Task-dependent activations of human auditory cortex during spatial discrimination and spatial memory tasks

NeuroImage 59 (2012) 4126–4131

Contents lists available at SciVerse ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Task-dependent activations of human auditory cortex during spatial discrimination and spatial memory tasks Teemu Rinne a, c,⁎, Sonja Koistinen a, Suvi Talja a, Patrik Wikman a, Oili Salonen b a b c

Institute of Behavioural Sciences, University of Helsinki, Finland Helsinki Medical Imaging Center, Helsinki University Central Hospital, Finland Advanced Magnetic Imaging Centre, Aalto University School of Science, Finland

a r t i c l e

i n f o

Article history: Received 27 June 2011 Revised 10 October 2011 Accepted 18 October 2011 Available online 29 October 2011 Keywords: Auditory cortex Spatial processing Attention Functional magnetic resonance imaging Humans

a b s t r a c t In the present study, we applied high-resolution functional magnetic resonance imaging (fMRI) of the human auditory cortex (AC) and adjacent areas to compare activations during spatial discrimination and spatial n-back memory tasks that were varied parametrically in difficulty. We found that activations in the anterior superior temporal gyrus (STG) were stronger during spatial discrimination than during spatial memory, while spatial memory was associated with stronger activations in the inferior parietal lobule (IPL). We also found that wide AC areas were strongly deactivated during the spatial memory tasks. The present AC activation patterns associated with spatial discrimination and spatial memory tasks were highly similar to those obtained in our previous study comparing AC activations during pitch discrimination and pitch memory (Rinne et al., 2009). Together our previous and present results indicate that discrimination and memory tasks activate anterior and posterior AC areas differently and that this anterior–posterior division is present both when these tasks are performed on spatially invariant (pitch discrimination vs. memory) or spatially varying (spatial discrimination vs. memory) sounds. These results also further strengthen the view that activations of human AC cannot be explained only by stimulus-level parameters (e.g., spatial vs. nonspatial stimuli) but that the activations observed with fMRI are strongly dependent on the characteristics of the behavioral task. Thus, our results suggest that in order to understand the functional structure of AC a more systematic investigation of task-related factors affecting AC activations is needed. © 2011 Elsevier Inc. All rights reserved.

Introduction In the prevailing auditory models, human auditory cortex (AC) consists of anatomically and functionally separate fields organized in two (or more) parallel processing streams (McLachlan and Wilson, 2010; Rauschecker, 2011; Recanzone and Cohen, 2010). While the anatomical and functional details may differ from one model to another, it is quite generally accepted that spatial and nonspatial auditory information is processed in separate streams. Several previous human fMRI studies have compared AC activations during spatial and nonspatial conditions and typically report activations associated with spatial processing in the posterior superior temporal gyrus (STG) or inferior parietal lobule (IPL), while activations in the anterior STG and Heschl's gyrus (HG) are associated with nonspatial (e.g., pitch) processing (Arnott et al., 2004; Barrett and Hall, 2006; Warren and Griffiths, 2003). However, interpretation of the activation differences observed during spatial and nonspatial conditions

⁎ Corresponding author at: Institute of Behavioural Sciences, PO Box 9, FI-00014 University of Helsinki, Finland. E-mail address: teemu.rinne@helsinki.fi (T. Rinne). 1053-8119/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2011.10.069

is not necessarily straightforward. Activations of the posterior STG during spatial conditions could be due to, for example, segregation of auditory sources (Smith et al., 2010; Zatorre et al., 2002) and not due to processing of spatial information as such. Further, as previous fMRI studies have shown that attention-engaging tasks strongly modulate AC activations (e.g., Degerman et al., 2006; Hall et al., 2000; Mayer et al., 2009; Petkov et al., 2004; Rinne, 2010; Rinne et al., 2005, 2009; Woodruff et al., 1996) and as the effects of different tasks on activations to spatial and nonspatial sounds have not been systematically studied, it is possible that the differences observed in previous studies between spatial and nonspatial conditions are affected by (uncontrolled) taskrelated factors. For example, it is possible that a comparison of two different spatial tasks or two different nonspatial tasks would show an anterior–posterior distribution of AC activations similar to those reported by previous studies comparing spatial and nonspatial tasks. Indeed, the results of our recent study demonstrated distinct functional differences between anterior and posterior AC during two pitch tasks performed on sounds with a fixed spatial location (Rinne et al., 2009). In that study, we compared AC activations during pitch discrimination and n-back pitch memory tasks that were varied parametrically in difficulty. During n-back memory tasks, subjects were required to indicate when a sound belonged to the same pitch category

T. Rinne et al. / NeuroImage 59 (2012) 4126–4131

4127

(low, medium or high) as the one presented 1–3 trials (depending on the difficulty level) before. In pitch-discrimination tasks, the sounds were otherwise similar but the last half of each sound was slightly (depending on the difficulty level) lower or higher in pitch than the first half. Subjects were required to indicate when the two halves of a sound had the same pitch. Comparison of pitch discrimination and pitch memory tasks showed increased anterior AC activations during pitch discrimination, while posterior AC activations (IPL) were enhanced during pitch memory tasks. These results indicate that the anterior–posterior distinction is present also between two pitch tasks using nonspatial stimuli. In the present study, we compared AC activations during spatial discrimination, spatial memory and visual discrimination tasks. Our experimental setup was similar to that used in our previous study introduced above. In the spatial discrimination task, with three difficulty levels, subjects were required to indicate when two halves of a sound (amplitude modulated random noise) had the same spatial location (Fig. 1a). In the spatial n-back memory tasks, subjects indicated when a sound belonged to the same spatial category (left, middle, or right) as the one presented 1–3 trials before (b). Further,

we presented the spatially varying sounds used in the auditory tasks and similar spatially fixed sounds during a demanding visual task. The task and its difficulty level were indicated by visual non-verbal task-instruction symbols (c–h). Based on the results of our previous study using pitch tasks, we assumed that these two auditory spatial tasks performed on spatially-varying noise bursts (with no clear pitch) would be associated with different anterior-posterior activation patterns. We also assumed that a comparison of activations to spatially varying and spatially fixed sounds presented during the visual task would reveal AC activations associated with processing of spatial stimulus information.

a

Stimuli and tasks Spatial discrimination task

Subjects Subjects (N = 17, 10 women, all right-handed) were 21–48 years (mean 25 years) of age. All subjects had normal hearing, normal or corrected-to-normal vision, and no history of psychiatric or neurological illnesses. An informed written consent was obtained from each subject prior to the experiment. The study protocol was approved by the Ethical Committee of the Hospital District of Helsinki and Uusimaa, Finland.

Spatial discrimination tasks

c d e

b

Materials and methods

Spatial memory task (2-back)

Spatial memory tasks

f g h

Fig. 1. (a, b) Spatially varying sounds were presented during auditory and visual task blocks. In addition, there were visual task blocks with spatially fixed sounds (not shown). The spatially varying sounds were 200 ms in duration and consisted of two successive 90-ms parts (each part included a 5 ms linear onset and offset ramps) separated by a 20-ms gap. In spatial discrimination tasks (a), the first and last half of each sound were either presented from the same location or were separated by 30°, 45° or 60° depending on the difficulty level. Subjects were required to press a button when the two halves of a sound had the same location (target). In spatial memory tasks (b), the two halves of a sound were separated by 0° or 15° and subjects were required to respond when a sound belonged to the same spatial category (left, middle or right) as the one presented 1, 2 or 3 trials before (target in a 2-back task is illustrated). The task and its difficulty level were indicated by task instruction symbols presented on a screen from 7 s before each block onset until the end of the block (c–h). Nonverbal symbols were used to minimize activation differences associated with processing of the task instruction. A letter “Λ” (Lambda) or “V” (not shown in the figure) in the middle of task instruction symbols indicated the task modality, auditory or visual, respectively. Spatial discrimination tasks were indicated by one red dot (c–e), while spatial memory tasks were indicated by two red dots (f–h). In spatial discrimination tasks, task difficulty level was indicated by the position (yellow rectangles) of the red dot (the leftmost position=easy, second position from the left=medium, rightmost position=hard). For the spatial memory tasks, the distance between two red dots indicated the relative serial positions of the sounds to be compared. For the visual tasks, in addition to the letter “V", two red dots were presented at second and third position from the left.

Subjects were presented with amplitude modulated random noise bursts (sinusoidal modulation at 33.33 Hz with modulation depth of 100%, diotic presentation, duration 200 ms, onset-to-onset interval 0.8–1 s) consisting of two successive 90-ms parts (each part included a 5 ms linear onset and offset ramps) separated by a 20-ms gap. The noise bursts were convolved with a generic head-related transfer function to create 17 equidistant virtual spatial locations between ±120° (step 15°) in azimuth. In spatial discrimination tasks (Fig. 1a; locations between ±90°), the first half of each sound occurred randomly in one location and the second half was either presented from the same or different location separated from the location of the first half by 30°, 45° or 60° (depending on the difficulty level). Subjects’ task was to indicate by pressing a button with their right hand when the two halves of a sound were presented from the same location. In spatial (n-back) memory tasks (Fig. 1b), the sounds were presented randomly from three spatial categories: Left (75°–120° to the left of midline), Right (75°–120° to the right of midline), or Middle (locations between ± 15° of midline). There was a 0° or 15° location difference between the first and last part of the sounds. Subjects’ were required to indicate when a sound belonged to the same spatial category (left, middle, or right) as the one presented 1–3 trials before (depending on the difficulty level). There were 2–4 targets in each block. The sounds were delivered with an UNIDES ADU2a audio system (Unides Design, Helsinki, Finland) via plastic tubes through a porous EAR tip (ER3, Etymotic Research) acting as an earphone. The noise of the scanner (~102 dB SPL, A-weighted measurement inside the head coil, first harmonic at about 1 kHz) was attenuated by the earplugs, circumaural ear protectors, and viscoelastic mattress inside and around the headcoil and under the subject. The experiment was controlled using Presentation software (Neurobehavioral Systems, Albany, CA). The sounds were presented in 15.5-s blocks alternating with 9-s breaks with no stimuli. During the breaks, the subjects focused on a fixation mark (white X on gray background, Red 190, Green 190, Blue 190) presented for 2 s in the middle of a screen (viewed through a mirror fixed to the head coil). The fixation mark was replaced by task-instruction symbols (Figs. 1c–h) 7 s before the start of the next block. The task-instruction symbols were presented until the end of the block. The non-verbal symbols indicated the task modality

4128

T. Rinne et al. / NeuroImage 59 (2012) 4126–4131

(auditory/visual), auditory task type (discrimination/memory) and task difficulty level (for details, see legend of Fig. 1). In the visual task, the subjects were instructed to ignore the sounds and to detect occasional slight luminance changes of a flickering gray rectangle (R 186, G 186, B 186) underlying the task-instruction symbols (Figs. 1c–h). The target rectangle was slightly brighter (R 194, G 194, B 194). Otherwise the visual stimulus sequences (stimulus duration 200 ms, onset-to-onset interval 0.8–1 s, 2–4 targets in each block) were similar to the auditory sequences. Sound sequences presented during the visual tasks were identical to the ones presented during spatial discrimination or spatial memory tasks. In addition, at the end of each fMRI run, there were 8 visual task blocks with monophonic spatially fixed sounds. These sounds were created by averaging the left and right ear channels of the sounds used in the spatial discrimination and memory tasks and playing this average signal to both ears. This averaging resulted in monophonic sounds with a fixed (middle of the head) location but with slightly varying source characteristics. For each task type and difficulty level (2 auditory tasks with 3 difficulty levels, visual task with spatially varying sounds of the auditory tasks, and visual task with spatially fixed sounds), 16 blocks were presented resulting in 128 blocks altogether. Pre-fMRI training In order to reveal differences in brain activity between easy and hard tasks and between spatial discrimination and spatial memory tasks, the hardest difficulty levels were made intentionally highly demanding. Before the fMRI session, subjects were carefully trained to perform the tasks (1–2 h of training in 2 sessions 1–5 days before scanning). During training, it was emphasized to the subjects that maximum effort in performance is essential especially during the most difficult levels. Analysis of the behavioral data Mean hit rates and reaction times were calculated separately for each task. Responses occurring 200–1300 ms relative to target onset (in the discrimination task, onset of the second part of the target pair) were accepted as hits. Hit rate (HR) was defined as the number of hits divided by the number of targets. Other responses were considered as false alarms. False alarm rate (FaR) was defined as the number of false alarms divided by the number of non-targets. HRs and FaRs were used to compute the d’ (index of stimulus detectability, d′ = z(HR) − z(FaR)). Mean reaction time was calculated only for hits. Behavioral results were analyzed using one-way repeated measures ANOVAs with factor Task Difficulty (3 levels). fMRI data acquisition and analysis fMRI data was acquired with a 3.0T GE Signa system retrofitted with an Advanced NMR operating console and an 8-channel head coil. Functional images were acquired using a T2*­weighted gradient­echo echo­planar (GE­EPI) sequence (TR 2000 ms, TE 32 ms, flip angle 90°, voxel matrix 96 × 96, FOV 20 cm, slice thickness 2.1 mm with no gap, in-plane resolution 2.1 mm × 2.1 mm, number of slices 24). Based on an anatomical scout image (sagittal slices, slice thickness 3 mm, in-plane resolution 0.94 mm× 0.94 mm), the middle EPI slices were aligned along Sylvian fissures (Fig. 2). The functional scanning was divided in two 25 min runs resulting in approximately 2 × 763 images. Between the runs, there was a short break during which subjects remained in the scanner and were instructed not to move their heads or speak. After the functional scans, a fluid-attenuated inversion recovery image using the same imaging slices but with denser in-plane resolution was acquired (FLAIR; TR 10000 ms, TE

IPL IPL HG IFG

IPL

ula

Ins

STG

Fig. 2. Inflated left-hemisphere cortical surface (light gray, gyri; dark gray, sulci). The areas imaged in the present study are illustrated in lighter grayscale. The EPI slices were aligned along Sylvian fissures to cover the STG, HG, anterior insula, and most of the IPL of both hemispheres. The imaged area did not completely cover the IPL, inferior frontal gyrus (IFG) and superior temporal sulcus in all cases.

120 ms, voxel matrix 320 × 192, slice thickness 2.1 mm, in­plane resolution 0.39 mm× 0.39 mm). Finally, at the end of the session, high-resolution anatomical images were acquired (voxel matrix 176 × 256 × 256, resolution 1 mm × 1 mm × 1 mm). Global voxel-wise analysis was performed using the tools developed by the Analysis Group at the Oxford Centre for Functional MRI of the Brain (FMRIB) and implemented within FMRIB's software library (FSL, release 4.1.3, www.fmrib.ox.ac.uk/fsl). First, data from the two runs were combined into one file for motion correction. The motion-corrected data was again split into two separate files, highpass filtered (cutoff 50 s), and spatially smoothed (Gaussian kernel of 7 mm full-width half-maximum). First-level statistical analysis was carried out using FMRIB's improved linear model. Based on the timing information recorded during the experiment, each functional image was labeled as either spatial discrimination (3 difficulty levels), spatial memory (3 levels), visual task with spatially varying sounds of the discrimination task, visual task with spatially varying sounds of the memory task, visual task with spatially fixed sounds, or baseline (9-s breaks with no sound stimuli). The hemodynamic response function was modeled with a gamma function (mean lag 6 s, SD 3 s) and its temporal derivative. Contrasts were specified to create Z-statistic images testing for task and difficulty effects. A second-level statistical analysis using fixed-effects combined the data from the two runs. For analysis across participants (third level analysis), the data were anatomically normalized in the following steps: First, cortical surfaces were extracted from high-resolution anatomical images, transformed to spherical standard space, and anatomically normalized on the basis of the cortical gyral and sulcal patterns using FreeSurfer (release 5.0.0, http://surfer.nmr.mgh.harvard.edu). Next, the three-dimensional (3D) spherical cortical surfaces were rotated and projected to a two-dimensional (2D) space separately for each hemisphere using equal area Mollweide projection (Python libraries matplotlib and basemap, http://matplotlib.sourceforge.net). This procedure produced 3D-to-2D anatomical transformation matrices for each subject that were then applied separately for each subject to transform the results of the 3D second-level statistical analysis to 2D. Finally, the group analysis (FMRIB's local analysis of mixed effects, N = 17) was run on these flattened data. Z-statistic images were thresholded using clusters determined by Z > 2.3 and a (corrected) cluster significance threshold of P b 0.05 (using Gaussian random field theory).

T. Rinne et al. / NeuroImage 59 (2012) 4126–4131

4129

Results Behavior In spatial discrimination (d′ values in easy, medium and hard tasks: 1.77, 2.15, 1.56), spatial memory (1-back, 2-back and 3-back: 3.80, 2.21, 1.56), and visual tasks (spatially varying sounds and spatially fixed sounds: 3.94, 4.02) subjects successfully performed the demanding tasks during the fMRI. Note that in the spatial discrimination task, even the easiest condition proved to be quite demanding in the scanner. In the spatial discrimination and spatial memory tasks, mean hit rates (HRs) were 54 and 62% and mean reaction times 686 ms and 727 ms, respectively. Mean HRs in easy, medium and hard spatial discrimination tasks were 60, 60 and 41%, respectively, and in 1-back, 2-back and 3-back spatial memory tasks 92, 58 and 36%, respectively. HRs decreased with increasing task difficulty both in the spatial discrimination (F(2,32) = 31, P b 0.0001, linear trend, F(1,16) = 32, P b 0.0001) and spatial memory tasks (F(2,32) = 168, P b 0.0001, linear trend, F(1,16) = 324, P b 0.0001). Mean RTs in easy, medium and hard spatial discrimination tasks were 683, 687 and 689 ms, respectively, and in 1-back, 2-back and 3-back spatial memory tasks 692, 743 and 745 ms, respectively. In the spatial memory task, RTs increased with increasing difficulty (F(2,32) = 5.5, P b 0.01, linear trend, F(1,16) = 8.4, P b 0.05). Voxel-wise analysis of fMRI data

Fig. 3. Activations (a, f) shown on flattened mean 2D cortical surface (N = 17, threshold Z > 2.3, cluster-corrected P b 0.05 unless otherwise specified). (a) Activations to sounds in the absence of auditory attention (blue) were isolated by contrasting activations during the visual task (spatially varying sounds) with activations during the 9-s breaks with no sounds. General effects of auditory tasks were isolated by contrasting all auditory tasks with the visual task (red). Areas showing significant activations in both contrasts are shown in yellow. (b) Activations specific to spatial discrimination (blue) and spatial memory (red) tasks were extracted by comparing each auditory task (all difficulty levels) with activations during the visual task. (c) Areas where activations were stronger during spatial discrimination than spatial memory tasks (blue) and areas where activations were stronger during spatial memory than spatial discrimination tasks (red). (d) Effects of task difficulty (linear contrasts) on spatial discrimination (blue and yellow, threshold Z > 1.96, nonsignificant) and spatial memory (red and yellow, Z > 2.3, cluster-corrected P b 0.05) tasks. (e) Results of linear inverse contrast revealing areas where activations decreased with increasing task difficulty during spatial discrimination (blue) and spatial memory (red). (f) Comparison of activations to spatially varying and fixed sounds during visual task (threshold Z > 1.96, nonsignificant). (g) Anatomical labels. STG superior temporal gyrus, HG Heschl's gyrus, IPL inferior parietal lobule (consisting of angular gyrus and supramarginal gyrus). Note that although anterior–posterior direction is indicated, the interpretation of the bottomtop axis is less straightforward due to cortical curvature (for HG, lateral is down and medial is up).

AC activations to sounds in the absence of auditory attention were isolated by contrasting activations during all visual task blocks with spatially varying sounds with activations during the 9-s breaks with no sounds. In both hemispheres, widespread AC regions in the anterior and posterior STG were activated by the task-irrelevant sounds (Fig. 3a, blue and yellow; for anatomical labels, see Figs. 2 and 3g). The spatially varying sounds presented during the visual task were either similar to sounds used in the spatial discrimination (locations ± 90°, within-pair difference 0°–60°) or in the spatial memory tasks (locations ± 120°, within-pair difference 0° or 15°). As these conditions were not associated with systematic AC activation differences, all visual task blocks with spatially varying sounds were collapsed together in the analyses. General effects of auditory tasks were isolated by contrasting all auditory tasks with the visual tasks with spatially varying sounds. Distinct activation clusters associated with the auditory tasks were detected in the posterior STG, IPL and anterior insula in both hemispheres (Fig. 3a, red and yellow). Activations specific to the two auditory tasks were extracted by comparing each auditory task (all difficulty levels) with activations during the visual task with spatially varying sounds. Enhanced activations during spatial discrimination were detected bilaterally in the posterior STG and anterior insula and in the left anterior STG (Fig. 3b, blue and yellow). Activation increases during spatial memory tasks, in turn, were found bilaterally in more posterior parts of STG as well as in the IPL, and anterior insula (Fig. 3b, red and yellow). Direct comparison of the two auditory tasks with each other revealed stronger activations during spatial discrimination than during spatial memory in the anterior insula, anterior STG, HG, and posterior STG (Fig. 3c, blue), while areas in the IPL and anterior insula were more activated during the spatial memory tasks than during spatial discrimination (Fig. 3c, red). Effects of auditory task difficulty were examined with linear contrasts. Significant activation increases with increasing task difficulty were detected in the anterior insula and IPL during the spatial memory tasks (Fig. 3d, red and yellow). During spatial discrimination, in turn, significant activations (not shown) were detected only in the right insula. However, with a more lenient threshold (Z > 1.96, nonsignificant) distinct clusters showing higher activity

4130

T. Rinne et al. / NeuroImage 59 (2012) 4126–4131

for more difficult discrimination tasks (Fig. 3d, blue and yellow) were seen within the right hemisphere areas (in the anterior insula and IPL) where activations were also modulated by memory load. (Note that in Fig. 3d different thresholds are used for the spatial discrimination and spatial memory tasks.) Inverse linear contrasts showed that activations decreased with increasing memory load in areas extending from the anterior insula, anterior STG, and HG to posterior STG and IPL (Fig. 3e, red). The corresponding inverse linear contrast for spatial discrimination revealed no significant effects. Contrast of activations to spatially varying vs. spatially fixed sounds during visual task revealed no significant AC activation increases (Fig. 3f). Even with a more lenient threshold (Z> 1.96, nonsignificant) no systematic activation increases to spatially varying sounds were detected in the STG areas posterior to HG that have been associated with spatial processing in previous studies. Discussion In the present study, spatial discrimination and spatial memory tasks were associated with distinct AC activation patterns (Figs. 3b–e). Activations in the anterior STG were stronger during spatial discrimination than during spatial memory, while spatial memory was associated with stronger IPL activations (c). Further, activations in the IPL and anterior insula increased with increasing memory load (d). We also found that wide areas of the anterior insula, anterior STG, HG, and posterior STG were strongly deactivated during the spatial memory tasks. In these areas, activations decreased with an increasing memory load (e). These AC activation patterns associated with spatial discrimination and spatial memory tasks are highly similar to those obtained in our previous study comparing AC activations during pitch discrimination and pitch memory (Rinne et al., 2009). In both studies, the discrimination tasks were associated with increased activations in anterior AC, the memory tasks were associated with increased IPL activations, and activations of anterior AC including HG decreased with increasing memory load. Note that in the present study, the spatial discrimination and spatial memory tasks were performed on spatially varying noise bursts with no clear pitch and, in our previous study, the pitch discrimination and pitch memory tasks were performed on pitch varying sounds with a fixed spatial location. Thus, it is evident that the anterior-posterior activation patterns obtained in these two studies are not simply due to differences between spatial and nonspatial processing. Rather, these results show that activations in anterior and posterior AC depend in a different manner on the characteristics of the behavioral task. Although the present study did not focus on a direct comparison of spatial and nonspatial conditions, we also measured activations to both spatially varying and spatially fixed sounds presented in separate blocks during the visual task. In contrast with the results of several previous fMRI studies (e.g., Barrett and Hall, 2006; Krumbholz et al., 2005; Warren and Griffiths, 2003), no AC areas showed significantly stronger activations to spatially varying than to spatially fixed sounds. Notably, even with a more lenient threshold (Z > 1.96, nonsignificant; Fig. 3f), we found no systematic activation increases to spatially varying sounds in the posterior STG (‘where’ stream). In the present study, the spatially varying and fixed sounds were presented during an attention-demanding visual detection task, while previous studies have often used passive listening or easy auditory tasks. In such conditions, attention to sounds may depend on the stimuli (e.g., spatially varying sounds may be more interesting, distracting or attention-catching than spatially invariant sounds) and, further, the way sounds are processed in AC may depend on attention (Sussman et al., 2002). Thus, when the behavioral task is not carefully controlled, it is possible that the auditory system may actually be engaged in a different task or processing mode

during spatial and nonspatial stimulation. To our knowledge, only one previous fMRI study has investigated AC activations to spatially varying and fixed sounds during a visual task that was deliberately designed to maintain attention away from the sounds. In contrast to the present results, Deouell et al. (2007) detected stronger activations in the posterior STG (planum temporale) to spatially varying (five equiprobable locations) sounds as compared with spatially fixed (one location) sounds presented during a visual task. They used sparse imaging to eliminate scanner noise during sound presentation. It is possible that the present study failed to detect significant differences between activations to spatially varying and fixed sounds because we used continuous imaging rather than sparse imaging. First, activations of the auditory system to the continuous scanner noise could have concealed any differences between AC activations to spatially varying and fixed sounds. This, however, is unlikely as we were able to detect significant taskdependent effects. Second, it is possible that the scanner noise acted as an acoustic masker making the spatially varying and fixed sounds (noise bursts) less distinguishable from the background and, thus, less obtrusive during the visual task. In the present study, this could have made it easier to focus on the visual task and to ignore the spatially varying and fixed sounds, while, in the study by Deouell et al. (2007), the lack of scanner noise during presentation of sounds could have made it harder to completely ignore the sounds. Subsequent studies should carefully investigate the role of behavioral tasks and attention on AC activations to spatially varying sounds. In the present study, activations in the STG areas posterior to HG were enhanced both during spatial discrimination and spatial memory as compared with visual task with similar sounds (Fig. 3b). However, STG activations immediately posterior to HG (planum temporale, PT) were more associated with the discrimination task than with the memory task (b, c). It could be argued that this was because the discrimination task required more spatial processing as both parts of the sounds had to be processed in detail before it was possible to judge whether they were presented in the same or different location, while in the memory task the analysis could be stopped as soon as the spatial category was obtained. Thus, the present results seem to suggest that while the STG areas immediately posterior to HG and areas closer to IPL are both involved in spatial processing, these areas could have different roles in spatial tasks. However, in our previous study (Rinne et al., 2009) we found exactly the same pattern of activations during pitch discrimination and pitch memory tasks performed on spatially fixed sounds. In that study, areas of the posterior STG were activated by both pitch discrimination and pitch memory, but the areas immediately posterior to HG were more associated with pitch discrimination. Together these results are in line with the suggestion that areas of the posterior STG act as a more general computational resource involved in segregation of auditory sources and do not support spatial processing per se (Griffiths and Warren, 2002; Smith et al., 2010). In our previous study comparing pitch discrimination and pitch memory tasks, we found that activations in areas of IPL and anterior insula were strongly associated with the pitch memory tasks (Rinne et al., 2009). In these areas activations increased with increasing memory load. Further, in these areas activations also tended to increase with increasing task difficulty in the pitch discrimination task (this effect was only seen with a nonsignificant threshold). Interestingly, in the present study, we saw the same pattern: Activations in IPL and anterior insula increased with increasing memory load in the spatial working memory task (Fig. 3d, red) and activations in the same areas also tended to increase with increasing task difficulty in the spatial discrimination task (nonsignificant; blue and yellow). The effects observed in the anterior insula and IPL during pitch and spatial memory tasks could be associated with auditory working memory (Alain et al., 2010; Brechmann et al., 2007; Gaab et al., 2006; Martinkauppi et al., 2000) or categorical processing required by these tasks. However, although the requirements for working memory and categorical processing were probably lower during the discrimination tasks than during the memory tasks, activations in these

T. Rinne et al. / NeuroImage 59 (2012) 4126–4131

areas tended to increase with increasing task difficulty also during pitch and spatial discrimination tasks. Together these results suggest, at least tentatively, that the task-difficulty effects seen in anterior insula and IPL are not associated with working memory as such but could be due to some more general aspect of the demanding auditory tasks. In our previous study, we unexpectedly found that activations in wide AC areas decreased with increasing memory load in the (categorical) pitch memory task (Rinne et al., 2009). We suggested that this deactivation was due to an active interruption of default pitch processing as soon as the pitch category was resolved in order to save resources and time for the actual memory task. In the present study, however, we found that this effect is not specific to pitch tasks as a similar deactivation was detected also during spatial (categorical) memory tasks performed on noise bursts (no clear pitch; Fig. 3e, red). These deactivations observed during categorical n-back memory tasks demonstrate the complex dynamics of AC activations during active listening tasks. Our results are partly in line with the prevailing dual-stream model of auditory processing assuming an anterior “what” (anterior STG and HG) and posterior “where” (posterior STG and IPL) pathway. However, the present anterior-posterior differences in AC activations were observed between spatial discrimination and spatial memory tasks both performed on spatially varying sounds and both requiring analysis of spatial information. Our previous study (Rinne et al., 2009) found similar anterior-posterior differences between pitch discrimination and pitch memory tasks. While these results are not necessarily inconsistent with idea that areas of anterior and posterior AC have different roles in processing of spatial and nonspatial information, they indicate that the pattern of AC activations observed with fMRI is strongly dependent on the characteristics of the listening task and cannot be explained only by stimulus-level differences (e.g., between spatial and nonspatial conditions). Together our present and previous results underline the importance of investigating task-dependent modulations of AC activations in order to understand the functional organization of human AC (Hickok and Poeppel, 2007).

Author contributions TR designed the experiment with help from SK and PW, analyzed the data, and wrote the manuscript. SK implemented and performed the experiment with help from PW. ST implemented data analysis tools. OS participated in the analysis of anatomical MRI data.

Acknowledgments This work was supported by the Academy of Finland (grants #1135900 and #1141006) and the National Doctoral Programme of Psychology in Finland.

4131

References Alain, C., Shen, D., Yu, H., Grady, C., 2010. Dissociable memory- and response-related activity in parietal cortex during auditory spatial working memory. Front. Psychol. 1. Arnott, S.R., Binns, M.A., Grady, C.L., Alain, C., 2004. Assessing the auditory dual-pathway model in humans. NeuroImage 22, 401–408. Barrett, D.J.K., Hall, D.A., 2006. Response preferences for "what" and "where" in human non-primary auditory cortex. NeuroImage 32, 968–977. Brechmann, A., Gaschler-Markefski, B., Sohr, M., Yoneda, K., Kaulisch, T., Scheich, H., 2007. Working memory specific activity in auditory cortex: potential correlates of sequential processing and maintenance. Cereb. Cortex 17, 2544–2552. Degerman, A., Rinne, T., Salmi, J., Salonen, O., Alho, K., 2006. Selective attention to sound location or pitch studied with fMRI. Brain Res. 1077, 123–134. Deouell, L.Y., Heller, A.S., Malach, R., D'Esposito, M., Knight, R.T., 2007. Cerebral responses to change in spatial location of unattended sounds. Neuron 55, 985–996. Gaab, N., Gaser, C., Schlaug, G., 2006. Improvement-related functional plasticity following pitch memory training. NeuroImage 31, 255–263. Griffiths, T.D., Warren, J.D., 2002. The planum temporale as a computational hub. Trends Neurosci. 25, 348–353. Hall, D.A., Haggard, M.P., Akeroyd, M.A., Summerfield, A.Q., Palmer, A.R., Elliott, M.R., Bowtell, R.W., 2000. Modulation and task effects in auditory processing measured using fMRI. Hum. Brain Mapp. 10, 107–119. Hickok, G., Poeppel, D., 2007. The cortical organization of speech processing. Nat. Rev. Neurosci. 8, 393–402. Krumbholz, K., Schönwiesner, M., Rübsamen, R., Zilles, K., Fink, G.R., von Cramon, D.Y., 2005. Hierarchical processing of sound location and motion in the human brainstem and planum temporale. Eur. J. Neurosci. 21, 230–238. Martinkauppi, S., Rämä, P., Aronen, H.J., Korvenoja, A., Carlson, S., 2000. Working memory of auditory localization. Cereb. Cortex 10, 889–898. Mayer, A.R., Franco, A.R., Canive, J., Harrington, D.L., 2009. The effects of stimulus modality and frequency of stimulus presentation on cross-modal distraction. Cereb. Cortex 19, 993–1007. McLachlan, N., Wilson, S., 2010. The central role of recognition in auditory perception: a neurobiological model. Psychol. Rev. 117, 175–196. Petkov, C.I., Kang, X., Alho, K., Bertrand, O., Yund, E.W., Woods, D.L., 2004. Attentional modulation of human auditory cortex. Nat. Neurosci. 7, 658–663. Rauschecker, J.P., 2011. An expanded role for the dorsal auditory pathway in sensorimotor control and integration. Hear. Res. 271, 16–25. Recanzone, G.H., Cohen, Y.E., 2010. Serial and parallel processing in the primate auditory cortex revisited. Behav. Brain Res. 206, 1–7. Rinne, T., 2010. Activations of human auditory cortex during visual and auditory selective attention tasks with varying difficulty. Open Neuroimaging J. 4, 187–193. Rinne, T., Pekkola, J., Degerman, A., Autti, T., Jääskelainen, I.P., Sams, M., Alho, K., 2005. Modulation of auditory cortex activation by sound presentation rate and attention. Hum. Brain Mapp. 26, 94–99. Rinne, T., Koistinen, S., Salonen, O., Alho, K., 2009. Task-dependent activations of human auditory cortex during pitch discrimination and pitch memory tasks. J. Neurosci. 29, 13338–13343. Smith, K.R., Hsieh, I.H., Saberi, K., Hickok, G., 2010. Auditory spatial and object processing in the human planum temporale: no evidence for selectivity. J. Cogn. Neurosci. 22, 632–639. Sussman, E., Winkler, I., Huotilainen, M., Ritter, W., Näätänen, R., 2002. Top-down effects can modify the initially stimulus-driven auditory organization. Cogn. Brain Res. 13, 393–405. Warren, J.D., Griffiths, T.D., 2003. Distinct mechanisms for processing spatial sequences and pitch sequences in the human auditory brain. J. Neurosci. 23, 5799–5804. Woodruff, P.W., Benson, R.R., Bandettini, P.A., Kwong, K.K., Howard, R.J., Talavage, T., Belliveau, J., Rosen, B.R., 1996. Modulation of auditory and visual cortex by selective attention is modality-dependent. Neuroreport 7, 1909–1913. Zatorre, R.J., Bouffard, M., Ahad, P., Belin, P., 2002. Where is 'where' in the human auditory cortex? Nat. Neurosci. 5, 905–909.