|
|
发表于 2007-5-23 15:55:44
|
显示全部楼层
Learning-related changes of brain activation in the visual ventral stream: An fMRI study of mirror reading skill
Hiroko Mochizuki-Kawaia, b, e, , , Takashi Tsukiuraa, Satoshi Mochizukic and Mitsuru Kawamurab, d
aCognitive and Behavioral Science Group, Neuroscience Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba CENTRAL 2, 1-1-1 Umezono, Tsukuba 305-8568, Japan
bDepartment of Neurology, Showa University School of Medicine, Tokyo 142-8666, Japan
cInstitute of Psychology, University of Tsukuba, Tsukuba 305-8572, Japan
dCREST, Japan Science and Technology Corporation, Kawaguchi 332-0012, Japan
eNational Institute of Floricultural Science, National Agriculture and Food Research Organization (NIFS-NARO), Tsukuba 305-8519, Japan
Accepted 1 September 2006. Available online 23 October 2006.
Abstract
A previous neuroimaging study has indicated that the visual dorsal stream may contribute to accurate reading of mirror-reversed words. However, the role of the visual ventral stream in the learning of mirror reading skill remains ambiguous. In the present fMRI study, we investigated learning-related changes in brain activation in the visual ventral stream in a mirror reading task. Subjects participated in three successive runs of the mirror reading task, in each of which they were asked to read mirror-reversed words and normal words as accurately and as quickly as possible. The behavioral data for the mirror reading condition showed significant improvement in reaction time but not in performance accuracy across the three runs. The activation data showed different learning-associated patterns related to the right and left visual ventral streams. On the right side, activity related to the reading of mirror stimuli was significantly greater than that related to normal stimuli in the first run only, whereas on the left side it was greater in all runs. Additional correlation analysis between response time data and percentage signal changes only in the mirror reading condition showed significant correlation on the right visual ventral stream in the first run only, whereas that on the left visual ventral stream was found only in the third run. The dissociable response between the right and left visual ventral streams may reflect learning-related changes in reading strategy and may be critical in improving the speed of reading mirror-reversed words.
Keywords: fMRI; Mirror reading; Speed; Fusiform gyrus; Visual ventral stream; Visual dorsal stream
Article Outline
1. Introduction
2. Results
2.1. Behavioral data
2.2. fMRI data
3. Discussion
3.1. Right visual ventral stream
3.2. Left visual ventral stream
3.3. Learning model for mirror reading skill
4. Experimental procedures
4.1. Subjects
4.2. Word stimuli
4.3. Experimental design
4.4. Data acquisition and analysis
Acknowledgements
References
1. Introduction
Previous neuropsychological and neuroimaging studies have investigated brain activations during skill learning by using several different tasks and reported that skill-learning processes are mediated by several brain areas, including the frontal, parietal, striatal, and cerebellar regions (e.g., Mochizuki-Kawai et al., 2004 and Sakai et al., 1998). The mirror reading task has been employed to investigate the brain regions related to perceptual skill learning (e.g., Kassubek et al., 2001, Poldrack et al., 2001 and Yamadori et al., 1996), and the degree of skill learning in this task has been evaluated in terms of the accuracy and speed of reading. Previous functional magnetic resonance imaging (fMRI) studies have indicated that the visual dorsal and ventral streams may contribute differentially to the learning of mirror reading skill (Kassubek et al., 2001 and Poldrack et al., 1998).
For the visual dorsal stream, Dong et al. (2000) reported a significant correlation between fMRI signal changes in the right superior parietal cortex and behavioral data (accuracy) when performing the mirror reading task. This finding suggests that the dorsal occipito-parietal activations may reflect the spatial transformation process of mirror-reversed stimuli and contribute to the accuracy rather than the speed of reading mirror-reversed words (Dong et al., 2000). In contrast, Poldrack et al., 1998 and Poldrack et al., 2001 suggested that the ventral occipito-temporal stream may be concerned with the linguistic recognition process, which includes the identification of letter form, phonological representation, and/or lexical processing. Dong et al. (2000) investigated brain activations in the visual ventral stream during the reading of Japanese kana (phonogram) mirror words. In their study, subjects reversed individual letters one by one and individually matched the reversed letters with the normal ones in the mirror reading condition, whereas they identified several letters as a word form (not individual letters) in the normal reading condition. These findings suggest that the strategy involved in the letter identification process in the mirror reading condition may cause greater activation in the left visual ventral stream.
Previous studies have reported that the letter form identification process may be strongly related to the speed of recognition of mirror-reversed letters (Hamm et al., 2004 and Heil et al., 1998). Hamm et al. (2004) measured the event-related brain potential during a parity judgment task, in which subjects were visually presented with rotated mirror/normal letters and required to judge whether the letters presented were mirror-reversed or normal (Hamm et al., 2004). The behavioral data suggest that not only simple rotation of letters in the picture plane (spatial rotation) but also matching of rotated and normal letters (visual form matching) may be needed in the parity judgment task and, additionally, that the latter process may cause delayed response time for mirror-reversed stimuli. Another electrophysiological study involving a parity judgment task showed that subjects completed the preparation of their hand response before completing spatial rotation of the presented stimuli (Heil et al., 1998). This finding suggests that response preparation for judgment may be mainly dependent on visual form processing, which may affect the response time required to judge letter parity. Thus, we hypothesized that the letter identification process based on visual form processing, which may be mediated in the visual ventral stream, may play an important role in the reading of mirror-reversed letters, whereas the spatial transformation (rotation) process, which may be mediated in the visual dorsal stream, may play a supplementary role in the reading of mirror-reversed letters only when this cannot be successfully completed by visual form processing. According to the previous findings, the former process may be reflected in reading speed, and the latter in accuracy of reading. However, Dong et al. (2000) failed to identify significant correlations between improvement in reading speed and activity in the visual ventral stream. Other neuroimaging studies have not discriminated between accurate and rapid reading when considering the contribution of the visual ventral stream (Kassubek et al., 2001 and Poldrack et al., 2001).
In the present fMRI study, we investigated the brain activations associated with learning-related change in reading speed in the mirror reading task. Previous similar neuroimaging and neuropsychological studies have reported a trend for reading times to decrease markedly in the early stage of learning, and for accuracy to show a relative improvement after the initial change in reading speed, although these changes were not examined statistically (Daum et al., 1993, Martone et al., 1984, Poldrack et al., 2001 and Yamadori et al., 1996). Thus, in the present study, we investigated the changes in brain activation related to the mirror reading task in the early learning phase only, during which the speed (but not the accuracy) of reading mirror words may improve markedly. In the present study, subjects participated in three successive runs of a block-designed fMRI experiment, and learning-related changes in brain activations were measured across the three runs. Each fMRI run consisted of mirror, normal, and visual fixation (baseline) conditions (Fig. 1A). In the mirror condition, subjects were presented with mirror-reversed Japanese words (Fig. 1B), whereas in the normal condition they were presented with normal words. In both conditions, subjects were required to judge whether the presented words were real or non-words and to press the appropriate response buttons. At the first step, we identified brain activations in the contrast of the mirror and normal conditions in each run. Second, we defined the activation clusters detected from the first-step analyses as regions of interests (ROIs), in which the mean percentage signal changes were analyzed by two-way ANOVA with factors of stimulus type and run. To better characterize the relationship between activity in the ROIs and reading speed, we analyzed correlation coefficients between the percent signal changes and reading times only in mirror words. The learning-dependent change of brain activations in the visual ventral stream may be related to improvement in the speed of reading mirror words and reflect a change in the process of identification of mirror letters. We believe that the findings in the present study may provide a new insight into the cognitive and neural model of the learning of mirror reading skill.
Display Full Size version of this image (31K)
Fig. 1. (A) Experimental design. Eight of fifteen subjects started with the mirror block and the other 7 subjects started with the normal block. Baseline blocks (visual fixation) were placed between the mirror and normal blocks. (B) Examples of real word (meaning economics) and non-word stimuli. (The mirror stimuli consisted of each letter reversed into the mirror version.) Pronounceable non-words were created from real words by replacing one of the last three letters. Each stimulus appeared only once for each subject. In the mirror and normal blocks, ten words, including two or three non-words, were presented in a random order (duration of presentation: 3 s; ISI: 2 s).
2. Results
2.1. Behavioral data
Behavioral data are shown in Table 1. In the mean response times, a two-way ANOVA with the factors of stimulus type and run revealed significant main effects of stimulus type [F(1,28) = 105.24; p < 0.01] and run [F(2,56) = 11.53; p < 0.01], and interaction between these factors [F(2,56) = 5.22; p < 0.01]. Post hoc comparisons using Tukey's method showed a significant decrease in response times across the three runs only in the mirror condition. The mean response times in the second and third runs were significantly shorter than that in the first run (p < 0.01). However, the response times for normal stimuli did not change across the three runs. In all three runs, the response times for mirror words were significantly longer than those for normal words (p < 0.01). In the mean scores of correct response (percentage), a two-way ANOVA showed a significant main effect of stimulus type [F(1,28) = 4.65; p < 0.05]. Subjects responded correctly in the normal condition more often than in the mirror condition. We did not identify a significant main effect of run [F(2,56) = 0.79] or significant interaction between the two factors [F(2,56) = 1.01].
Table 1.
Behavioral data during mirror reading task across 3 runs Conditions Run Reaction time (ms) mean (SD) Correct response (%) mean (SD)
Mirror 1 2085 (134.4) 78.7 (14.8)
2 1971 (224.9) 80.0 (14.4)
3 1905 (182.4) 78.2 (14.9)
Normal 1 1352 (177.5) 88.8 (6.3)
2 1287 (198.6) 84.8 (10.7)
3 1321 (225.3) 87.3 (8.0)
2.2. fMRI data
The fMRI data obtained during the reading of mirror words were compared to those obtained during the reading of normal words. In the first run, we found significant activations in the bilateral superior parietal lobule (BA 7), right middle and inferior frontal gyri (BA 6 and 44), right fusiform gyrus (BA 19), right cerebellar vermis, left medial frontal lobe (BA 6), and left cerebellar hemisphere. In the second run, significant activations were identified in the bilateral superior parietal lobule (BA 7), right inferior frontal gyrus (BA 45), right precentral gyrus (BA 6), right inferior temporal gyrus (BA 37), right fusiform gyrus (BA 19), left precentral and medial frontal lobe (BA 6), and left inferior occipital gyrus (BA 19). The comparison in the third run showed significant activations in the bilateral inferior frontal (BA 44/45) and fusiform (BA 19) gyri, right superior parietal lobule (BA 7), left middle frontal (BA 6) and occipital gyri (BA 19), and left supramarginal gyrus (BA 40). These activations are summarized in Table 2.
Table 2.
Significantly activated brain regions during mirror reading relative to normal
--------------------------------------------------------------------------------
Locations
--------------------------------------------------------------------------------
BA
--------------------------------------------------------------------------------
Coordinates
--------------------------------------------------------------------------------
t value
--------------------------------------------------------------------------------
Voxel size
--------------------------------------------------------------------------------
x y z
1st run
Right Middle frontal gyrus 6 30 1 52 4.95 29
Inferior frontal gyrus 44 51 7 25 5.02 35
Fusiform gyrus 19 44 − 68 − 10 5.80 266
Superior parietal lobule 7 30 − 58 36 4.71 70
Cerebellar vermis 2 − 71 − 23 4.77 69
Left Medial frontal lobe 6 − 4 14 44 5.02 56
Superior parietal lobule 7 − 28 − 62 40 5.29 51
Cerebellar hemisphere − 38 − 67 20 6.27 492
2nd run
Right Inferior frontal gyrus 45 48 15 21 8.95 271
Precentral gyrus 6 26 0 41 5.22 81
Inferior temporal gyrus 37 46 − 61 − 10 5.88 31
Fusiform gyrus 19 32 − 80 − 9 5.71 76
Superior parietal lobule 7 28 − 54 43 10.18 743
Left Medial frontal lobe 6 − 8 12 44 6.16 43
Precentral gyrus 6 − 42 3 22 8.63 465
Superior parietal lobule 7 − 28 − 64 40 7.53 332
Inferior occipital gyrus 19 − 40 − 80 − 6 7.20 248
3rd run
Right Inferior frontal gyrus 44 42 8 14 5.05 38
Inferior frontal gyrus 45 44 22 10 5.50 39
Fusiform gyrus 19 40 − 65 − 10 7.87 238
Superior parietal lobule 7 20 − 64 44 4.81 85
Left Middle frontal gyrus 6 − 28 − 2 41 4.28 26
Inferior frontal gyrus 44 − 36 3 26 5.78 46
Fusiform gyrus 19 − 42 − 76 − 6 8.73 478
Supramarginal gyrus 40 − 42 − 37 35 4.50 22
Middle occipital gyrus 19 − 32 − 75 22 6.11 429
BA: Brodmann area.
To identify learning-related changes in brain activation, the percent signal changes in the regions activated significantly in the subtraction analyses for each run were extracted from the mirror and normal conditions, and the mean signal changes were analyzed by two-way ANOVAs. The two-way ANOVAs revealed significant interactions between stimulus type and run only in the right fusiform gyrus (32, − 80, − 9) [F(2,56) = 4.00; p < 0.05, see Fig. 2A], right inferior temporal gyrus [F(2,56) = 3.99; p < 0.05, see Fig. 2B], and left fusiform gyrus [F(2,56) = 4.20; p < 0.05, see Fig. 2C]. Post hoc comparisons using Tukey's method showed that the percent signal changes in these areas for the mirror stimuli only decreased significantly from the first run to the second or third runs (p < 0.01). The decreasing pattern of signal changes was similar to that of the behavioral data for response time. In the right fusiform and inferior temporal gyri, the percent signal changes for the mirror stimuli were significantly greater than those for the normal stimuli only in the first run (p < 0.01). In the left fusiform gyrus, the signal changes for the mirror stimuli were significantly greater than those for the normal stimuli in the first (p < 0.01), second (p < 0.01), and third runs (p < 0.10; significant trend).
Display Full Size version of this image (62K)
Fig. 2. Activated images during the mirror reading condition relative to normal, and mean percent signal changes in each region: (A) right fusiform gyrus (32, − 80, − 9); (B) right inferior temporal gyrus (46, − 61, − 10); (C) left fusiform gyrus (− 42, − 76, − 6).
In the bilateral superior parietal lobule (right: 20, − 64, 44; left: − 28, − 64, 40), right fusiform gyrus (40, − 65, − 10), left middle frontal gyrus, left middle and inferior occipital gyri, and left cerebellar hemisphere, two-way ANOVAs showed significant main effects of stimulus type (p < 0.01) and run (p < 0.01). However, no significant interaction between the two factors was identified in these areas. In other regions, including the bilateral inferior frontal and precentral gyri, medial frontal lobe, superior parietal lobule (right: 30, − 58, 36; right: 28, − 54, 43; left: − 28, − 62, 40), right cerebellar vermis, right fusiform gyrus (44, − 68, 10), middle frontal gyrus, and left supramarginal gyrus, we found a significant main effect of run (p < 0.01), but not of stimulus type. Furthermore, no interaction was found between run and stimulus type.
In the bilateral visual ventral stream, we detected significant interaction between stimulus type and run. To identify the precise roles of these areas in reading speed, we calculated the percent signal changes during reading mirror words in each peak voxel and examined the correlation between the percent signal changes and reaction times only in the mirror trials by each run (Fig. 3). In the right fusiform gyrus, the percent signal change was significantly correlated with the reaction time data only in the first run (r = 0.596, p < 0.05), whereas the correlation coefficients were not significant in the second and third runs. In the left fusiform gyrus, the percent signal change was significantly correlated with the reaction time in the third run (r = 0.559, p < 0.05). We found no significant correlation in the right inferior temporal gyrus across the three runs (first run: r = − 0.043, second run: r = − 0.012, third run: r = 0.081).
Display Full Size version of this image (23K)
Fig. 3. Results of correlation analyses between reaction times and percent signal changes in the right and left fusiform gyri during reading mirror words.
3. Discussion
In the present study, we investigated learning-dependent changes in brain activation during the learning of mirror reading skill. The behavioral data showed that the reading times for mirror words decreased significantly across the three runs, whereas the accuracy of reading mirror words did not change across the runs. The findings suggest that subjects may acquire skill for the rapid reading of mirror-reversed words but not for accurate reading. Thus, the changes in brain activations in the present study may reflect the learning of rapid reading rather than of accurate reading. Learning-related decrease in brain activations specific to the mirror reading condition was identified in the right fusiform gyrus (BA 19), inferior temporal gyrus (BA 37), and left fusiform gyrus (BA 19), which are involved in part of the visual ventral stream. The findings suggest that the bilateral visual ventral stream may contribute to the rapid reading of mirror-reversed words. In addition, we found different patterns in the results of ROI analysis between the right and left visual ventral streams. In the right visual ventral stream, the percent signal changes for mirror stimuli were significantly higher than those for normal stimuli only in the first run; whereas in the left visual ventral stream, the percent signal changes for mirror stimuli were significantly higher across the three runs. Furthermore, in the correlation analysis between percent signal changes in the right and left fusiform gyrus and response time data, a significantly positive correlation was observed in the right fusiform gyrus in the first run only, whereas that in the left fusiform gyrus was identified only in the third run. The findings from ROI and correlation analyses suggest that the right and left visual ventral streams may contribute differentially to the learning of mirror reading skill.
3.1. Right visual ventral stream
In the right visual ventral stream, including the right fusiform (BA 19) and inferior temporal gyri (BA 37), the percent signal change in brain activation for the mirror reading condition decreased in association with improvement in the speed of reading mirror-reversed words. These areas were significantly activated in the contrast of the mirror with the normal reading condition in the first run, but not in the second and third runs. In the right fusiform gyrus, we found a significantly positive correlation between percent signal changes and reaction time in the first run only. These findings suggest that the right visual ventral stream may contribute to the very early stage of the learning of rapid reading in mirror reading skill.
A combined study of neuropsychology with neuroimaging for a patient with alexia reported that the right occipito-temporal area, which was preserved in this patient, may play a pivotal role in letter-by-letter reading (Cohen et al., 2004). The phenomenon of letter-by-letter reading is often observed in patients with alexia or dyslexia, who are impaired in reading words or sentences presented visually. Although these patients cannot recognize the visual form of words presented as a whole (Salmelin et al., 1996), they can understand the meaning of words using the strategy of letter-by-letter reading. Cohen et al. (2004) investigated the brain activations during letter-by-letter reading in their patient with alexia, who had damages in the left occipito-temporal region, and reported the right occipito-temporal activations. The findings indicate that the visual form of letters may be processed in the right visual ventral stream and that the patient with alexia may identify each word serially, based on stored representations of the letters. The activation patterns of the right visual ventral stream observed in the present study may reflect that subjects identified the mirror-reversed word serially using the letter-by-letter reading strategy in the first run of the mirror reading task, but being able to recognize the mirror-reversed words without using this strategy in the second and third runs. The learning-related change of reading strategies may cause decreased activity in the right visual ventral stream and improvement in the speed of reading mirror-reversed words. Our results indicate that the process of serial identification of letters, which is represented by a letter-by-letter reading strategy, may be an important factor for reading speed in the mirror reading task only in the very early stage of learning.
3.2. Left visual ventral stream
We found a learning-related decrease of brain activations in a posterior part of the left visual ventral stream (fusiform gyrus). The percent signal changes for mirror stimuli in this area were significantly higher than those for normal stimuli in all runs. The pattern of brain activations in the left visual ventral stream was different from that in the right visual ventral stream, in which significant difference in brain activation between the mirror and normal conditions was observed in the first run only. In addition, correlation analyses between percent signal changes and reaction time showed a significantly positive correlation only in the third run. These findings suggest that the left visual ventral stream may play a different role from that of the right visual ventral stream in the learning of mirror reading skill.
McCandliss et al. (2003) suggested that the left fusiform area may contribute to the processing of visually presented words and to the rapid reading of words. They proposed that fast and parallel identification of several letters may be an important factor in smooth reading and may be mediated in the left fusiform gyrus (McCandliss et al., 2003). The activation in the visual ventral stream found in the present study may reflect learning of the parallel strategy in the letter identification processes, and learning of the parallel strategy, which was mediated in the left visual ventral stream, may contribute to the improvement in reaction time in the latter runs. A relatively longer time and more practice may be needed to acquire the parallel identification skill than the serial identification skill in the reading of mirror-reversed words. However, once the parallel identification skill is acquired, it may allow us to read mirror words more rapidly than the serial identification skill. In the acquisition of mirror reading skill, the serial strategy of letter identification may be required in the relatively early stage of learning, whereas the parallel one may contribute to the learning in the relatively latter stage. The activation patterns found in the present study indicate that the two different mechanisms of serial and parallel identification of letters may support the process of learning mirror reading skill in different ways.
3.3. Learning model for mirror reading skill
The present findings support our hypothesis that the letter identification process based on visual form processing, which may be mediated in the visual ventral stream, may play an important role in the reading of mirror-reversed letters. Previous studies have also reported that the letter identification process based on visual form processing may be strongly related to the speed of recognition of mirror-reversed letters (Hamm et al., 2004 and Heil et al., 1998). In contrast, the process of spatial transformation (rotation) of mirror-reversed letters, which may be mediated in the visual dorsal stream, may play a supplementary role in reading these letters only when this cannot be achieved by visual form processing. A previous study by Dong et al. (2000) reported the relationship between reading accuracy and the visual dorsal stream. Thus, the process of learning mirror reading skill may be organized mainly by two different systems, one of which is the letter identification process based on the visual form processing mediated in the visual ventral stream, and the other is the spatial transformation (rotation) process mediated in the visual dorsal stream. The former system may be strongly related to the process of rapid reading and function as a primary system of the learning of mirror reading skill, whereas the latter system to the accurate reading skill and function as a secondary system in the acquisition of mirror reading skill. The coordination of these systems may be required to acquire mirror reading skill. In addition, our results suggest that improvement in reading speed in the mirror reading task may be achieved by means of at least two different mechanisms, i.e., serial and parallel identification of mirror-reversed letters. The former process may be related to the early stage learning of mirror reading skill and the activations in the right visual ventral stream, whereas the latter process to the advanced stage of mirror reading learning and the activations in the left visual ventral stream. The dissociable response between the right and left visual ventral streams may reflect learning-related changes in reading strategy and may be critical in improving the speed of reading mirror-reversed words.
4. Experimental procedures
4.1. Subjects
Fifteen normal volunteers (mean age 21.93 years; 10 men and 5 women) participated in this study. All subjects were right-handed and had scores above + 80 on the Edinburgh Handedness Inventory (Oldfield, 1971). Written informed consent was obtained from all subjects based on the Declaration of Helsinki (1975). All the procedures of this study received prior approval from the Institutional Human Research Review Committee and MRI Research Review Board of our institute.
4.2. Word stimuli
Two hundred and forty Japanese nouns written in kana phonogram were chosen from the vocabulary of the NTT Database series Vol. 7 (Amano and Kondo, 2000). The words were divided into 24 sets, each consisting of ten words. There was no significant difference in word frequency among the sets [F(23,216) = 0.14]. Half of the 24 sets were used as mirror word stimuli and the other half as normal stimuli. Each word consisted of six to eight Japanese kana letters (Fig. 1B). Each letter of the mirror words was individually reversed into the mirror form to avoid oculomotor effects, which have often been observed in previous studies of mirror reading (e.g., Poldrack et al., 1998). In previous studies, subjects had to read the mirror-reversed words from right to left (e.g., Poldrack et al., 1998), whereas in the present study they could read the mirror stimuli silently from left to right. Thus, in the present experiment, the direction in which the mirror words were read was the same as that of the normal words. Two or three words in each set were used as pronounceable non-words, which were created by replacing just one letter with another. Each stimulus was presented only once per subject.
4.3. Experimental design
Subjects participated in three successive fMRI runs separated by a time interval of about 10 min. Each run consisted of 16 blocks, including four mirror, four normal, and eight baseline blocks (visual fixation), which were placed between the mirror and normal blocks (Fig. 1A). The presentation order was counterbalanced between the mirror and normal conditions across subjects. In the mirror and normal blocks, ten words, including two or three non-words, were presented in random order (duration of presentation: 3 s; ISI: 2 s). Subjects were required to judge whether the presented items were real or non-words and to press the left button with their index finger if the item was a real word or the right button with their middle finger if it was a non-word. In the baseline block, subjects were instructed to press either the left or right button as soon as a visual fixation was presented. When subjects pressed the response button within 3000 ms, the item disappeared immediately.
Each stimulus was presented through a projector and back-projected onto a screen placed above the subject's feet. Subjects were able to see the screen through a mirror fixed in a head cage. The timing of stimulus presentation was controlled by the TTL trigger pulse from the MRI scanner. The pulse signal was mediated by a response box made by us. The stimuli were presented using SuperLab Pro (http://www.cedrus.com/) software with Windows 2000 OS. The responses from each subject were recorded with a system of response buttons we devised.
4.4. Data acquisition and analysis
All MRI data were acquired by 3 T GE Signa LX MRI scanner (General Electric, Milwaukee, WI, USA). The subjects were positioned in the scanner, and their heads were immobilized with support cushions and a neck support for medical use. Before fMRI scanning, a structural localizer (SPGR) in the sagittal plane was acquired with the following parameters: TR = 68 ms, TE = 1.6 ms, FOV = 24 × 24 cm2, matrix sizes = 256 × 128, flip angle = 30°, slice thickness/gap = 7/3 mm. For fMRI scanning, a gradient echo planner imaging (EPI) sequence modified by VP real-time reconstruction system was used for functional imaging with the following parameters: TR = 5000 ms, TE = 29.8 ms, FOV = 20 × 20 cm2, matrix sizes = 64 × 64, flip angle = 90°, NEX = 1, slice thickness/gap = 4/1 mm. Twenty-seven horizontal slices were obtained, and 160 sequential images were collected in each run. After fMRI scanning, high-resolution T2 anatomical images were collected (TR = 24 ms, TE = 5 ms, FOV = 20 × 20 cm2, matrix size = 256 × 192, flip angle = 90°) at the same slices as for the EPI scans (slice thickness/gap = 4/1 mm, 27 horizontal slices).
The acquired data were analyzed using SPM99 (Wellcome Department of Cognitive Neurology, http://www.fil.ion.ucl.ac.uk/spm). First, time series of EPI volumes were corrected for differences in slice acquisition times and realigned for motion corrections. Second, these realigned images were coregistered with the T2 anatomical images of each subject. Third, a parameter file for spatial normalization into MNI brain template was constructed from the individual T2 anatomical images with a standard SPM T2 template. All EPI volumes coregistered were spatially normalized into an MNI space with the parameter file and resliced into a resolution of 2 × 2 × 4 mm voxels. All EPI data normalized spatially were spatially smoothed using a Gaussian kernel of FWHM = 8 mm.
In the subtraction analysis, we assessed average activations across subjects by carrying out a two-step random-effect analysis. In the first step, specific effects to the mirror reading condition were tested by applying an appropriate contrast of the mirror with normal conditions to the parameter estimates for each condition, resulting in t-statistics for every voxel. In this way, the contrast images were determined for each individual subject in each run. In the second step, we carried out a one-sample t-test on the contrast images. Activations were considered significant if they had a cluster of 20 or more voxels above t > 3.79 (uncorrected for multiple comparisons; p < 0.001) with maximum t value. Activated locations were converted from MNI to Talairach and Tournoux (1988) space with MNI2TAL tool (http://www.mrc-cbu.ac.uk/Imaging/Common).
To identify the precise role in activated regions detected from the subtraction analyses, we compared percent signal changes in all activated regions among runs by using the software of MarsBar (http://www.mrc.cbu.cam.ac.uk/Imaging/marsbar.html). First, we extracted mean percent signal changes from each activated cluster detected by the contrasts of the mirror with normal conditions in each run for each subject. Mean percent signal changes were calculated from mean intensities in the mirror or normal condition. The mean percent signal changes in each cluster of all activated regions were compared using two-way ANOVA with factors of stimulus type and run to identify the regions showing learning-dependent change in brain activations. Subsequently, we performed correlation analysis in regions that showed a significant interaction between stimulus type and run. We extracted the signal intensities from a peak voxel of each region with the significant interaction and examined the correlation coefficients between the percent signal changes and reaction time data only in the mirror trials by each run.
Acknowledgments
H. Mochizuki-Kawai is supported by the Japan Society for the Promotion of Science (JSPS) Research Fellowships for Young Scientists. This study was supported by a Grant-in-Aid to H. Mochizuki-Kawai (01927) for scientific research from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This study was supported in part by a Showa University Grant-in-Aid for innovative collaborative research projects, a Special Research Grant-in-Aid for Development of Characteristic Education from the Japanese Ministry of Education, Culture, Sports, Science and Technology, and a Grant-in-Aid for Scientific Research on Priority Areas (c) from the Japanese Ministry of Education, Culture, Sports, Science and Technology (No. 15590910). Part of this study was also supported by the Independent Administrative Organization under the Ministry of Economy, Trade and Industry (METI) of Japan.
References
Amano and Kondo, 2000 N. Amano and K. Kondo, NTT Database Series [Lexical Properties of Japanese] Vol. 7: Frequency, Sanseido Press, Tokyo (2000).
Cohen et al., 2004 L. Cohen, C. Henry, S. Dehaene, O. Martinaud, S. Lehericy, C. Lemer and S. Ferrieux, The pathophysiology of letter-by-letter reading, Neuropsychologia 42 (2004), pp. 1768–1780. SummaryPlus | Full Text + Links | PDF (491 K) | View Record in Scopus | Cited By in Scopus
Daum et al., 1993 I. Daum, H. Ackermann, M.M. Schugens, C. Reimold, J. Dichgans and N. Birbaumer, The cerebellum and cognitive functions in humans, Behav. Neurosci. 107 (1993), pp. 411–419. Abstract | PDF (857 K) | Full Text via CrossRef | View Record in Scopus | Cited By in Scopus
Dong et al., 2000 Y. Dong, H. Fukuyama, M. Honda, T. Okada, T. Hanakawa, K. Nakamura, Y. Nagahama, T. Nagamine, J. Konishi and H. Shibasaki, Essential role of the right superior parietal cortex in Japanese kana mirror reading: an fMRI study, Brain 123 (2000), pp. 790–799. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus
Hamm et al., 2004 J.P. Hamm, B.W. Johnson and M.C. Corballis, One good turn deserves another: an event-related brain potential study of rotated mirror-normal letter discriminations, Neuropsychologia 42 (2004), pp. 810–820. SummaryPlus | Full Text + Links | PDF (250 K) | View Record in Scopus | Cited By in Scopus
Heil et al., 1998 M. Heil, M. Rauch and E. Hennighausen, Response preparation begins before mental rotation is finished: evidence from event-related brain potentials, Acta Psychol. 99 (1998), pp. 217–232. SummaryPlus | Full Text + Links | PDF (303 K) | View Record in Scopus | Cited By in Scopus
Kassubek et al., 2001 J. Kassubek, K. Schmidtke, H. Kimmig, C.H. Lucking and M.W. Greenlee, Changes in cortical activation during mirror reading before and after training: an fMRI study of procedural learning, Cogn. Brain Res. 10 (2001), pp. 07–217.
Martone et al., 1984 M. Martone, N. Butters, M. Payne, J.T. Becker and D.S. Sax, Dissociations between skill learning and verbal recognition in amnesia and dementia, Arch. Neurol. 41 (1984), pp. 965–970. View Record in Scopus | Cited By in Scopus
McCandliss et al., 2003 B.D. McCandliss, L. Cohen and S. Dehaene, The visual word form area: expertise for reading in the fusiform gyrus, Trends Cogn. Sci. 7 (2003), pp. 93–299.
Mochizuki-Kawai et al., 2004 H. Mochizuki-Kawai, M. Kawamura, Y. Hasegawa, S. Mochizuki, R. Oeda, K. Yamanaka and H. Tagaya, Deficits in long-term retention of learned motor skills in patients with cortical or subcortical degeneration, Neuropsychologia 42 (2004), pp. 1858–1863. SummaryPlus | Full Text + Links | PDF (144 K) | View Record in Scopus | Cited By in Scopus
Oldfield, 1971 R.C. Oldfield, The assessment and analysis of handedness: the Edinburgh Inventory, Neuropsychologia 9 (1971), pp. 97–113. Abstract | Abstract + References | PDF (1067 K) | View Record in Scopus | Cited By in Scopus
Poldrack et al., 1998 R.A. Poldrack, J.E. Desmond, G.H. Glover and J.D. Gabrieli, The neural basis of visual skill learning: an fMRI study of mirror reading, Cereb. Cortex 8 (1998), pp. 1–10. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus
Poldrack et al., 2001 R.A. Poldrack, J.D. Gabrieli, J.E. Desmond and G.H. Glover, Characterizing the neural mechanisms of skill learning and repetition priming: evidence from mirror reading, Brain 124 (2001), pp. 67–82. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus
Sakai et al., 1998 K. Sakai, O. Hikosaka, S. Miyauchi, R. Takino, Y. Sasaki and B. Putz, Transition of brain activation from frontal to parietal areas in visuomotor sequence learning, J. Neurosci. 18 (1998), pp. 1827–1840. View Record in Scopus | Cited By in Scopus
Salmelin et al., 1996 R. Salmelin, E. Service, P. Kiesila, K. Uutela and O. Salonen, Impaired visual word processing in dyslexia revealed with magnetoencephalography, Ann. Neurol. 40 (1996), pp. 157–162. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus
Talairach and Tournoux, 1988 J. Talairach and P. Tournoux, Co-planar Stereotactic Atlas of the Human Brain: 3-dimensional Proportional System: An Approach to Cerebral Imaging, Georg. Thieme, Stuttgart (1988).
Yamadori et al., 1996 A. Yamadori, T. Yoshida, E. Mori and H. Yamashita, Neurological basis of skill learning, Cogn. Brain Res. 5 (1996), pp. 49–54. Abstract | PDF (458 K) | View Record in Scopus | Cited By in Scopus
Corresponding author. Present Address: National Institute of Floricultural Science, National Agriculture and Food Research Organization (NIFS-NARO), Tsukuba 305-8519, Japan. Fax: +81 29 838 6842. |
|
楼主: 诗雨寒梦
狗仔卡
显身卡
楼主
发表于 2007-5-22 12:45:03
发表于 2007-5-22 17:41:17
