Purpose
To segment the layers of subcortical nuclei (LGN, SC) with fMRI, and to investigate selective pathway damage in glaucoma and amblyopia patients.
Methods
Eighteen early-stage open-angle glaucoma patients, eighteen adult unilateral amblyopia patients and eighteen controls were enrolled in the study. The M and P stimuli were used to segment the M and P functional layers of LGN. The M stimulus was low spatial frequency sine wave patterns at low luminance contrast, and counter-phase flickering at high temporal frequency. The P stimulus was high spatial frequency, isoluminant red/green square wave patterns, reversing contrast at low temporal frequency. MRI data were collected with 3T scanner (Siemens Verio) and were analyzed using Brainvoyger QX.The M layers of the LGN was identified as voxels showing stronger response to the M stimulus than to the P stimulus, and vice versa for the identification of the P layers.
Results
In normal group, the M layers were located more medially, interiorly and anteriorly compare to the location of the P layers, which is consistent with the human LGN anatomy. In glaucoma group, nearly all LGN voxels showed stronger responses to the P stimulus. In the M layers of LGN, a significant interaction was found between the factors of stimuli and groups but there was no interaction between stimuli and groups in the P layers. In amblyopia group, fewer LGN voxels showed strong responses to the P stimulus in the LGNs. In the P layers of LGN, a significant interaction was found between the factors of stimuli and groups but no such interaction in M layers. These results indicate that neuronal responses to the M and P stimuli were selectively reduced in the LGNs of glaucoma patients and amblyopia subjects, that is a greater loss of signals to the M stimulus in the M layers in glaucoma and a greater loss of signals to the P stimulus in the P layers in amblyopia. FMRI signals measured at superior depth showed the same significant interaction as in LGNs of patients. In glaucoma patients, the selective subcortical damage were well correlated with OCT parameters and visual field loss (p<0.05).
Conclusions
Using fMRI, we demonstrated clear segregation of M and P divisions in human LGN noninvasively.This fMRI data provide in-vivo and direct evidence to support selective function loss in the two eye diseases and provide a promising paradigm to measure functional defects related to the two parallel pathways.