| Part Ⅰ:Magnocellular and Parvocellular Pathway Impairment in Early-Stage Glaucoma:a fMRI studyPurpose.This study used functional magnetic resonance imaging (fMRI) to sketch the function of magnocellular and parvocellular pathway in normal and early-stage open-angle glaucoma subjects and to verify the selective functional loss hypothesis.Methods.According to Modified Glaucoma Stage System (GSS), early-stage open-angle glaucoma (OAG) and age, gender-matched normal subjects were enrolled in our study. OAG included primary open-angle glaucoma (POAG) and normal-tension glaucoma (NTG). Isoluminant red/green counterphase-flickering checkerboard with high spatial frequency (12cycles in tangential direction,6.5cycles in radial direction), low temporal frequency (0.5Hz) was applied to activate parvocellular pathway (P stimulus) and achromatic counterphase-flickering checkerboard with low spatial frequency (6cycles in tangential direction,3cycles in radial direction), high temporal frequency (10Hz) was used to outline the magnocellular pathway (M stimulus). Early visual processes such as lateral geniculate body (LGN), superior colliculus (SC) as well as high-level processes such as middle temporal (MT) region, occipital-ventral-temporal (OVT) clusters were selected as Regions of Interests (ROIs) due to their proven bias to either pathway. LGN ROI was defined by hemifield checkoard counterphase-flickering at7.5Hz and MT ROI defined by moving/stationary ring stimulus. The fMRI images were analysed by SPM, AFNI and Brainvoyager QX. SPSS16.0was used to perform further analysis.Results.Eighteen patients with early-stage open-angle glaucoma (ten patients with POAG and eight patients with NTG) and eighteen normal subjects participated in this study. Age, gender, refractive error showed no significant differences among normal group, POAG group, NTG group. Mean defect (MD) parameters of test eye in Octopus perimetry had no significant difference between POAG group (1.923±0.335) and NTG group (2.525±0.460). The volumn of LGN ROI was199.7±11.36mm3in normal group,147±7.80mm3in OAG group,140.8±9.76mm3in POAG group and156.0±12.66mm3in NTG group. Independent t-test showed significant difference bwteen normal and OAG group (t=3.820, p<0.01) and no difference between POAG and NTG group (t=0.9672, p=0.3403). Acorrding to the GLM analysis, beta values for each voxel to M stimulus and P stimulus were extracted. We defined each voxel as M-perference voxel when beta values difference (M-P) above zero while as P-perference voxel when beta values difference (M-P) below zero. In normal group, the relative proportions (the proportion of M-perference voxels:38.35%±4.36%) and the spatial arrangment of M-, P-perference voxels (M-perference voxels were located in ventral and medial part) in LGN were consistent with anatomic datas. The glaucoma group showed the similar spatial arrangment but decreased ratio of M-perference voxels (30.74%±5.19%) in LGN. Based on the spatial arragement of M-and P-perference voxels in normal group, we defined M-part and P-part in the two patient groups. Using ANOVA to compare bold signals between groups in M and P part of LGN, the interaction between stimuli conditions and groups was significant in M-part (F=7.054,p=0.010) but no significance was seen in P-part (F=0.994, p=0.322), which indicated M-part’s response to M stimulus declined most in patient group. SC showed the similar results as M-part of LGN with significant interaction (F=7.55, P<0.01). However in MT and OVT clusters, no significant interaction was seen. According to MD differences between nasal and temporal visual fields, we defined better LGN in each patient corresponding to smaller MD and worse LGN in each patient corresponding to larger MD. Using similar analysis methods above, the interaction between stimuli conditions and groups (better/worse LGN) was significant only in M-part (F=28.581, p<0.001). Pearson analysis showed response difference between M-part and P-part to M stimulus was negatively correlated with MD (r=-0.41797, p=0.011) while response difference between M-part and P-part to P stimulus was not correlated with MD (r=-0.090143, p=0.60109).Conclusions.Using functional magnetic resonance imaging (fMRI) to sketch the two parallel pathway’s function, we demonstrated clear segregation of M and P divisions in human LGN. In subcortical visual processes (LGN, SC) of early-stage open-angle glaucoma, magnocelluar function significantly more impaired than parvocelluar’s. But in high-level visual processes, this defective bias is not obvious. This fMRI data provide the first physiological evidence to support selective function loss in early glaucoma and pave the way for a promising paradigm to measure functional defects of diseases related to the two parallel pathways. Part II:Application of a New Magnocellular Pathway Isolating Paradigm with Parvocellular Pathway Saturated in Preperimetric GlaucomaPurpose.This study explored a new magnocellular pathway isolating paradigm with parvocellular pathway saturated and evaluated functional loss in preperimetric glaucoma with the paradigm.Methods.Eleven eyes (eleven patients) with preperimetric open-angle glaucoma and matched eyes of eleven age, gender-matched normal subjects participated in the study. Glaucoma subjects were tested foveally and peripherally (12°). Speak of peripheral test location, one nasal quadrant and one temporal quadrant for each subject were chosen. Control subjects were tested in matching locations. Contrast sensitivity (reciprocal of contrast threshold) was assessed for direction discrimination of the luminance-modulated grating with low-spatial-frequency (0.5cyc/deg) moving horizontally in three temporal-modulated conditions (3Hz,8Hz,15Hz) on the top of isoluminant red-green grating which was applied to saturate parvocellular pathway. Michelson Contrasts was applied to calculate the contrast of luminance-modulated grating and3-up-l-down staircase strategy was used to obtain contrast threshold in each run. Repeated measures analyses of variance (RMANOVAs) were carried out with test conditions (temporal frequency) and test location (nasal/temporal quadrant) as within-subject factors, groups (normal/glaucoma) as between-subject factors while contrast sensitivity as the dependent variable, and Bonferroni correction for multiple comparisons was used in all post hoc analyses. Receiver operating characteristic curve (ROC) was applied to evaluate the diagnostic efficiency of this paradigm.Results.The contrast sensitivity in both fovea and periphery showed an "inverted V" shape with highest sensitivity in intermediate temporal frequency (8Hz) both in normal and glaucoma subjects. Fovea:The contrast sensitivity in the three temporal frequency conditions:21.89±2.96(3Hz),29.57±2.90(8Hz),20.62±2.02(15Hz) in normal group,22.02±1.51(3Hz),29.85±2.52(8Hz),21.55±1.61(15Hz) in glaucoma group. Compared to the normal group, the contrast sensitivity in all conditions for glaucoma patients was not very much changed. RMANOVAs showed significant difference in temporal frequency (F=20.622, p<0.001), but no significant difference between groups and its interaction with temporal frequency (F=0.041, p=0.960). By Bonferroni correction, significant difference (p<0.05) existed between 3Hz and8Hz,8Hz and15Hz. Periphery:The contrast sensitivity in the three temporal frequency conditions:14.64±0.71(3Hz),24.20±1.45(8Hz),18.94±0.89(15Hz) in normal group,14.35±0.89(3Hz),19.66±1.48(8Hz),14.02±1.13(15Hz) in glaucoma group. The contrast sensitivity was decreased compared to the control group, especially in intermediate and high temporal frequency conditions. RMANOVAs showed statistically significant differences in temporal frequency (F=50.799, p<0.001) and interaction between groups and temporal frequency (F=5.653, p<0.05), but no significant difference between groups (F=2.665, p=0.118). Test location (nasal/temporal quadrant) and its interaction with groups were not significant either. RMANOVAs was used seperately in three temporal frequency conditions, only in the high temporal frenquecy condition (15Hz) significant difference (F=6.112, p=0.023) between groups can be found, which is consistent with magocellular pathway function loss. ROC analysis demostrated the areas under curve (AUCs) were0.522、0.693、0.787in3Hz,8Hz,15Hz of peripheral locations respectively.Conclusion.In preperimetric glaucoma, the decline of contrast sensitivity to stimuli which isolated the magnocellular pathway function can be detected peripherally. The findings of this study implied a new psychophysics paradigm to measure magnocellular function and confirmed the viewpoint that selective evalution of magnocellular function could facilitate the earlier detection of ganglion cell function loss in glaucoma before the development of visual field defects with standard clinical perimetry. Part III:Roles of Magnocellular and Parvocellular Pathways in Binocular RivalryPurpose.To investigate the roles of magnocellular and parvocellular pathways in binocular rivalry and to provide a possible explanation for neural mechanism of amblyopia.Methods.Gratings with different colors and forms were presented dichoptically. The stimuli were placed at the upper, right, lower, left side of the fixation cross in sequence trail by trail, with a distance of2°between the centers of the stimuli and the fixation. The diameter of the gratings were2°. The luminance of the grating was made to change sinusoidally, just as that on top of the chromatic grating there were achromatic sine wave drifting. The drifting directions of luminance-modulated gratings dichoptically remained orthogonally. This sinusoidal luminance drifting were presented with spatial frequency of0.5,1,2cycle/degree; temporal frequency of1,8,15Hz; contrast10%or20%(only with spatial frequency of0.5cycle/degree and temporal frequency of15Hz). Eighty trials for each condition were divided into two sessions and there were20sessions in total. In each trial, the stimuli were presented for2seconds. After presentation, subjects were asked to response with mouse for moving direction of luminance wave and the pattern (red/grey circular graing, green/grey radial grating or patchy pattern). Motion intergration was defined when response of motion direction was the three between the orginal moving direction of circular grating and radial grating. Repeated measures analyses of variance (RMANOVAs) were carried out with test conditions (spatial frequency, temporal frequency) as within-subject factors, while trial numbers of intergrated motion as the dependent variable, and Bonferroni correction for multiple comparisons was used in all post hoc analyses. RMANOVAs was also used to compare the differences between10%and20%contrast conditions.Results.Eight subjects (four males and four females) were enrolled in our study. The proportion of perceived single type of pattern (red/grey circular graing or green/grey radial grating) was87.14%±1.90%, while the proportions of perceived motion integration of luminance drifting was30.59%±2.38%. The proportion of motion intergration in each condition was:21.58%±5.65%(0.5cycle/degree,1Hz,10%contrast),43.77%±8.73%(0.5cycle/degree,8Hz,10%contrast),53.03%±8.35%(0.5cycle/degree,15Hz,10%contrast),19.30%±4.78%(1cycle/degree,1Hz,10%contrast),26.27%±5.78%(1cycle/degree,8Hz,10%contrast), 47.53%±7.96%(1cycle/degree,15Hz,10%contrast),15.50%±5.51%(2cycle/degree,1Hz,10%contrast),24.97%±6.30%(2cycle/degree,8Hz,10%contrast),39.74%±7.76%(2cycle/degree,15Hz,10%contrast),35.94%±8.56%(0.5cycle/degree,1Hz,20%contrast). As temporal frquency of luminance drifting became higher, more motion intergration can be detected. There was a decreased tendency of motion intergration with increasing spatial frequency in all temporal frequency conditions. Significant F-values was apparent for temporal frequency (F=8.256, p=0.004) and spatial frequency (F=4.437, p=0.032). Bonferroni correction showed significant differences could be found between8Hz and15Hz (p<0.01),0.5cycle/degree and2cycle/degree (p<0.01),1cycle/degree and2cycle/degree (p<0.01), and marginal significance could be found between1Hz and15Hz (p=0.052). Although there was no significance between10%and20%contrast conditions (F=2.415, p=0.164), compared to trial nums of intergrated motion in20%contrast condition (mean=28.75, SEM=6.850), more motion intergration took place in10%contrast condition (mean=35.75, SEM=6.981).Conclusion.Motion intergration in binocular rivalry bias high temporal modulation, low spatial frequency and low contrast. Our findings support the view that binocular rivalry mainly happens in parvocellular pathway while magnocellular pathway makes little contribution to rivalry, which seems to explain the neural mechanism of selective functinal damage in monocular amblyopia. |
PDF Full Text Request