5- or 3-mm slice thickness (3D Turbo FLASH sequence, TR = 15 ms, TE = 6.9 ms) was acquired for each subject.
Data Processing
After removing the first image of each session to allow for T1 magnetization equilibrium, functional images were realigned (Jenkinson and others 2002) using the FSL software (http://www.fmrib.ox.ac.uk/fsl) in order to correct for small head movements. In 1 session for 1 subject artifactual activation within the ventricles and around the edges of the brain occurred; as these phenomena are typical of excessive head movement, this session was removed from the analysis. Such artifacts were not present in any of the remaining data sets analyzed. Functional images were registered to the participant's structural scan and then to the MNI 152-average brain. Registration was checked to ensure accurate positioning of the raw EPI data relative to each subject's anatomy. Finally, each image was smoothed with a 5-mm full-width half-maximum Gaussian filter. The FSL software was used to compute individual subject analyses, in which each of the 3 conditions were modeled separately, and mixed-effects group analyses were used to identify those areas that were significantly activated in each condition. Statistical images were thresholded at Z > 2.3 and P < 0.05 (cluster thresholding, corrected for multiple comparisons).
Analyses
The standard, temporal, and spectral conditions were directly contrasted to identify those areas of auditory cortex differentially responsive to increased variation in the temporal and spectral domains. ROI analyses were then conducted to compare activation levels between the hemispheres for those regions identified. In the first instance, ROIs were defined functionally, on the basis of activation clusters from the group analysis, to test the hypothesis that responses to greater temporal variation are left lateralized, whereas those to increased spectral variation are right lateralized.
To test responses in those areas described by Zatorre and Belin (2001) as demonstrating left and right lateralizations (i.e., HG and anterior STG, respectively), we also used a more regionally specific ROI approach. ROIs in HG were defined on the basis of individual anatomical information. In accordance with cytoarchitectonic data, the borders of HG were delineated anteriorly as the first transverse sulcus and posteriorly as Heschl's sulcus (Penhune and others 1996; Rivier and Clarke 1997). In cases where more than one gyrus existed in the same hemisphere, only the region anterior to the second transverse sulcus was included (Rademacher and others 1993; Rivier and Clarke 1997). As PAC is thought to occupy only the posteromedial two-thirds of HG (Rademacher and others 1993; Penhune and others 1996; Hall and others 2002), we divided HG into posteromedial and anterolateral portions. Posteromedial HG was defined as two-thirds of the medio-lateral extent in the X plane and two-thirds of the posteroanterior extent in the Y plane. Anterolateral HG (which is not thought to contain PAC) was defined as encompassing one-third of the lateral extent of HG in the X plane and one-third of the anterior extent in the Y plane. Previously, Wallace and others (2002) have described this region as the anterolateral area. As auditory areas outside of HG (e.g., anterior STG) have no easily definable anatomical landmarks (Rivier and Clarke 1997), spherical ROIs with a 5-mm radius were centered on coordinates from the group analysis.
Once selected, ROIs of a specific type were employed as masks to extract 2 types of information, namely, the number of significantly activated voxels (at Z > 2.3, P < 0.01) and the mean percent signal change (averaged over the ROI) in the blood oxygen level–dependent (BOLD) response. To test for lateralization, paired samples t-tests were run on these data for the left and right hemispheres in each subject. Laterality indices (LI) were also calculated: LI = (right BOLD signal – left BOLD signal)/(right BOLD signal + left BOLD signal). LI values ranged from –1.0, indicating a complete left lateralization, to +1.0, indicating a complete right lateralization. In calculating LI, negative BOLD values (relative deactivation in the particular contrast) were set to zero so that the LI reflected a complete lateralization to the opposite hemisphere; this prevented the analysis defining lateralization on the basis of relative deactivation rather than activation. LI values for each region were tested against zero using 1-sample Wilcoxon tests (nonparametric) as they were expected to have a non-Gaussian distribution. There has been some question as to the most appropriate method of both defining ROIs and determining laterality in fMRI data (Rutten and others 2002; Adcock and others 2003); each technique has its own limitations. For example, LI calculations simply reflect a ratio of counts and are artificially "capped" at maximum and minimum values; they do not take into account the underlying variation. Therefore, we used the aforementioned combination of measures in order to obtain a more robust estimate of laterality than that afforded by any single measure alone, the use of which could potentially have biased results.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Analysis of the imaging data revealed a main effect of auditory stimulation in extensive regions of the primary and nonprimary auditory cortices for all 3 conditions (standard, temporal, and spectral) relative to silence. In both hemispheres, areas of activation were identified as HG, the STG, and the upper bank of the posterior superior temporal sulcus (STS).
Temporal versus Spectral Contrasts
Contrasting the temporal condition with the spectral condition (temporal > spectral) revealed significant activation along HG in both hemispheres (Fig. 2a). At the group level, the cluster in the left hemisphere contained more voxels than that in the right (850 vs. 504 voxels; Table 1). To examine this lateralization in more detail, data were extracted using whole activation cluster masks from the group analysis. Left and right cluster masks were flipped about the x axis and added together to ensure that they were symmetrical in size and location. Across individuals, the number of activated voxels (Z > 2.3) was on average greater in the left hemisphere cluster, although this difference narrowly missed significance (t(13) = 2.00, P = 0.068). The mean percent BOLD signal change was then calculated for each mask in each subject. A paired t-test revealed that activation for the temporal > spectral contrast was significantly greater in the left hemisphere than the right (t(13) = 2.82, P = 0.014) (Fig. 3). Twelve of the 14 subjects individually demonstrated this left lateralization, and the group mean LI of –0.24 ± 0.08 was significantly different from zero (Z = 2.48, P = 0.013).
|
|
|
At the group level, the reverse contrast (spectral > temporal) revealed a significant activation cluster in the right posterior STS but not in the left (197 vs. 0 voxels; Fig. 2a and Table 1). This cluster mask was flipped about the x axis and used to examine lateralization at the individual level. Data extracted from this region in the left and right hemispheres showed that the number of activated voxels (Z > 2.3) was significantly greater in the right hemisphere cluster (t(13) = 2.20, P = 0.046). The mean percent BOLD signal change was also calculated for each mask in each subject, and a paired t-test revealed that activation was significantly greater in the right hemisphere than the left (t(13) = 5.63, P < 0.001) (Fig. 4). Twelve of the 14 subjects individually demonstrated this right lateralization, and the group mean LI of +0.71 ± 0.10 was significantly different from zero (Z = 3.10, P = 0.002). In the original PET study (Zatorre and Belin 2001
), the same contrast revealed significant activation in the anterior STG bilaterally. To test for a similar response, we also examined our data using a relaxed statistical threshold. At an uncorrected voxel threshold of P < 0.01, we identified activation in both the right (Z = 3.39) and left anterior STG (Z = 3.03) (Table 1). We examined lateralization in this region for the spectral > temporal contrast using spherical ROIs with a radius of 5 mm, centered on the coordinates of these peaks (left x = –52, y = 10, z = –10; right x = 54, y = 12, z = –14). Although the mean percent BOLD signal change was on average greater in the right hemisphere, this difference did not reach significance. Differences in the number of active voxels (Z > 2.3) per ROI in each hemisphere were also not significant.
|
Temporal > Standard Contrast
Contrasting the temporal with the standard condition (temporal > standard) revealed significant bilateral activation in a large region centered on HG and including surrounding regions of the STG (Fig. 2b). Both left and right clusters of activation included the posteromedial portion of HG, which is thought to be the location of PAC (Rademacher and others 1993; Penhune and others 1996; Hall and others 2002). At the group level, the spatial extent of the left hemisphere cluster was greater than that observed on the right (1512 vs. 1383 voxels; Table 2), and across individuals the number of activated voxels (Z >2.3) per cluster mask was significantly greater on the left (t(13) = 2.65, P = 0.020). A paired t-test revealed that mean percent BOLD signal change for the temporal > standard contrast was also significantly greater in the left hemisphere than the right (t(13) = 3.05, P = 0.010) (Fig. 5a). Eleven of the 14 subjects individually demonstrated this left lateralization, and the group mean LI of –0.19 ± 0.06 was significantly different from zero (Z = 2.29, P = 0.022).
|
|
In the previous study by Zatorre and Belin (2001)
, responses to increased temporal variation were apparently restricted to the core auditory areas. Using our high-resolution magnetic imaging scans we decided to examine this finding in more detail and with respect to each individual subject's anatomy. Activation was examined within posteromedial and anterolateral HG, which were defined as described in Materials and Methods for each subject separately. As ROIs were defined on an individual anatomical basis, they were not identical in size in each hemisphere. The number of significantly active voxels in each ROI was therefore calculated as a proportion of the total number of voxels in the ROI. In posteromedial HG, the proportion of activated voxels in the left was significantly greater than that in the right, at Z > 2.3 (t(13) = 3.56, P = 0.004). Also, 12 of 14 individuals had a relative left lateralization in mean percent BOLD signal change, with a group mean LI of –0.16 ± 0.07 that was significantly different from zero (Z = 2.36, P = 0.018) (Fig. 5b); a paired t-test comparing left and right mean signal changes narrowly missed significance (t(13) = 2.07, P = 0.059). In anterolateral HG, differences between the hemispheres in the proportion of activated voxels were not significant. The mean LI calculated on the basis of signal change in anterolateral HG was 0.00 ± 0.09, which was also not significant, and an equal number of subjects demonstrated left and right lateralization (Fig. 5c). These results indicate that the left lateralization observed for the temporal > standard contrast is more consistent in the posteromedial portion of HG, the putative location of PAC, and support the previous findings that increased temporal variation activates primarily core areas. Spectral > Standard Contrast
Contrasting the spectral with the standard condition (spectral > standard) revealed significant activation in lateral and anterior STG bilaterally, including the anterolateral but not the posteromedial portion of HG (Fig. 2c). In the right hemisphere, this cluster spread anteriorly along the STG to a greater extent than it did in the left (784 vs. 413 voxels; Table 2). As before, whole activation cluster masks were flipped about the x axis and added together to test the lateralization of responses related to this contrast. Across individuals, the number of activated voxels (Z > 2.3) was on average greater in the right hemisphere cluster, although this difference did not reach significance. A paired t-test also did not reveal a significant difference in mean percent BOLD signal change between the hemispheres for this contrast (Fig. 6a). The mean LI was +0.11 ± 0.10 and not significantly different from zero; 8 of 14 subjects individually demonstrated a right hemisphere lateralization.
|
In the previous study (Zatorre and Belin 2001
), it was reported that responses to increasing spectral variation were right lateralized in anterior STG. We therefore decided to examine lateralization specifically in this region. For this analysis, spherical ROIs with a 5-mm radius were placed in anterior STG in both hemispheres to compute local means. These were centered on the coordinates of local maxima in this area from the group analysis for the spectral > standard contrast. As there was an absence of suprathreshold activity at the group level in the left anterior STG, the ROI from the right hemisphere was flipped about the x axis (center voxel coordinates: x = ±56, y = 12, z = –12; these coordinates are similar to those identified in this region by the spectral > temporal contrast also, Table 1). A second pair of ROIs was centered on peak coordinates from the group analysis (spectral > standard contrast) in lateral STG to provide a comparison (center voxel coordinates: left x = –54, y = –14, z = 8; right x = 64, y = –12, z = 4; Table 2). In anterior STG, there was a significant right lateralization in mean percent BOLD signal change for the spectral > standard contrast (t(13) = 2.76, P = 0.016). Eleven of 14 subjects demonstrated this pattern of activation, with a group mean LI of +0.46 ± 0.16 that was significantly different from zero (Z = 2.28, P = 0.023) (Fig. 6b). The average number of activated voxels (Z > 2.3) in this ROI was greater in the right hemisphere, although this difference did not reach significance. In lateral STG left–right differences were not significant (mean LI was –0.03 ± 0.14; 6 of the 14 subjects demonstrated a right lateralization) (Fig. 6c). These findings suggest that there was a significant and specific right lateralization at the subject level in the anterior STG for the spectral > standard contrast, which was not apparent from the initial ROI analysis based on large functional activation clusters.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
The results of our fMRI study show that auditory stimuli with increased temporal variation activate HG bilaterally, whereas stimuli with increased spectral variation activate the STG bilaterally and a region of the right posterior STS. Relative activation in HG was left lateralized in response to increased rate of temporal alternation, whereas activity in posterior STS and anterior STG was right lateralized with increased number of spectral elements. Exploratory analyses also suggested that the left lateralization observed in response to increased temporal variation is greatest in the posteromedial portion of HG, which contains putative PAC.
Differentiating Primary and Nonprimary Auditory Cortices
The present data are consistent with research documenting the involvement of core auditory cortex in fine-grained temporal processing and the proposed hierarchical analysis of sound (Griffiths and others 1998; Gandour and others 2002; Boemio and others 2005). Recent imaging studies have shown that responses to pure tones are greatest in PAC, whereas areas outside this core region are more responsive to increasing temporal structure and complexity in the spectral domain (Wessinger and others 2001; Patterson and others 2002). Distinctions between primary and nonprimary auditory cortices have also been made on the basis of sustained and transient temporal coding (Seifritz and others 2002) and phasic response patterns (Harms and others 2005).
Areas found to be specifically activated by increased spectral variation in our study corresponded to nonprimary auditory cortex. Previously, Thivard and others (2000) demonstrated that lateral "belt" areas are more sensitive to stimuli containing spectral changes. Activity in nonprimary auditory cortex overlapping anterolateral HG has also been shown to correlate with pitch salience (Penagos and others 2004), and the observations of others support the notion that nonprimary regions are specialized for pitch processing and the spectral analysis of sounds (Patterson and others 2002; Warren and others 2003, 2005). Similarly, areas of the right STS, which was identified in this study by comparing spectral and temporal processing directly, have been highlighted as particularly responsive to aspects of vocal information such as prosody and intonation (Belin and others 2000, 2002; Lattner and others 2005).
Functional Asymmetry in Auditory Cortex
Our data show that when the temporal and spectral conditions were directly compared, resulting patterns of activation were asymmetric. Relative to the spectral condition, responses to the temporal condition were significantly greater in the left hemisphere (in HG), whereas suprathreshold activation in response to the converse contrast (i.e., spectral > temporal) occurred only in the right hemisphere (in STS). Given that more than one parameter varies between the 2 conditions, however, it is not possible to ascertain whether it is variation in the temporal or spectral domain, or both, that is responsible for these effects. Therefore, each condition was also separately compared with the standard condition to test more specifically whether responses to increased temporal variation or increased spectral variation were lateralized.
Increased temporal variation, relative to the standard condition, resulted in activation that was greater in spatial extent in the left hemisphere. ROI analyses based on functional activation clusters confirmed a significant left lateralization in signal change, which was relatively consistent across individuals. For the spectral > standard contrast, signal change extracted for each subject from functionally defined large cluster ROIs did not show a significant right lateralization. We speculate that this is due to the inclusion of anatomical regions with potentially different functional sensitivity within such relatively large ROIs, which may have obscured smaller, regionally specific effects; it has previously been shown that restricting analyses to specifically defined regions systematically improves laterality calculations (Rutten and others 2002). Indeed, our analysis using smaller ROIs centered on peak activation voxels in anterior STG confirms the pattern seen at the group level, that is, significantly greater activation in the right hemisphere compared with the left.
Using a specific ROI approach, we were also able to show that the left hemisphere lateralization observed in response to increased temporal variation was potentially greater in the posteromedial portion of HG, which has previously been shown to contain PAC (Rademacher and others 1993; Penhune and others 1996; Rivier and Clarke 1997). The anterolateral portion of HG did not show any consistent lateralization. The spherical ROI analysis showed significant right hemisphere lateralization in anterior STG specific to increased spectral variation. This asymmetry was relatively consistent across individuals and concordant with the finding of Zatorre and Belin (2001) that spectral changes recruit anterior regions of the STG bilaterally, with a greater response on the right. It must be noted, however, that although lateralization effects were significant in posteromedial HG and anterior STG, it was not the case that every subject consistently demonstrated lateralization in response to increased temporal or spectral variation.
Our findings are consistent with evidence that regions within the superior temporal cortex are independently specialized for processing different dimensions of auditory stimuli, in the absence of linguistic information. They support the notion that the left hemisphere is specialized for temporal processing. Increased variation in the temporal domain produced activity in and around HG, which exhibited a consistent left lateralization. Evidence for the left hemisphere specialization for rapid temporal processing has been reported before (Mills and Rollman 1980; Belin and others 1998; Nicholls and others 1999; Zaehle and others 2004; Yamasaki and others 2005). Our finding that specific regions within the right hemisphere are more responsive to increased spectral variation is also consistent with patient data, for example, which demonstrate that lesions affecting the right temporal cortex impair specific spectral processing skills (Sidtis and Volpe 1988; Robin and others 1990; Alcock and others 2000; Johnsrude and others 2000; Hyde and Peretz 2004; Murayama and others 2004). Left–right asymmetries are even noticeable in the peripheral auditory pathways of infants and neonates as early in processing as at the level of the cochlea (Khalfa and others 1998; Morlet and others 1999; Sininger and Cone-Wesson 2004).
There are a number of hypotheses that might account for these findings. The first of these is the spectral–temporal trade-off hypothesis of Zatorre and Belin (2001). The present data, particularly those concerning the direct contrasts between the temporal and spectral conditions, support the notion that the left hemisphere has finer temporal resolution at the expense of spectral resolution, whereas the right has finer spectral resolution at the expense of temporal resolution. The authors speculate that these differences might be related to structural asymmetries between the hemispheres. For example, it has been proposed that wider column spacing and greater axon myelination may facilitate rapid temporal processing in the left hemisphere (Penhune and others 1996; Anderson and others 1999; Buxhoeveden and Casanova 2000; see Hutsler and Galuske 2003, for a full review of structural asymmetries).
A similar explanation of these asymmetries, provided by the hypothesis of Poeppel and colleagues, is that they arise due to differences in the time frames over which auditory information is processed by the hemispheres. In this framework (Poeppel 2003; Boemio and others 2005), functional asymmetries in auditory processing are driven by differences in temporal structure between stimuli. Indeed, evidence exists to suggest that the right hemisphere lateralization, observed in this and other studies in response to dynamic pitch variation, need not depend on the spectral properties of the stimulus (Griffiths and others 1998; Patterson and others 2002). Furthermore, rather than having structural asymmetries as their substrate, Boemio and others (2005) propose that hemispheric differences in temporal processing, in areas such as the STS, might be based on differential input from the STG.
Another explanation for our findings is that there may have been differences in sequential integration between each of the conditions. In the temporal condition, for example, it might be the case that the sequence of rapidly alternating pure tones was perceived as 2 separate auditory streams, with tones being perceptually grouped according to attributes such as their pitch rather than their order in time. However, it is unlikely that auditory stream segregation accounts for our findings because the stimulus duration (6 s) was shorter and the frequency difference between tones (500 Hz) smaller than typically needed for this percept to occur (Bregman 1978; Rose and Moore 2000). Furthermore, stream segregation is most commonly observed when stimuli are patterned, for example, in ABA_ABA triplets (van Noorden 1975), and when listeners are specifically instructed to "listen out" for 2 separate streams (Rose and Moore 2000), and neither of these conditions apply to the present study.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Here we have extended observations showing that specific regions within the superior temporal cortex process acoustic cues asymmetrically even in the absence of linguistic information. Furthermore, we have tested this effect at the individual level, which will facilitate relating varying functional asymmetries to differences in structure. There are a number of possible explanations for this apparent hemispheric specialization; however, we believe that our data cannot be explained on the basis of differences in auditory stream segregation between the temporal and spectral conditions. When selecting subjects we controlled for linguistic and musical experience, both of which are known to affect functional asymmetry (Ohnishi and others 2001
; Gandour and others 2003); we also believe, therefore, that differences in such experience are unlikely to account for the lateralization patterns observed here. In conclusion, our data do support the proposal that the left hemisphere is specialized for processing auditory stimuli containing rapid temporal changes and that the right hemisphere is specialized for processing changes in frequency information over a longer time window. These conclusions are in accord with both the spectral–temporal trade-off hypothesis (Zatorre and Belin 2001) and the notion of asymmetric temporal integration windows (Poeppel 2003; Boemio and others 2005). |
Acknowledgments |
|---|
We thank Robert Zatorre for the kind donation of auditory stimuli. This work was supported by the Wellcome Trust (HLJ and DVMB) and the Medical Research Council (KEW and PMM).
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Adcock JE, Wise RG, Oxbury JM, Oxbury SM, Matthews PM. 2003. Quantitative fMRI assessment of the differences in lateralization of language-related brain activation in patients with temporal lobe epilepsy. Neuroimage 18:423–438.[CrossRef][ISI][Medline]
Alcock KJ, Wade D, Anslow P, Passingham RE. 2000. Pitch and timing abilities in adult left-hemisphere-dysphasic and right-hemisphere-damaged subjects. Brain Lang 75:47–65.[CrossRef][ISI][Medline]
Anderson B, Southern BD, Powers RE. 1999. Anatomic asymmetries of the posterior superior temporal lobes: a postmortem study. Neuropsychiatry Neuropsychol Behav Neurol 12:247–254.
Belin P, Zatorre RJ, Ahad P. 2002. Human temporal-lobe response to vocal sounds. Brain Res Cogn Brain Res 13:17–26.[CrossRef][Medline]
Belin P, Zatorre RJ, Hoge R, Evans AC, Pike B. 1999. Event-related fmri of the auditory cortex. Neuroimage 10:417–429.[CrossRef][ISI][Medline]
Belin P, Zatorre RJ, Lafaille P, Ahad P, Pike B. 2000. Voice-selective areas in human auditory cortex. Nature 403:309–312.[CrossRef][Medline]
Belin P, Zilbovicius M, Crozier S, Thivard L, Fontaine A, Masure MC, Samson Y. 1998. Lateralization of speech and auditory temporal processing. J Cogn Neurosci 10:536–540.[Abstract]
Bever TG, Chiarello RJ. 1974. Cerebral dominance in musicians and nonmusicians. Science 185:537–539.
Boemio A, Fromm S, Braun A, Poeppel D. 2005. Hierarchical and asymmetric temporal sensitivity in human auditory cortex. Nat Neurosci 8:389–395.[CrossRef][ISI][Medline]
Bregman AS. 1978. Auditory streaming is cumulative. J Exp Psychol Hum Percept Perform 4:380–387.[CrossRef][ISI][Medline]
Buxhoeveden D, Casanova M. 2000. Comparative lateralisation patterns in the language area of human, chimpanzee, and rhesus monkey brains. Laterality 5:315–330.[CrossRef][ISI][Medline]
Efron R. 1963. Temporal perception, aphasia and d‘ej’a vu. Brain 86:403–424.
Gandour J, Tong Y, Wong D, Talavage T, Dzemidzic M, Xu Y, Li X, Lowe M. 2004. Hemispheric roles in the perception of speech prosody. Neuroimage 23:344–357.[CrossRef][ISI][Medline]
Gandour J, Wong D, Dzemidzic M, Lowe M, Tong Y, Li X. 2003. A cross-linguistic fmri study of perception of intonation and emotion in Chinese. Hum Brain Mapp 18:149–157.[CrossRef][ISI][Medline]
Gandour J, Wong D, Hsieh L, Weinzapfel B, Van Lancker D, Hutchins GD. 2000. A crosslinguistic pet study of tone perception. J Cogn Neurosci 12:207–222.
Gandour J, Wong D, Lowe M, Dzemidzic M, Satthamnuwong N, Tong Y, Li X. 2002. A cross-linguistic fmri study of spectral and temporal cues underlying phonological processing. J Cogn Neurosci 14:1076–1087.
Glover GH. 1999. Deconvolution of impulse response in event-related bold fmri. Neuroimage 9:416–429.[CrossRef][ISI][Medline]
Griffiths TD, Buchel C, Frackowiak RS, Patterson RD. 1998. Analysis of temporal structure in sound by the human brain. Nat Neurosci 1:422–427.[CrossRef][ISI][Medline]
Hall DA, Haggard MP, Akeroyd MA, Palmer AR, Summerfield AQ, Elliott MR, Gurney EM, Bowtell RW. 1999. "Sparse" temporal sampling in auditory fmri. Hum Brain Mapp 7:213–223.[CrossRef][ISI][Medline]
Hall DA, Johnsrude I, Haggard MP, Palmer AR, Akeroyd MA, Summerfield AQ. 2002. Spectral and temporal processing in human auditory cortex. Cereb Cortex 12:140–149.
Harms MP, Guinan JJJ, Sigalovsky IS, Melcher JR. 2005. Short-term sound temporal envelope characteristics determine multisecond time patterns of activity in human auditory cortex as shown by fmri. J Neurophysiol 93:210–222.
Hickok G, Love T, Swinney D, Way EC, Buxton RB. 1997. Functional mr imaging during auditory word perception: a single-trial presentation paradigm. Brain Lang 58:197–201.[CrossRef][ISI][Medline]
Hsieh L, Gandour J, Wong D, Hutchins GD. 2001. Functional heterogeneity of inferior frontal gyrus is shaped by linguistic experience. Brain Lang 76:227–252.[CrossRef][ISI][Medline]
Hutsler J, Galuske RA. 2003. Hemispheric asymmetries in cerebral cortical networks. Trends Neurosci 26:429–435.[CrossRef][ISI][Medline]
Hyde KS, Peretz I. 2004. Brains that are out of tune but in time. Psychol Sci 15:356–360.[CrossRef][ISI][Medline]
Jenkinson M, Bannister P, Brady M, Smith S. 2002. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 17:825–841.[CrossRef][ISI][Medline]
Johnsrude IS, Penhune VB, Zatorre RJ. 2000. Functional specificity in the right human auditory cortex for perceiving pitch direction. Brain 123:155–163.
Khalfa S, Micheyl C, Veuillet E, Collet L. 1998. Peripheral auditory lateralization assessment using teoaes. Hear Res 121:29–34.[CrossRef][ISI][Medline]
Lattner S, Meyer ME, Friederici AD. 2005. Voice perception: sex, pitch, and the right hemisphere. Hum Brain Mapp 24:11–20.[CrossRef][ISI][Medline]
Liegeois-Chauvel C, de Graaf JB, Laguitton V, Chauvel P. 1999. Specialization of left auditory cortex for speech perception in man depends on temporal coding. Cereb Cortex 9:484–496.
Mills L, Rollman GB. 1980. Hemispheric asymmetry for auditory perception of temporal order. Neuropsychologia 18:41–48.[CrossRef][ISI][Medline]
Morlet T, Goforth L, Hood LJ, Ferber C, Ducleaux R, Berlin CI. 1999. Development of human cochlear active mechanism asymmetry: involvement of the medial olivocochlear system? Hear Res 134:153–162.[CrossRef][ISI][Medline]
Murayama J, Kashiwaga T, Kashiwaga A, Mimura M. 2004. Impaired pitch production and preserved rhythm production in a right brain-damaged patient with amusia. Brain Cogn 56:36–42.[CrossRef][ISI][Medline]
Narain C, Scott SK, Wise RJ, Rosen S, Leff A, Iversen SD, Matthews PM. 2003. Defining a left-lateralized response specific to intelligible speech using fmri. Cereb Cortex 13:1362–1368.
Nicholls ME, Schier M, Stough CK, Box A. 1999. Psychophysical and electrophysiologic support for a left hemisphere temporal processing advantage. Neuropsychiatry Neuropsychol Behav Neurol 12:11–16.[ISI][Medline]
Ohnishi T, Matsuda H, Asada T, Aruga M, Hirakata M, Nishikawa M, Katoh A, Imabayashi E. 2001. Functional anatomy of musical perception in musicians. Cereb Cortex 11:754–760.
Oldfield RC. 1971. The assessment and analysis of handedness: the Edinburgh Inventory. Neuropsychologia 9:97–113.[CrossRef][ISI][Medline]
Palmer AR, Bullock DC, Chambers JD. 1998. A high-output, high quality sound system for use in auditory fmri. Neuroimage 7:S359.
Patterson RD, Uppenkamp S, Johnsrude IS, Griffiths TD. 2002. The processing of temporal pitch and melody information in auditory cortex. Neuron 36:767–776.[CrossRef][ISI][Medline]
Penagos H, Melcher JR, Oxenham AJ. 2004. A neural representation of pitch salience in nonprimary auditory cortex revealed with functional magnetic resonance imaging. J Neurosci 24:6810–6815.
Penhune VB, Zatorre RJ, MacDonald JD, Evans AC. 1996. Interhemispheric anatomical differences in human primary auditory cortex: probabilistic mapping and volume measurement from magnetic resonance scans. Cereb Cortex 6:661–672.
Poeppel D. 2003. The analysis of speech in different temporal integration windows: cerebral lateralization as ‘asymmetric sampling in time’. Speech Commun 41:245–255.
Rademacher J, Caviness VSJ, Steinmetz H, Galaburda AM. 1993. Topographical variation of the human primary cortices: implications for neuroimaging, brain mapping, and neurobiology. Cereb Cortex 3:313–329.
Rivier F, Clarke S. 1997. Cytochrome oxidase, acetylcholinesterase, and nadph-diaphorase staining in human supratemporal and insular cortex: evidence for multiple auditory areas. Neuroimage 6:288–304.[CrossRef][ISI][Medline]
Robin DA, Tranel D, Damasio H. 1990. Auditory perception of temporal and spectral events in patients with focal left and right cerebral lesions. Brain Lang 39:539–555.[CrossRef][ISI][Medline]
Rose MM, Moore BCJ. 2000. Effects of frequency and level on auditory stream segregation. J Acoust Soc Am 108:1209–1214.[CrossRef][ISI][Medline]
Rutten GJM, Ramsey NF, van Rijen PC, van Veelen CWM. 2002. Reproducibility of fmri-determined language lateralization in individual subjects. Brain Lang 80:421–437.[CrossRef][ISI][Medline]
Schwartz J, Tallal P. 1980. Rate of acoustic change may underlie hemispheric specialisation for speech perception. Science 207:1380–1381.
Scott SK, Blank CC, Rosen S, Wise RJ. 2000. Identification of a pathway for intelligible speech in the left temporal lobe. Brain 123:2400–2406.
Scott SK, Wise RJ. 2004. The functional neuroanatomy of prelexical processing in speech perception. Cognition 92:13–45.[CrossRef][ISI][Medline]
Seifritz E, Esposito F, Hennel F, Mustovic H, Neuhoff JG, Bilecen D, Tedeschi G, Scheffler K, Di Salle F. 2002. Spatiotemporal pattern of neural processing in the human auditory cortex. Science 297:1706–1708.
Shtyrov Y, Kujala T, Palva S, Ilmoniemi RJ, Naatanen R. 2000. Discrimination of speech and of complex nonspeech sounds of different temporal structure in the left and right cerebral hemispheres. Neuroimage 12:657–663.[CrossRef][ISI][Medline]
Sidtis JJ, Volpe BT. 1988. Selective loss of complex-pitch or speech discrimination after unilateral lesion. Brain Lang 34:235–245.[CrossRef][ISI][Medline]
Sininger YS, Cone-Wesson B. 2004. Asymmetric cochlear processing mimics hemispheric specialization. Science 305:1581.
Takegata R, Nakagawa S, Tonoike M, Naatanen R. 2004. Hemispheric processing of duration changes in speech and non-speech sounds. Neuroreport 15:1683–1686.[CrossRef][ISI][Medline]
Talavage TM, Edmister WB. 2004. Nonlinearity of fmri responses in human auditory cortex. Hum Brain Mapp 22:216–228.[CrossRef][ISI][Medline]
Tallal P, Piercy M. 1973. Defects of non-verbal auditory perception in children with developmental aphasia. Nature 241:468–469.[CrossRef][Medline]
Tallal P, Piercy M. 1974. Developmental aphasia: rate of auditory processing and selective impairment of consonant perception. Neuropsychologia 12:83–93.[CrossRef][ISI][Medline]
Tallal P, Piercy M. 1975. Developmental aphasia: the perception of brief vowels and extended stop consonants. Neuropsychologia 13:69–74.[CrossRef][ISI][Medline]
Tervaniemi M, Hugdahl K. 2003. Lateralization of auditory-cortex functions. Brain Res Brain Res Rev 43:231–246.[CrossRef][Medline]
Thivard L, Belin P, Zilbovicius M, Poline JB, Samson Y. 2000. A cortical region sensitive to auditory spectral motion. Neuroreport 11:2969–2972.[ISI][Medline]
van Noorden LPAS. 1975. Temporal coherence in the perception of tone sequences [thesis]. Eindhoven, The Netherlands: Eindhoven University of Technology.
Wallace MN, Johnston PW, Palmer AR. 2002. Histochemical identification of cortical areas in the auditory regions of the human brain. Exp Brain Res 143:499–508.[CrossRef][ISI][Medline]
Warren JD, Jennings AR, Griffiths TD. 2005. Analysis of the spectral envelope of sounds by the human brain. Neuroimage 24:1052–1057.[CrossRef][ISI][Medlin