Study on Orienting of Visuospatial Attention by Task and Rest Functional Magnetic Resonance Imaging

全文

(1)Title of Thesis. Study on Orienting of Visuospatial Attention by Task and Rest Functional Magnetic Resonance Imaging. August, 2014. Yujie Li. The Graduate School of Natural Science and Technology (Doctor’s Course) OKAYAMA UNIVERSITY.

(2)

(3) Abstract. Orienting of visuospatial attention was investigated under different cue-type and inter-stimulus interval (ISI) conditions by functional magnetic resonance imaging (fMRI). In addition, alterations within such visuospatial attention-related network in normal older adults and patients with Alzheimer’s disease (AD) were also examined by combining task and rest fMRI. Four experiments were conducted and a modified version of Posner paradigm was used in the present study. Important findings are summarized below: 1.. In the first experiment, by comparing the neural correlates of orienting of visuospatial attention under unlasting versus lasting cue conditions, the neuroimaging data revealed increased activation of the left intraparietal sulcus (IPS) in the unlasting cue condition. We suggest that the difference in activity in the left IPS between lasting and unlasting cue conditions is the neural correlate of spatial working memory.. 2.. Our second experiment focuses on the effects of ISI on orienting of visuospatial attention. The neuroimaging data revealed reduced activity in the posterior cingulate cortex (PCC) as ISI becomes longer. We propose that, as time went on, the strength of visual-spatial bias reduced, resulting in decreased PCC activation.. 3.. In the third experiment, we combined task and rest fMRI to investigate the age-dependent alterations of resting-state functional connectivity within the visuospatial attention-related network. Our results showed that marked reduction in the number of connections occurred in elderly subjects but was not uniform throughout the brain: significant loss of communications in the anterior portion of the brain and between the anterior and posterior cerebral cortex, preserved communications in the posterior portion of the brain. Moreover, the older adults exhibited weakened resting-state functional connectivity between supplementary motor area. I.

(4) and left anterior insular cortex. These findings suggest that disrupted functional connectivity of the neural network for visuospatial attention in the elderly brain may underlay the decline in cognitive performance. 4.. In the last experiment, we combined task and rest fMRI to investigate the AD-associated deterioration of task-evoked activity and resting-state functional connectivity within the visuospatial attention-related network. Our results showed that, compared with healthy control (HC), patients with AD had significant decreased activity in left dorsolateral prefrontal cortex and left precentral sulcus, which may correlate with executive dysfunction. Moreover, AD patients exhibited prominent reduction in the number of resting-state connections, which may reflect ineffective integration of information within the visuospatial attention-related network.. In conclusions, our four experiments revealed the neural basis of orienting of visual-spatial attention under different cue-type and ISI conditions, and disclosed the brain mechanisms underlying the alterations within such visuospatial attention-related network in normal older adults and patients with Alzheimer’s disease.. II.

(5) Contents Abstract ............................................................................................................................. I Contents ......................................................................................................................... III Chapter 1 Introduction..................................................................................................... 1 1.1 Orienting of attention ................................................................................................. 3 1.2 Resting-state ................................................................................................................ 4 1.3 The aim of the dissertation ......................................................................................... 6 1.4 The contents of the dissertation ................................................................................. 6 Chapter 2 fMRI study on orienting of visuospatial attention elicited by lasting and unlasting cues ................................................................................................................... 9 2.1 Introduction ................................................................................................................11 2.2 Materials and Methods ............................................................................................. 12 2.2.1 Subjects ............................................................................................................... 12 2.2.2 Stimuli and experiment paradigm .................................................................... 12 2.2.3 fMRI scanning..................................................................................................... 13 2.2.4 fMRI data analysis ............................................................................................. 14 2.3 Results........................................................................................................................ 15 2.3.1 Behavioral results............................................................................................... 15 2.3.2 fMRI results ........................................................................................................ 19 2.4 Discussion .................................................................................................................. 19 2.4.1 Brain activation during lasting cue tasks and unlasting cue tasks ............... 19 2.4.2 Different left IPS activation triggered by lasting and unlasting cues............ 21 2.5 Conclusions ................................................................................................................ 22 Chapter 3 fMRI study on orienting of visuospatial attention triggered by different inter-stimulus interval ....................................................................................................25 3.1 Introduction ............................................................................................................... 27 3.2 Methods ...................................................................................................................... 27 3.2.1 Subjects ............................................................................................................... 27 3.2.2 Stimuli and experiment paradigm .................................................................... 28 3.2.3 fMRI scanning..................................................................................................... 29 3.2.4 fMRI data analysis ............................................................................................. 29 3.3 Results........................................................................................................................ 31. III.

(6) 3.3.1 Behavioral results............................................................................................... 31 3.3.2 fMRI results ........................................................................................................ 31 3.4 Discussion .................................................................................................................. 36 3.4.1 Brain activation in different ISI conditions...................................................... 36 3.4.2 Decreased PCC activation along with increased ISI ....................................... 37 3.5 Conclusions ................................................................................................................ 38 Chapter 4 fMRI study on orienting of visuospatial attention in normal older adults ...39 4.1 Introduction ............................................................................................................... 41 4.2 Methods ...................................................................................................................... 42 4.2.1 Subjects ............................................................................................................... 42 4.2.2 Experimental design .......................................................................................... 43 4.2.3 Data acquisition .................................................................................................. 44 4.2.4 fMRI data analysis ............................................................................................. 45 4.2.4.1 Preprocessing ............................................................................................... 45 4.2.4.2 Task activation and ROIs ............................................................................ 45 4.2.4.3 Functional connectivity analysis ................................................................ 46 4.3 Results........................................................................................................................ 47 4.3.1 Behavioral results............................................................................................... 47 4.3.2 Results of task-fMRI data .................................................................................. 49 4.3.3 Results of rest-fMRI data ................................................................................... 49 4.4 Discussion .................................................................................................................. 50 4.5 Conclusions ................................................................................................................ 53 Chapter 5 fMRI study on orienting of visuospatial attention in Alzheimer’s disease....55 5.1 Introduction ............................................................................................................... 57 5.2 Methods ...................................................................................................................... 58 5.2.1 Subjects ............................................................................................................... 58 5.2.2 Experimental design .......................................................................................... 59 5.2.3 Data acquisition .................................................................................................. 60 5.2.4 fMRI data analysis ............................................................................................. 61 5.2.4.1 Preprocessing ............................................................................................... 61 5.2.4.2 Task activation and ROIs ............................................................................ 61 5.2.4.3 Functional connectivity analysis ................................................................ 62 5.3 Results........................................................................................................................ 63 5.3.1 Behavioral results............................................................................................... 63. IV.

(7) 5.3.2 Results of task-fMRI data .................................................................................. 63 5.3.3 Results of rest-fMRI data ................................................................................... 65 5.4 Discussion .................................................................................................................. 67 5.5 Conclusions ................................................................................................................ 68 Chapter 6 Conclusions and future challenges ................................................................69 6.1 Conclusions ................................................................................................................ 71 6.2 Future challenges ...................................................................................................... 72 6.2.1 Effect of normal aging and Alzheimer’s disease on attention system ............ 72 6.2.2 Interactions of spontaneous neuronal activity with task- or stimuli-evoked activity in normal aging and Alzheimer’s disease ..................................................... 72 6.2.3 Systems for early detection of dementia ........................................................... 73 Appendix A: A simple introduction to functional magnetic resonance imaging .............75 Appendix B: Analytical software for fMRI data ..............................................................81 B.1 Statistical Parametric Mapping (SPM) ................................................................... 81 B.2 Resting-State fMRI Data Analysis Toolkit (REST) ................................................ 85 B.3 Data Processing Assistant for Resting-State fMRI (DPARSF).............................. 88 Publications .....................................................................................................................91 References........................................................................................................................93 Acknowledgements ........................................................................................................101. V.

(8)

(9) Chapter 1 Introduction. Summary First, this chapter introduces the concept of orienting of attention and frontoparietal network underlying orienting of visuospatial attention. Then, the concept of resting-state and intrinsic connectivity networks at rest (resting-state networks) are also introduced. At the end, the aim and contents of the dissertation are briefly described.. 1.

(10)

(11) 1.1 Orienting of attention. 1.1 Orienting of attention Attention is the behavioral and cognitive process of selectively concentrating on one aspect of the environment while ignoring other things. For example, imagine that you visit a museum while a guide explains the ukiyo-e series Thirty-six Views of Mount Fuji by the Japanese artist Katsushika Hokusai. The guide’s words cue you to attend to different aspects of the ukiyo-e series, such as its color, spatial configuration or meaning. Attention is important because there are serial bottlenecks in human information processing, which means that it is no longer possible to continue processing everything in parallel. Attention can help us to select which pieces of information to focus on and which to neglect.. Figure 1.1. Posner task paradigm (A) and frontoparietal network (B).. Orienting of attention, defined as “the aligning of attention with a source of sensory input or an internal semantic structure stored in memory”, represents the ability to prioritize sensory input by selecting a modality or location [1]. Orienting of attention to a location improves efficiency of processing stimuli appearing at that location [2]. Orienting of visual-spatial attention is often studied within the context of the Posner task paradigm [1]. In this paradigm, a spatially informative symbolic central cue (e.g., arrow) is presented to indicate the location of an upcoming peripheral. 3.

(12) 1. Introduction. target. Targets appearing in predicted location (valid) are detected more rapidly and accurately than those that are not (invalid). Over the years, a considerable number of functional magnetic resonance imaging (fMRI) studies have revealed a frontoparietal network involved in controlling orienting of visuospatial attention using Posner paradigm. This network consists of anterior insular cortex (AIC), supplementary motor area (SMA) and dorsolateral prefrontal cortex (DLPFC) as well as intraparietal sulcus (IPS), frontal eye field (FEF) and middle temporal area (MT+) [3-5]. Previous studies showed that cue type and cue-target interval (also called inter-stimulus interval; ISI) might influence the cueing effect (invalid versus valid) [1,3,5-7]. However, no study to date has directly compared the neural correlates of orienting of visual-spatial attention under different cue type and ISI conditions. Moreover, normal aging and Alzheimer’s disease exhibited interaction with cueing effect [8,9], but the neural basis underlying the interaction remains unclear. These issues will be further elaborated and discussed in the following chapters.. 1.2 Resting-state In the past, investigators mainly focused on the brain activity elicited by task or stimuli. However, our brain is active all the time even when we are asleep and under anesthesia. Studies focusing on brain’s energy consumption have revealed that the resting human brain expends around 20% of the body’s energy, most of which is used to support ongoing neuronal signaling. Task-induced increases in neuronal metabolism are usually small (<5%) when compared with this large resting energy expenditure [10]. In other words, research on task- or stimuli-related brain activity can only assist in understanding a minor component of total brain activity. Therefore, if we hope to improve our knowledge about the brain function, we must take into consideration the component that consumes most of the brain’s energy: spontaneous neuronal activity [10].. 4.

(13) 1.2 Resting-state. Figure 1.2. Resting-state networks. (A) Default mode network; (B) Dorsal attention network; (C) Left frontoparietal network; (D) Right frontoparietal network; (E) Visual network; (F) Auditory network; (G) Sensorimotor network; (H) Self-referential network.. Recently, resting-state fMRI (rs-fMRI) has become a powerful tool for understanding the spontaneous neuronal activity [10-12]. Using rs-fMRI, one of the most important finding is that fMRI blood oxygenation level dependent (BOLD) signal from functionally related brain regions exhibit temporally coherence at rest [12]. This phenomenon was first described by Biswal et al. and termed as resting state functional connectivity. Since synchronous and spontaneous fluctuations in BOLD signal in the sensorimotor cortex has been revealed in Biswal et al.’s seminal study, functional connectivity has also been observed in visual, auditory, executive control and several other functional networks [13-15]. These large-scale coherent spatial patterns, known as resting-state networks (RSNs), were found high spatial similarity with task-induced activated patterns, such as. 5.

(14) 1. Introduction. shown in Biswal et al.’s work. Therefore the RSNs were named according to the function of its set of involved regions and considered to reflect the intrinsic functional architecture of the brain [10,16]. Furthermore, several networks were examined consistently across different subjects, reliably over time and stably through various sleep states [14,17]. Because considerable evidence suggests that functional connectivity is susceptible to various diseases such as Alzheimer’s disease (AD), resting-state fMRI has already been shown to be of great potential value for clinical applications [18]. In normal aging and Alzheimer’s disease, decline in visuospatial attention has been commonly observed [8,9]. Since resting-state functional connectivity exhibits association with cognitive performance [10], investigation on resting-state connections within visuospatial attention-related network provides insight into neural mechanisms underlying age- and AD-related changes. These issues will be further elaborated and discussed in the following chapters.. 1.3 The aim of the dissertation The main aim of this dissertation research was to examine the effects of cue type and ISI on orienting of visuospatial attention. Also of interest was to investigate the alterations of resting-state functional connectivity within visuospatial attention-related network in normal aging and Alzheimer’s disease. To achieve these aims, four related experiments were conducted.. 1.4 The contents of the dissertation The dissertation contains descriptions of the four experiments briefly introduced below. Chapter 2 describes the first experiment, in which we investigated the neural mechanisms involved in the processing of two types of cues using fMRI. Consistent with previous studies, the frontoparietal network exists for both types of cues. Furthermore, the neuroimaging data revealed. 6.

(15) 1.4 The contents of the dissertation. increased activation of the left IPS in the unlasting cue condition. Chapter 3 describes the second experiment, in which we investigated the brain activity changes along with increased ISI using fMRI. As the result, although, subjects exhibited no difference in the different ISI conditions, neuroimaging data revealed reduced activity in the posterior cingulate cortex (PCC) as ISI becomes longer. Chapter 4 describes the third experiment, in which we combined task and rest fMRI to investigate the age-dependent alterations of resting-state functional connectivity within the visuospatial attention -related network. Twenty-three young subjects and nineteen elderly subjects participated in this study and a modified Posner paradigm was used to define the region of interest (ROI). The results showed that marked reduction in the number of connections occurred with age but was not uniform throughout the brain. Moreover, the older adults exhibited weakened resting-state functional connectivity between supplementary motor area and left anterior insular cortex. Chapter 5 describes the fourth experiment, in which we combined task and rest functional magnetic resonance imaging (fMRI) to investigate the AD-associated deterioration of task-evoked activity and resting-state functional connectivity within the visuospatial attention-related network. Nineteen elderly subjects and ten AD patients participated in this study and performed a modified version of Posner paradigm. Our results showed that, compared with healthy control (HC), patients with AD had significant decreased activity in left dorsolateral prefrontal cortex and left precentral sulcus. Moreover, AD patients exhibited prominent reduction in the number of resting-state connections. In chapter 6, conclusions of the dissertation and future challenges are put forward. Finally, the appendices present topics related to the experiments, describing the principles of fMRI and tools for fMRI experiment data processing: Statistical Parametric Mapping (SPM), Data Processing Assistant for Resting-State fMRI (DPARSF) and Resting-State fMRI Data Analysis Toolkit (REST).. 7.

(16)

(17) Chapter 2. fMRI study on orienting of visuospatial attention elicited by lasting and unlasting cues. Summary In experimental paradigms, voluntary orienting of visual-spatial attention is conventionally achieved through the Posner task in which predictive central cues remain present until target offset (lasting cue) or disappear prior to target onset (unlasting cue). Previous studies have implied that lasting and unlasting cues elicit different levels of activity in the intraparietal sulcus (IPS). However, no study to date has directly compared the neural correlates of visual-spatial attention under unlasting versus lasting cue conditions. We investigated the neural mechanisms involved in the processing of both types of cues using functional magnetic resonance imaging (fMRI). Behaviorally, subjects exhibited no difference in the two cue type conditions. However, neuroimaging data revealed increased activation of the left IPS in the unlasting cue condition. We propose that in the unlasting cue condition, the time interval between the onsets of cue and target requires subjects to maintain a location in spatial working memory. This hypothesis suggests that the difference in activity in the left IPS between lasing and unlasting cue conditions is the neural correlate of spatial working memory.. Keywords: Posner task; Visual-spatial attention; Intraparietal sulcus (IPS); Functional magnetic resonance imaging (fMRI); Spatial working memory. 9.

(18)

(19) 2.1 Introduction. 2.1 Introduction Voluntary orienting of visual-spatial attention is often studied within the context of the Posner task paradigm [1]. In this paradigm, a spatially informative symbolic central cue is presented to indicate the location of an upcoming peripheral target. Targets appearing in predicted location (valid) are detected more rapidly and accurately than those that are not (invalid) [1,5,19]. In the original Posner task [1], cues appeared centrally on the screen and remained present until target offset (we term this the “lasting cue” in the present study). In other studies [3,19,20], cues appeared centrally and then disappeared, followed by varying interstimulus intervals (ISIs) before target onset (we term this the “unlasting cue” in the present study). Over the years, a considerable number of functional magnetic resonance imaging (fMRI) studies have revealed a dorsal fronto-parietal network involved in controlling visual-spatial attention using this paradigm [4,20,21]. This network primarily consists of the intraparietal sulcus (IPS) and the frontal eye field (FEF). Using unlasting cues, Coull and Nobre found significant activation of the right IPS in response to valid targets following highly predictive cues (80% validity) [3]. However, results of a different study by Doricchi et al. revealed no activation in the right IPS under the same conditions using lasting cues [5]. Based on these conflicting results, we believe that different mechanisms are involved in the processing of lasting and unlasting cues in the fronto-parietal network. However, to our knowledge, there has been no direct comparison of the cerebral correlates underlying the perception of the two types of cues, and the neural mechanisms responsible for these potentially divergent types of perception remain unclear. Our aim was to reveal the brain regions activated by lasting and unlasting cues in the Posner task. Furthermore, we aimed to explore differences in brain activity elicited by processing the two types of cues with the goal of determining what underlies these discrepancies. To address these objectives, we designed an fMRI experiment including two tasks based on the Posner paradigm—one with. 11.

(20) 2. Lasting and unlasting cue. lasting cues and the other with unlasting cues.. 2.2 Materials and Methods. 2.2.1 Subjects Nineteen healthy male volunteers (ages 21-32; mean 22.6) took part in the fMRI experiment. All subjects had normal or corrected-to-normal vision. The study was approved by the ethics committee of Okayama University, and written informed consent was obtained before the study.. Figure 2.1. Diagrammatic representation of the visual display and trial design. Subjects maintained fixation on the central cross throughout the duration of the experiment. Cues (unidirectional arrow) indicated the appearance of an upcoming target. The target (letter X) appeared in one of the boxes. Cues correctly indicated the location of the upcoming target (valid trial) in 90% of the trials.. 2.2.2 Stimuli and experiment paradigm We used two variations of the Posner task [1]. Stimuli were presented through a projector onto a. 12.

(21) 2.2 Materials and Methods. paper screen located in front of the subjects’ feet. Subjects viewed the screen via a 45 degree angled mirror attached to the head-coil of the MRI setup. Spatial arrangement, timing, and events of the different trial types used in the two tasks are reported in Figure 2.1. The fixation display in both tasks consisted of a central fixation cross (size 1°×1°) and 2 peripheral boxes (size 4.4°×4.4°), one centered 7°to the left and the other 7°to the right of the central fixation. The subject was presented with an arrow (size 5.8°×5.6°) in the center of the visual field pointing to the right or left, serving as a cue. The target stimulus was the letter X (size 4.4°×4.4°) appearing in one of the peripheral boxes. Each trial began with the fixation display followed by an arrow appearing at the center of the visual field. Subjects were instructed to attend to the cue information in preparation for responding to the upcoming target. In the lasting cue task, the arrow was presented for 200, 400, or 800 ms to prevent temporal orienting. In the unlasting cue task, the arrow was presented for 200 ms and followed by an ISI of 200, 400, or 800 ms. The target appeared for 100 ms on the side indicated by the arrow 90% of the time (valid trial) and on the opposite side 10% of the time (invalid trial). Subjects were instructed to indicate whether a target appeared in the left or right box by pressing the left or right key with the forefinger or middle finger of their right hand, respectively. The duration of each trial was three seconds. There were equal numbers of left and right directional cues and sided targets. Each subject participated in two sessions—one for the lasting cue task and the other for the unlasting cue task. The experimental details were explained to each subject prior to the MRI scan. In total, there were 120 trials for each session. Subjects were instructed to hold their gaze on the central fixation cross throughout the trial and to press the key with high accuracy and speed.. 2.2.3 fMRI scanning All subjects were imaged using a 1.5 T Philips scanner vision whole-body MRI system (Okayama University Hospital, Okayama, Japan), which was equipped with a head coil. The imaging area. 13.

(22) 2. Lasting and unlasting cue. consisted of 32 functional gradient-echo planar imaging (EPI) axial slices (voxel size=3×3×4 mm3, TR=3000 ms, TE=50 ms, FA=90°, 64×64 matrix) that were used to obtain T2*-weighted fMRI images in the axial plane. For each task, we obtained 124 functional volumes and excluded the first 4 scans from analysis. Before the EPI scan, a T1-weighted 3D MP-RAGE sequence was acquired for anatomical alignment (TR=2300 ms, TE=2.98 ms, TI=900 ms, voxel size=1×1×1 mm3).. 2.2.4 fMRI data analysis Data were analyzed using Statistical Parametric Mapping software SPM8 (Wellcome Department of Cognitive Neurology, London, UK; http://www.fil.ion.ucl.ac.uk/spm) running under the MATLAB environment (The Math-Works Inc., Natick, MA) [22]. To correct for differences in slice acquisition time, all images were synchronized to the middle slice. Subsequently, images were spatially realigned to the first volume and normalized to the MNI template supplied with SPM8 (resampled to 2×2×2 mm voxels) [23] using each subjects’ anatomic T1 volume, which had been coregistered to the mean functional image. Functional data were then spatially smoothed with an isotropic Gaussian Kernel of 8-mm full-width half-maximum (FWHM). The time series at each voxel for each subject were high-pass filtered at 128 s to remove low-frequency artifacts and temporally corrected for autocorrelations using the AR(1) model in SPM8. We used a 2-level analysis procedure for statistical inference based on the General Linear Model. In the first-level analysis, the time series of each subject were best fitted at every voxel using the onset timings of the 2 trial conditions: valid lasting cue (VL) and valid unlasting cue (VUL). All stimulus-functions (delta functions) were convolved with the SPM8 canonical hemodynamic response function (HRF) [24]. Regionally specific effects of the 2 trial conditions were tested using linear contrasts, resulting in 2 contrast images per subject. In the second-level analysis, a paired t-test was performed on these individual contrast images to evaluate differences between VUL and VL.. 14.

(23) 2.3 Results. Significantly activated voxels were identified based on a height threshold of p<0.01 uncorrected. Based on the contrast of unlasting versus lasting cue trials, we additionally conducted region of interest (ROI) analyses. We identified one sphere, with a radius of 8 mm, centered at the voxel of peak activation in the left IPS.. Table 2.1. Mean reaction times (ms) for the four experimental conditions. Standard errors of the mean are shown in parentheses. Valid trials. Invalid trials. Lasting cue. 375.37 (14.63). 434.69 (14.76). Unlasting cue. 370.28 (13.94). 446.26 (12.55). 2.3 Results. 2.3.1 Behavioral results All subjects responded to the task with accuracy above 95%. Reaction times (RTs) faster than 100 ms or slower than 1000 ms were excluded from analysis. Mean RTs for the four conditions (valid lasting, invalid lasting, valid unlasting and invalid unlasting) are shown in Table 2.1. The mean RTs for each subject were entered into a repeated measures ANOVA with the factors cue condition (lasting, unlasting) and validity condition (valid, invalid). The results indicated that: (1) RTs were not different in the 2 cue-type conditions (F(1,18)=0.276, P=0.606); (2) the main effect of cue validity was significant (F(1,18)=77.923, P<0.0001); (3) the cue-type × cue validity interaction was also significant (F(1,18)=12.653, P=0.002). The significant cueing effect (invalid versus valid) indicated that subjects oriented their attention based on the cue.. 15.

(24) 2. Lasting and unlasting cue. Figure 2.2. The activated brain regions from the group analysis of the lasting cues. L=left, R=right.. Figure 2.3. The activated brain regions from the group analysis of the unlasting cues. L=left, R=right.. 16.

(25) 2.3 Results. Table 2.2. Brain regions elicited by valid lasting cues. Cluster size. Hemisphere. Brain region. Brodmann area. 6622. L. Middle occipital gyrus. 18/19. R. Cerebellum. L. Inferior occipital gyrus. L. Cerebellum. L. Calcarine sulcus. L 189. MNI coordinates. Z-score. X. Y. Z. -50. -74. 0. 4.87. 28. -54. -24. 4.80. -32. -82. -6. 4.23. -20. -56. -20. 4.01. 18. -2. -86. -14. 3.94. Lingual gyrus. 17. -20. -96. -16. 3.68. R. Supramarginal gyrus. 40. 50. -24. 32. 4.42. 95. L. Parahippocampal gyrus. 35. -24. -32. -10. 4.17. 4729. L. Postcentral gyrus. 3/40/43. -38. -24. 44. 4.16. L. Inferior parietal lobule. 2. -56. -22. 50. 4.15. L. Precentral gyrus. 4. -32. -26. 64. 4.11. R. Supplementary motor area. 6. 4. 2. 66. 4.11. 140. R. Inferior temporal gyrus. 37. 48. -66. -8. 3.01. 44. R. Putamen. 32. -6. 8. 3.68. 285. R. Postcentral gyrus. 2. 48. -36. 64. 3.67. R. Superior parietal lobule. 7. 32. -52. 64. 3.15. 66. R. Thalamus. 8. -2. 0. 3.58. 399. R. Postcentral gyrus. 43. 66. -10. 32. 3.58. R. Precentral gyrus. 6. 64. 6. 26. 3.25. 25. L. Middle temporal gyrus. 21. -48. -42. 0. 3.47. 125. L. Precuneus. 7. -10. -64. 68. 3.29. 58. R. Frontal eye field. 6. 50. 0. 42. 3.20. 142. R. Calcarine sulcus. 18. 26. -102. 2. 2.95. R. Inferior occipital gyrus. 18. 26. -94. -2. 2.86. 40. R. Middle frontal gyrus. 46. 32. 44. 18. 2.87. 19. R. Superior temporal gyrus. 22. 64. -40. 18. 2.71. 18. Abbreviation: L=left; R=right.. 17.

(26) 2. Lasting and unlasting cue. Table 2.3. Brain regions elicited by valid unlasting cues. Cluster size. Hemisphere. Brain region. Brodmann area. 838. R. Calcarine sulcus. R. Inferior occipital gyrus. R. Cerebellum. R. Fusiform gyrus. 163. L. Cerebellum. 462. L. Angular gyrus. L. Z-score. X. Y. Z. 17. 14. -94. -2. 4.37. 19. 34. -76. -10. 3.90. 14. -82. -22. 3.59. 30. -64. -4. 3.43. -42. -46. -32. 3.98. 39/40. -44. -66. 36. 3.75. Intraparietal sulcus. 7. -36. -62. 52. 2.97. L. Middle temporal gyrus. 19. -50. -64. 14. 2.83. L. Supramarginal gyrus. 40. -58. -52. 28. 2.75. 135. L. Parahippocampal gyrus. 28/35. -20. -22. -18. 3.58. 304. L. Cuneus. 31. -6. -74. 22. 3.55. L. Calcarine sulcus. 31. -10. -70. 12. 3.21. L. Superior occipital gyrus. 31. -20. -64. 26. 3.02. 230. L. Inferior occipital gyrus. 19. -36. -68. -8. 3.46. 32. R. Hippocampus. 40. -20. -14. 3.39. 130. L. Paracentral lobule. 4. -12. -36. 70. 3.32. L. Postcentral gyrus. 3. -22. -36. 66. 3.13. 58. L. Fusiform gyrus. 19. -26. -50. -10. 3.18. 165. L. Precentral gyrus. 4/6. -32. -18. 50. 3.14. 22. R. Thalamus. 12. -32. 6. 2.98. 48. R. Parahippocampal gyrus. 28/35. 22. -18. -22. 2.86. 132. L. Lingual gyrus. 17/18. -18. -96. -12. 2.81. L. Middle occipital gyrus. 18. -22. -92. 0. 2.80. 40. R. Angular gyrus. 39. 46. -70. 34. 2.78. 18. R. Intraparietal sulcus. 7. 38. -66. 54. 2.73. 34. L. Inferior parietal lobule. 40. -48. -50. 42. 2.63. 19. L. Frontal eye field. 6. -40. 2. 46. 2.62. Abbreviation: L=left; R=right.. 18. MNI coordinates. 18/19/37.

(27) 2.4 Discussion. 2.3.2 fMRI results The results of whole brain analysis are shown in Table 2.2 and 2.3 (p<0.01 uncorrected, extent threshold 18 voxels). Valid lasting cues activated the left middle occipital gyrus (MOG), bilateral cerebellum, bilateral inferior occipital gyrus (IOG), bilateral calcarine sulcus (CaS), left lingual gyrus (LG), right supramarginal gyrus (SMA), left parahippocampal gyrus (Parahipp), bilateral postcentral gyrus (PoG), left inferior parietal lobule (IPL), bilateral precentral gyrus (PrG), right supplemental motor area (SMA), right inferior temporal gyrus (ITG), right putamen, right superior parietal lobule (SPL), right thalamus, left middle temporal gyrus (MTG), left precuneus (PCU), right frontal eye field (FEF), right middle frontal gyrus (MFG) and right superior temporal gyrus (STG) (Figure 2.2 and Table 2.2). Valid unlasting cues activated bilateral CaS, bilateral IOG, bilateral cerebellum, bilateral fusiform gyrus (FG), bilateral angular gyrus (AG), bilateral intraparietal sulcus (IPS), left MTG, left SMG, bilateral Parahipp, left cuneus, left superior occipital gyrus (SOG), right hippocampus (Hipp), left paracentral lobule (PCL), left PoG, left PrG, right thalamus, left LG, left MOG, left IPL and left FEF (Figure 2.3 and Table 2.3). Direct comparison of the effects of the valid unlasting and valid lasting cues revealed that, compared with valid lasting cues, valid unlasting cues elicited significantly more activation of the left IPS (Figure 2.4).. 2.4 Discussion. 2.4.1 Brain activation during lasting cue tasks and unlasting cue tasks Substantial existing evidence suggests that visual spatial attention is mediated through a monosynaptically interconnected network including not only the visual cortex [25,26] but also the frontal, parietal and cingulate cortices [3,5,20,27-29]. In the present study, we investigated brain regions activated by valid lasting and unlasting cues. Our fMRI data reveal that valid lasting cues. 19.

(28) 2. Lasting and unlasting cue. Figure 2.4. Brain regions showing significant increases in BOLD contrast during the unlasting and lasting cue trials (p<0.01 uncorrected, extent threshold 25 voxels). (a) Coronal section demonstrating activity in the left IPS. (b) Bar graph quantifying the parameter estimates for the left IPS for each experimental condition. Error bars represent the standard error of the mean. L=left, IPS=intraparietal sulcus, VL=valid lasting cues, VUL=valid unlasting cues.. evoked BOLD responses in the right FEF, right SPL and bilateral cerebellum areas. By contrast, valid unlasting cues triggered BOLD responses in the left FEF, bilateral IPS and bilateral cerebellum (Figure 2.2 and 2.3, Table 2.2 and 2.3). Based on neuropsychological and neuroimaging evidence, these regions of activation are consistent with theoretical proposals of a dorsal fronto-parietal network for attention. More specifically, Coull and Frith investigated the spatial and nonspatial aspects of attention and working memory. These authors suggested that the right IPS is recruited for both spatial and nonspatial attention and working memory [30]. Corbetta and colleagues. 20.

(29) 2.4 Discussion. demonstrated that activity in the FEF is not only linked with the execution of saccadic eye movement but also involved in voluntary orienting of visual-spatial attention [31]. Other studies have demonstrated transient neural activation in the SPL during spatial attention shifts [32,33]. Moreover, the function of the cerebellum in attention has recently come to light, and available data suggest that the posterior cerebellum supplies a temporal signal to cortical networks involved in spatial orienting [34]. Considering these previous findings, the results of current study suggest that both lasting and unlasting cues efficiently activate voluntary spatial attention.. 2.4.2 Different left IPS activation triggered by lasting and unlasting cues Another aim of the present study was to identify differences in brain activation caused by processing the two types of cues. Our results demonstrate that valid unlasting cues specifically enhance the BOLD response in the left IPS area compared with valid lasting cues (Figure 2.4). This distinct activation of the left IPS is likely related to its complex brain functions. While the IPS is known to play a critical role in spatial attention [4,30,35], previous studies have also associated function of this region with working memory [36-38]. Lepsien et al. investigated brain areas selectively supporting orienting of spatial attention to locations stored in working memory and found significant signal increases in the right IPS [36]. Corbetta and colleagues demonstrated that the IPS and FEF form a dorsal network that controls the endogenous allocation and maintenance of visuospatial attention (working memory) [37]. Furthermore, several links between spatial attention and spatial working memory have been established, suggesting that they are mediated by largely overlapping networks [38]. On the other hand, in order to eliminate the potentially confounding requirement for working memory involved in investigation of spatial attention using the Posner paradigm, Gitelman et al. presented arrow cues throughout the cue-target stimulus onset asynchrony (SOA; i.e., lasting cues) [4]. The practice of Gitelman prompted us to consider the increased demand on working memory. 21.

(30) 2. Lasting and unlasting cue. caused by the unlasting cue condition. In the unlasting cue task, the cue appeared centrally on the screen and triggered orienting of attention to the cued location. Subsequently, the cue disappeared and was followed by a time interval during which subjects maintained their attention on the cued location. In other words, the time interval is a memory interval, which requires location information to be maintained in spatial working memory. Thus, we hypothesize that this memory-related activity caused enhanced activation of the left IPS in the unlasting cue task compared with lasting cue task. In addition, Doricchi et al. [5] investigated the neural correlates of the spatial and expectancy components of endogenous and stimulus-driven orienting of attention in the standard Posner task with a central long-lasting cue. The authors observed no activation of the right IPS in response to valid targets following highly predictive cues (80% validity). Moreover, when subjected to trials with central unlasting cues, subjects in the study by Coull and Nobre showed significant activation of the right IPS in response to valid targets following highly predictive cues (80% validity) [3]. However, these previous findings do not precisely match our results presented in the current study. Whereas the previous study implied increased activity in the right IPS, we identified increased activity in the left IPS. We hypothesize that this discrepancy may result from the different roles served by left and right IPS, or may be caused by different experimental designs. Future research is required to determine the precise reason for the discrepancy between these two studies.. 2.5 Conclusions Based on a PubMed search, this study is the first to compare the neural correlates of attending to unlasting versus lasting cues in the Posner task. Our results demonstrated that although behavioral data showed no difference between the two cue type conditions, the fMRI data revealed higher activation of the left IPS in the unlasting cue condition. We propose that in the unlasting cue condition, the time interval between the onsets of cue and target requires subjects to maintain a. 22.

(31) 2.5 Conclusions. location in spatial working memory. This hypothesis suggests that the difference in activity in the left IPS between lasing and unlasting cue conditions is the neural correlate of spatial working memory.. 23.

(32)

(33) Chapter 3. fMRI study on orienting of visuospatial attention triggered by different inter-stimulus interval. Summary In experimental paradigms, voluntary orienting of visual-spatial attention is conventionally achieved through the Posner task in which predictive central cues are presented to indicate the location of an upcoming peripheral target, followed by varying inter-stimulus interval (ISI). Previous studies have indicated that the effects of ISI on spatial attention can occur. However, to date, brain mechanisms associated with ISI effects remain unclear. We investigated the brain activity changes along with increased ISI. Behaviorally, subjects exhibited no difference in the different ISI conditions. However, neuroimaging data revealed reduced activity in the posterior cingulate cortex (PCC) as ISI becomes longer. We propose that, as time went on, the strength of visual-spatial bias reduced, resulting in decreased PCC activation.. Keywords: Posner task; Visual-spatial attention; Posterior cingulate cortex (PCC); Functional magnetic resonance imaging (fMRI); Visual-spatial bias. 25.

(34)

(35) 3.1 Introduction. 3.1 Introduction Voluntary orienting of visual-spatial attention is often studied within the context of the Posner task paradigm [1]. In this paradigm, a spatially informative cue is presented to indicate the location of an upcoming peripheral target. Targets appearing in predicted locations (valid) are detected more rapidly and accurately than those that are not (invalid) [1]. Usually, different time intervals between cue and target, i.e. inter-stimulus interval (ISI), are utilized to prevent temporal orienting [19,39,40]. The effects of varying ISI on spatial orienting of attention have also been examined and discussed in previous studies [1,6,7]. In a work by Jonides [6], it has been demonstrated that the time course of efficiency changes over a few hundred milliseconds. Moreover, Shulman, Remington and Mclean [7] found the U-shaped function relating reaction time to ISI for all 4 target positions. Subsequently, Posner replicated Shulman et al.’s experiment and suggested that the effects of ISI on spatial attention can occur [1]. In addition, aging effect of ISI has been revealed in study by Greenwood, Parasuraman and Haxby [9]. More recently, a research which tested attentional orienting in patients with unilateral amygdala lesions showed altered ISI effects in patients compared with normal controls [41]. The latter two results indicate that ISI effects can be influenced by normal aging [9] and brain damage [41] both of which accompanied by changed brain function. However, to date, neural mechanisms associated with ISI effects even in healthy individuals remain unclear. Our aim is to reveal the brain activity changes along with increased ISI. To address this issue, we designed an fMRI experiment including central cue based on the Posner paradigm, since peripheral cue may entangle movements of the eyes [1].. 3.2 Methods 3.2.1 Subjects 27.

(36) 3. Effect of ISI. Nineteen healthy male volunteers (ages 21-32; mean 22.6) took part in the fMRI experiment. All subjects had normal or corrected-to-normal vision. The study was approved by the ethics committee of Okayama University, and written informed consent was obtained before the study.. Figure 3.1. Diagrammatic representation of the visual display and trial design. Subjects maintained fixation on the central cross throughout the duration of the experiment. Cues (unidirectional arrow) indicated the appearance of an upcoming target. The target (letter X) appeared in one of the boxes. Cues correctly indicated the location of the upcoming target (valid trial) 90% of the time.. 3.2.2 Stimuli and experiment paradigm We used a variations of the Posner task [1]. Stimuli were presented through a projector onto a paper screen located in front of the subjects’ feet. Subjects viewed the screen via a 45 degree angled mirror attached to the head-coil of the MRI setup. Spatial arrangement, timing, and events of the different trial types used in the two tasks are reported in Figure 3.1. The fixation display in both tasks consisted of a central fixation cross (size 1°×1°) and 2 peripheral boxes (size 4.4°×4.4°), one centered 7°to the left and the other 7°to the right of the central fixation. The subject was presented with an arrow (size 5.8°×5.6°) in the center of the visual field pointing to the right or left, serving as a cue. The target stimulus was the letter X (size 4.4°×4.4°) appearing in one of the peripheral boxes. Each trial began with the fixation display followed by an arrow appearing at the center of the visual field. Subjects were instructed to attend to the cue information in preparation for responding to the upcoming target. In each trial, the arrow was presented for 200 ms and followed by an ISI of 200,. 28.

(37) 3.2 Methods. 400, or 800 ms. The target appeared for 100 ms on the side indicated by the arrow 90% of the time (valid trial) and on the opposite side 10% of the time (invalid trial). Subjects were instructed to indicate whether a target appeared in the left or right box by pressing the left or right key with the forefinger or middle finger of their right hand, respectively. The duration of each trial was three seconds. There were equal numbers of left and right directional cues and sided targets. Each subject participated in one session. The experimental details were explained to each subject prior to the MRI scan. In total, there were 120 trials for one session. Subjects were instructed to hold their gaze on the central fixation cross throughout the trial and to press the key with high accuracy and speed.. 3.2.3 fMRI scanning All subjects were imaged using a 1.5 T Philips scanner vision whole-body MRI system, which was equipped with a head coil. The imaging area consisted of 32 functional gradient-echo planar imaging (EPI) axial slices (voxel size=3×3×4 mm3, TR=3000 ms, TE=50 ms, FA=90°, 64×64 matrix) that were used to obtain T2*-weighted fMRI images in the axial plane. For each task, we obtained 124 functional volumes and excluded the first 4 scans from analysis. Before the EPI scan, a T1-weighted 3D MP-RAGE sequence was acquired for anatomical alignment (TR=2300 ms, TE=2.98 ms, TI=900 ms, voxel size=1×1×1 mm3).. 3.2.4 fMRI data analysis Data were analyzed with Statistical Parametric Mapping software SPM8 (Wellcome Department of Cognitive Neurology, London, UK, http://www.fil.ion.ucl.ac.uk/spm) running under the MATLAB environment (The Math-Works Inc., Natick, MA) [22]. To correct for differences in slice acquisition time, all images were synchronized to the middle slice. Subsequently, images were spatially realigned to the first volume and normalized to the MNI template supplied with SPM8 (resampled to. 29.

(38) 3. Effect of ISI. 2×2×2 mm voxels) [23] using each subjects’ anatomic T1 volume, which had been coregistered to the mean functional image. Functional data were then spatially smoothed with an isotropic Gaussian Kernel of 8-mm full-width half-maximum (FWHM). The time series at each voxel for each subject were high-pass filtered at 128 s to remove low-frequency artifacts and temporally corrected for autocorrelations using the AR(1) model in SPM8. We used a 2-level analysis procedure for statistical inference based on the General Linear Model. In the first-level analysis, the time series of each subject were best fitted at every voxel using the onset timings of the 3 trial conditions: ISI of 200ms, 400ms or 800ms. All stimulus-functions (delta functions) were convolved with the SPM8 canonical hemodynamic response function (HRF) [24]. Regionally specific effects of the 3 trial conditions were tested using linear contrasts, resulting in 3 contrast images per subject. In the second-level analysis, these individual contrast images were entered into a 1×3 within-subjects ANOVA (ISI [200, 400, 800]). Correlation for nonsphericity was used to account for possible differences in error variance across conditions [42]. Within this ANOVA, we isolated brain activation that resulted from the following contrasts: (1) ISI of 200ms; (2) ISI of 400ms; (3) ISI of 800ms. Only effects surviving an uncorrected threshold of p<0.01 were interpreted. Furthermore, the main effect (ME) of ISI was also estimated. Only effects surviving an uncorrected threshold of p<0.001 and including 10 or more contiguous voxels were interpreted. Based on the result from ME analysis, we additionally conducted region of interest (ROI) analyses. We identified seven spheres, each with a radius of 4 mm, centered at the voxel of peak activation in the right postcentral gyrus (PoG), left inferior frontal gyrus (IFG), left insular cortex, left supramarginal gyrus (SMG), right posterior cingulate cortex (PCC), left PoG and right calcarine sulcus (CaS).. 30.

(39) 3.3 Results. Table 3.1. Mean reaction times (ms) for the six experimental conditions. Standard errors of the mean are shown in parenthesis. Valid trials. Invalid trials. ISI_200. 364.82 (12.95). 449.17 (10.79). ISI_400. 367.98 (14.56). 448.39 (14.23). ISI_800. 378.03 (15.44). 440.85 (15.05). 3.3 Results. 3.3.1 Behavioral results All subjects responded to the task with accuracy above 95%. Reaction times (RTs) faster than 100 ms or slower than 1000 ms were excluded from analysis. Mean RTs for the six conditions (valid ISI_200, valid ISI_400, valid ISI_800, invalid ISI_200, invalid ISI_400 and invalid ISI_800) are shown in Table 1. The mean RTs for each subject were entered into a repeated measures ANOVA with the factors ISI condition (200, 400 and 800ms) and validity condition (valid, invalid). The results indicated that: (1) RTs were not different in the 3 ISI conditions (F(2,36)=0.098, P=0.86); (2) the main effect of cue validity was significant (F(1,18)=88.085, P<0.0001); (3) the ISI × cue validity interaction was not significant (F(2,36)=3.04, P=0.063). The significant cueing effect (invalid versus valid) indicated that subjects oriented their attention based on the cue.. 3.3.2 fMRI results The results of whole brain analysis are shown in Table 3.2, 3.3 and 3.4 (p<0.01 uncorrected). When ISI was short (200ms), Valid cues activated bilateral fusiform gyrus (FG), right inferior occipital gyrus (IOG), left intraparietal sulcus (IPS), right posterior cingulate cortex (PCC), bilateral precuneus (PCU), bilateral postcentral gyrus (PoG), bilateral precentral gyrus (PrG), left middle. 31.

(40) 3. Effect of ISI. Figure 3.2. The activated regions from the group analysis of the ISI_200 conditions. L=left, R=right.. Figure 3.3. The activated regions from the group analysis of the ISI_400 conditions. L=left, R=right.. 32.

(41) 3.3 Results. Figure 3.4. The activated regions from the group analysis of the ISI_800 conditions. L=left, R=right.. temporal gyrus (MTG), left middle occipital gyrus (MOG), left cerebellum, right insula, right Putamen, right inferior parietal lobule (IPL), right superior parietal lobule (SPL), left inferior frontal gyrus (IFG) and left frontal eye field (FEF) (Figure 3.2). On the condition of ISI_400, Valid cues activated bilateral parahippocampal gyrus (Parahipp), bilateral Cerebellum, left PoG, bilateral PCC, left PrG, left MTG, left IPS, left FG, left PCU and right SPL (Figure 3.3). When ISI was long (800ms), Valid cues activated bilateral Parahipp, bilateral FG, right LG and left MTG (Figure 3.4). In addition, our main effect analysis showed that bilateral PoG, left IFG, left insula, left supramaginal gyrus (SMG), right PCC and right calcarine sulcus (CaS) were significantly associated with varying ISI (Figure 3.5).. 33.

(42) 3. Effect of ISI. Table 3.2. Brain activity under ISI-200 condition. Cluster size. Hemisphere. Brain region. Brodmann area. 8717. R. Fusiform gyrus. R. Z-score. X. Y. Z. 19. 34. -76. -10. 5.50. Inferior occipital gyrus. 17. 14. -94. -2. 5.34. L. Intraparietal sulcus. 39. -44. -62. 40. 3.64. R. Posterior cingulate cortex. 31. 18. -28. 44. 4.87. L. Precuneus. 31. -12. -50. 30. 3.48. L. Postcentral gyrus. 3/43. -66. -10. 18. 4.28. L. Precentral gyrus. 6. -60. 4. 14. 3.25. L. Middle temporal gyrus. 22/37. -50. -48. 0. 4.27. L. Middle Occipital gyrus. 37. -44. -74. 6. 3.13. R. Precental gyrus. 6. 64. 10. 28. 3.78. R. Postcentral gyrus. 64. -16. 40. 3.50. R. Insula. 13. 48. -18. 24. 3.17. L. Fusiform gyrus. 20. -48. -34. -24. 3.34. L. Cerebellum. -40. -54. -34. 2.42. 309. L. Precentral gyrus. -40. -16. 58. 3.48. 147. R. Putamen. 32. 2. 2. 3.47. 40. R. Inferior parietal lobule. 40. 46. -28. 30. 3.06. 52. R. Superior parietal lobule. 7. 38. -64. 52. 2.94. 52. R. Precuneus. 39. 42. -68. 38. 2.88. 21. L. Inferior frontal gyrus. 45. -58. 22. 4. 2.86. 47. L. Frontal eye field. 6. -46. 2. 44. 2.74. 1057. 257. 328. 307. 210. Abbreviation: L=left; R=right.. 34. MNI coordinates. 4/6.

(43) 3.3 Results. Table 3.3. Brain activity under ISI-400 condition. Cluster size. Hemisphere. Brain region. Brodmann area. 240. L. Parahippocampal gyrus. 35. 40. L. Cerebellum. 217. L. Postcentral gyrus. L 251. MNI coordinates. Z-score. X. Y. Z. -18. -20. -18. 3.97. -4. -54. -48. 3.79. 3. -12. -38. 72. 3.24. Posterior cingulate cortex. 31. -22. -34. 48. 2.89. L. Precentral gyrus. 4/6. -24. -18. 56. 3.48. 93. R. Posterior cingulate cortex. 31. 28. -28. 44. 3.31. 26. R. Cerebellum. 20. -48. -48. 3.13. 207. L. Middle temporal gyrus. 19/39. -38. -70. 36. 3.12. 75. R. Parahippocampal gyrus. 35. 20. -22. -16. 3.01. 48. L. Intraparietal sulcus. 7. -34. -60. 54. 2.94. 22. L. Fusiform gyrus. 19. -40. -82. -10. 2.85. 112. L. Posterior cingulate cortex. 30/31. -20. -64. 26. 2.84. L. Precuneus. 31. -8. -70. 24. 2.60. R. Superior parietal lobule. 7. 40. -66. 54. 2.78. 21. Abbreviation: L=left; R=right.. Table 3.4. Brain activity under ISI-800 condition. Cluster size. Hemisphere. Brain region. Brodmann area. 96. R. Parahippocampal gyrus. 105. L. 35. MNI coordinates. Z-score. X. Y. Z. 37. 30. -44. -14. 2.35. Fusiform gyrus. 19. -38. -84. -10. 3.04. R. Fusiform gyrus. 19. 34. -78. -12. 2.95. R. Lingual gyrus. 18. 28. -78. -6. 2.87. 26. L. Parahippocampal gyrus. 35. -18. -26. -18. 2.87. 18. L. Middle temporal gyrus. 39. -40. -70. 34. 2.74. Abbreviation: L=left; R=right.. 35.

(44) 3. Effect of ISI. 3.4 Discussion. 3.4.1 Brain activation in different ISI conditions Substantial existing evidence suggests that visual spatial attention is mediated through a monosynaptically interconnected network including not only the visual cortex [25,26] but also the frontal, parietal and cingulate cortices [3,5,20,27-29]. In the present study, we investigated brain regions activated by varying ISI. Our fMRI data reveal that valid cues evoked BOLD responses in the left FEF, left IPS, right SPL and left cerebellum areas on the condition of ISI_200. As time went on, decreased activity throughout frontal, parietal and occipital cortex could be observed clearly (Figure 3.2, 3.3 and 3.4, Table 3.2, 3.3 and 3.4). Based on neuropsychological and neuroimaging evidence, these regions of activation are consistent with theoretical proposals of a dorsal fronto-parietal network for attention. More specifically, Coull and Frith investigated the spatial and nonspatial aspects of attention and working memory. These authors suggested that the right IPS is recruited for both spatial and nonspatial attention and working memory [30]. Corbetta and colleagues demonstrated that activity in the FEF is not only linked with the execution of saccadic eye movement, but also involved in voluntary orienting of visual-spatial attention [31]. Other studies have demonstrated transient neural activation in the SPL during spatial attention shifts [32,33]. Moreover, the function of the cerebellum in attention has recently come to light, and available data suggest that the posterior cerebellum supplies a temporal signal to cortical networks involved in spatial orienting [34]. Considering these previous findings, the results of current study suggest that central informative cues efficiently activate voluntary spatial attention.. 36.

(45) 3.4 Discussion. Figure 3.5. Brain regions showing main effect of ISI (p<0.001 uncorrected, extent threshold 10 voxels). (a) Sagittal section showing activity in the left SMG, left IFG and left PoG. (b) Sagittal section showing activation in the left insula. (c) Sagittal section showing activity in the right CaS and right PCC. (d) Sagittal section showing activation in the right PoG. (e-k) Bar graphs showing the parameter estimates for the left SMG (e), left IFG (f), left PoG (g), left insula (h), right CaS (i), right PCC (j) and right PoG (k) in each experimental condition. Error bars represent the standard error of the mean. Asterisks indicate significance at p<0.05 (*), p<0.01 (**) and p<0.001 (***) uncorrected. L=left, R=right.. 3.4.2 Decreased PCC activation along with increased ISI While whole brain analysis shows reduced activity throughout frontal, parietal and occipital cortex as ISI becomes longer, the ROI analysis exhibits that increased ISI decreases or tends to decrease activation in right PoG, left insula, left SMG and right PCC (Figure 3.5). In the Posner task, the cue stimulus predicting the location of an upcoming target stimulus can direct subjects’ attention to the cued position and generate visual spatial biases [19]. In a study by Small et al. [19], the generation of. 37.

(46) 3. Effect of ISI. visual spatial biases has been revealed to be a role for PCC. And in the consequent research by Small et al. [40], the authors demonstrated the relationship between the attentional bias and activation in the PCC through monetary incentives. Specifically, Small et al. found that monetary incentives had a significant effect upon the link between the visual spatial bias and activity in the PCC in that degree of bias was consistently associated with greater activation [40]. In our study, at the beginning of one trial, the cue appeared centrally on the screen and triggered orienting of attention to the predicted location. Concurrently, strong visual spatial bias was generated to the cued position owing to the predictiveness of cue. As time went on, the strength of bias reduced, resulting in decreased PCC activation. Therefore, in the current study, our results showed decreased PCC activation along with increased ISI (Figure 3.5).. 3.5 Conclusions In the current study, we investigated the brain activity changes along with increased ISI. Behaviorally, subjects exhibited no difference in the different ISI conditions. However, neuroimaging data revealed reduced activity in the posterior cingulate cortex (PCC) as ISI becomes longer. We propose that, as time went on, the strength of visual-spatial bias reduced, resulting in decreased PCC activation. These observations improve our understandings of the dynamics of visuospatial attention-related network.. 38.

(47) Chapter 4 fMRI study on orienting of visuospatial attention in normal older adults. Summary Advanced aging is accompanied by decline in visuospatial attention. Previous neuroimaging and electrophysiological studies have demonstrated dysfunction of specific brain areas in normal older adults. However, little is known about the age-related changes in communications between brain regions that involved in visuospatial attention. Here, we combined task and rest functional magnetic resonance imaging (fMRI) to investigate the age-dependent alterations of resting-state functional connectivity within the task-related network. Twenty-three young subjects and nineteen elderly subjects participated in this study and a modified Posner paradigm was used to define the region of interest (ROI). Our results showed that marked reduction in the number of connections occurred in elderly subjects but was not uniform throughout the brain: significant loss of communications in the anterior portion of the brain and between the anterior and posterior cerebral cortex, preserved communications in the posterior portion of the brain. Moreover, the older adults exhibited weakened resting-state functional connectivity between supplementary motor area and left anterior insular cortex. These findings suggest that disrupted functional connectivity of the brain network for visuospatial attention in normal older adults may underlay the decline in cognitive performance.. Keywords: Posner task; Visual-spatial attention; Aging effect; Functional magnetic resonance imaging (fMRI); Resting-state. 39.

(48)

(49) 4.1 Introduction. 4.1 Introduction Visuospatial attention is often studied within the context of the Posner paradigm [1]. In this paradigm, a spatially informative cue is presented to indicate the location of an upcoming peripheral target. Targets appearing in predicted location (valid) are detected more rapidly and accurately than those that are not (invalid) [1]. Over the years, a considerable number of functional magnetic resonance imaging (fMRI) studies have revealed a frontoparietal network involved in controlling visuospatial attention using Posner paradigm [3-5]. This network consists of anterior insular cortex (AIC), supplementary motor area (SMA) and dorsolateral prefrontal cortex (DLPFC) as well as intraparietal sulcus (IPS), frontal eye field (FEF) and middle temporal area (MT+) [3-5]. Through advanced aging, decline in various cognitive domains has been commonly observed. Using Posner paradigm, previous studies have shown that elderly subjects responded to the target more slowly than young subjects [9,43,44]. One possible reason for this reduction in visuospatial attention is the dysfunction of specific gray matter areas. Electrophysiological studies have demonstrated that advanced aging has an impact on neural activity in those attention-related brain regions [45]. In addition to changes in task-relevant regions, alterations in communication between the different nodes of brain networks have also been found to be correlated with advanced aging [46-48]. For example, poorer performance in older adults on working memory tasks [47] and emotional memory tasks [46] has been shown to be associated with increases and decreases in connectivity within and/or between task-related networks. Therefore, investigation on interactions between brain areas provides insight into neural mechanisms underlying age-related changes. Recently, resting-state fMRI (rs-fMRI) has become a powerful tool for understanding the functional organization of the human brain [10-12]. Based on the synchrony of spontaneous fluctuations in blood oxygenation level dependent (BOLD) signal from functionally related brain regions, several large-scale coherent spatial patterns, namely resting-state networks (RSNs), have. 41.

(50) 4. Advanced aging. been identified [14]. Age-related changes in resting-state functional connectivity has been found within and/or between RSNs [49], such as default mode network (DMN) [50] and motor network [51]. Besides, age-associated differences in functional connectivity have also been revealed between brain areas pertaining to the visuospatial attention-related network [52]. Specifically, the authors identified ROIs in IPS, FEF and MT+ on the correlation map for a seed in IPS which selected from a previous functional correlation study [13], and observed age-related reduction in functional connectivity between IPS and MT+ [52]. However, each ROI comprised corresponding areas of both hemispheres rather than of one hemisphere, which led to be undetectable in inter-hemispherical connectivity. Furthermore, since SMA and AIC were not included in the analysis, how advanced aging affects the resting-state functional connectivity within the entire system that participates in visuospatial attention is still unclear. In the current study, we investigated age-related changes in resting-state functional connectivity between regions that associated with visuospatial attention by combining task and rest fMRI. A modified version of Posner paradigm was used to determine visuospatial attention-evoked brain activation based on which eleven spherical ROIs (seeds) were defined. These ROIs were centered in the SMA, bilateral AIC, bilateral DLPFC, bilateral FEF, bilateral IPS and bilateral MT+. Then, for each subject, mean time series within each ROI was extracted from the resting-state fMRI data and Pearson correlation coefficients were calculated between any possible pair of ROIs. Finally, age-dependent differences in resting-state functional connectivity were examined using two-sample t-tests, corrected for multiple comparisons.. 4.2 Methods. 4.2.1 Subjects. 42.

(51) 4.2 Methods. Twenty-three healthy young volunteers (ages 21-32; mean 22.7) and nineteen healthy older volunteers (ages 60-78; mean 66.5; MMSE score 29.5 ± 0.1) took part in the fMRI experiment. All subjects had normal or corrected-to-normal vision and reported that they were all right-handed. None of the subjects had a history of neurological or psychiatric dysfunction and experience of neuropsychological experiment. The study was approved by the ethics committee of Okayama University, and written informed consent was obtained before the study. Three young subjects with excessive head movements and one older subject for whom fMRI data acquisition failed were excluded. In addition, three other older subjects, although, failed to record behavioral data, the imaging data of them were considered to be available and analyzed in our study.. Figure 4.1. Diagrammatic representation of the visual display and trial design. Subjects maintained fixation on the central cross throughout the duration of the experiment. Cues (unidirectional arrow) indicated the appearance of an upcoming target. The target (letter X) appeared in one of the boxes. Cues correctly indicated the location of the upcoming target (valid trial) 90% of the time.. 4.2.2 Experimental design Resting-state fMRI data were first recorded with one scan when subjects were instructed to keep. 43.

(52) 4. Advanced aging. their eyes closed, not to fall asleep and not to think of anything in particular. This was followed by one scan during a simple visual spatial attention task consisting of 120 trials (Figure 4.1). Each trial began with the fixation display followed by an arrow appearing at the center of the visual field. This arrow was presented for 200 ms and served as a cue, instructing the subjects to pay attention to the left or right visual field. After an inter-stimulus interval (ISI) of 200, 400, or 800 ms, the target appeared for 100 ms on the side indicated by the arrow 90% of the time (valid trial) and on the opposite side 10% of the time (invalid trial). Subjects were instructed to indicate whether a target appeared in the left or right visual field by pressing the left or right key with the forefinger or middle finger of their right hand, respectively. The duration of each trial was 3000 ms and there were equal numbers of left and right directional cues and sided targets. Subjects were asked to hold their gaze on the central fixation cross throughout the trial and to press the key with high accuracy and speed. Stimuli were presented through a projector onto a paper screen located in front of the subjects’ feet. Subjects viewed the screen via a 45 degree angled mirror attached to the head-coil of the MRI setup.. 4.2.3 Data acquisition All subjects were imaged using a 1.5 T Philips scanner vision whole-body MRI system (Okayama University Hospital, Okayama, Japan), which was equipped with a head coil. The imaging area consisted of 32 functional gradient-echo planar imaging (EPI) axial slices (TR=3000 ms, TE=50 ms, FA=90°, acquisition matrix=80×79, FOV=240 mm2, slice thickness=4 mm, gap=0.5 mm) that were used to obtain T2*-weighted fMRI images in the axial plane. We obtained 176 functional volumes for the resting-state session and 124 functional volumes for the task run. The first 4 images of each functional scan were discarded to allow for equilibration of the magnetic field. After the EPI scans, a T1-weighted 3D magnetization-prepared rapid acquisition gradient echo (MP-RAGE) sequence was acquired (TR=9.4 ms, TE=4.6 ms, FA=10°, acquisition matrix=240×240, voxel size=1×1×1 mm3,. 44.

(53) 4.2 Methods. 200 contiguous axial slices).. 4.2.4 fMRI data analysis 4.2.4.1 Preprocessing The imaging data were analyzed using Statistical Parametric Mapping software (SPM8; Wellcome Department of Cognitive Neurology, London, UK; http://www.fil.ion.ucl.ac.uk/spm) running under the MATLAB environment (Version 7.8; The MathWorks, Inc., Natick, MA). To correct for differences in slice acquisition time, all images were synchronized to the middle slice. Subsequently, images were spatially realigned to the first volume due to head motion. Subjects who had head movements exceeding 3 mm on any axis or rotations greater than 3° were excluded. After the correction, the imaging data were normalized to the Montreal Neurological Institute (MNI) EPI template supplied with SPM8 (resampled to 3×3×3 mm3 voxels). Finally, the normalized images were smoothed with a Gaussian kernel of 8-mm full-width half-maximum (FWHM).. 4.2.4.2 Task activation and ROIs For task run, the time series at each voxel for each subject were high-pass filtered at 128 s to remove low-frequency artifacts and temporally corrected for autocorrelations using the AR(1) model in SPM8. Then statistical analysis, which was based on General Linear Model (GLM), was conducted. If only including the canonical hemodynamic response function (HRF) in GLM, delay-induced amplitude bias will be observed [53]. Hence, in current study, a “derivative boost” was utilized to counteract this effect, as suggested previously [54]. First, regressors were normally generated by convolving the stimulus-function corresponding to each experimental condition (valid cue or invalid cue) which was a sum of delta functions with the SPM8 canonical HRF and its temporal derivative. After orthogonalization and normalization of regressors, the model was fit to the data and beta. 45.

(54) 4. Advanced aging. images were obtained. Then, the “derivative boost” can be calculated for each voxel as follow:. H  sign( ˆ1 ) ˆ12  ˆ22 where. ˆ1. (4.1). and ˆ2 are the least-squares estimates of activation amplitude for the canonical HRF. and temporal derivative term respectively [54-56]. These resulting “combined” images from each group were entered into a second-level one-sample t-test to yield group-level activations. A p<0.05, false discovery rate (FDR) corrected for multiple comparison, was considered significant. Based on statistical parametric map for young group, eleven spherical regions of interest (ROIs) were created with a radius of 8 mm centered at the voxels with maxima local T values in attention-related regions.. 4.2.4.3 Functional connectivity analysis For resting-state run, using the DPARSF software (V2.3; http://rfmri.org/DPARSF) and REST toolkit (V1.8; http://restfmri.net/forum/REST), the preprocessed imaging data were removed the linear trend of time series and were temporally band-pass filtered (0.01-0.08 Hz) to reduce the effects of low-frequency drifts and high-frequency physiological noises. Then, several sources of spurious variance including the six estimated head motion parameters and the average time series in the cerebrospinal fluid and white matter regions were removed from the data through linear regression. For each subject, the mean time series of each ROI, which was defined based on attention-related activation maps, was obtained by simply averaging the time series of all voxels within that region. To measure the functional connectivity among regions, we calculated the Pearson correlation coefficients between any possible pair of regional time series, and then obtained a temporal correlation matrix (11×11) for each subject. We applied Fisher’s r-to-z transformation to improve the normality of the correlation matrix. Then, two-tailed one-sample t-tests were performed for all the possible 55 [i.e. (11×10)/2] pairwise correlations across subjects according to each group. 46.