[Application of Deep Learning Algorithm in the Grading Assessment of Corneal Fluorescein Staining].

Objective: To explore the application value of applying deep learning (DL) algorithm in the grading assessment of corneal fluorescein staining. Methods: A cross-sectional study was carried out, covering 600 corneal fluorescein staining photos acquired in the Contact Lens Clinic, West China Hospital, Sichuan University between 2020 and 2022. Out of the 600 photos, 500 were used to construct the algorithm and the remaining 100 were used for the validation of the algorithm and a comparative analysis of the difference in grading accuracy (ACC) and the length of diagnostic time between artificial intelligence (AI) and optometry students. One month after finishing the first grading analysis, assessment by AI and optometry students was conducted for a second time and results from the two rounds of assessment were compared to examine the intrarater agreement ( kappa value) of the two analyses. The grading analysis results of 3 experienced optometrists were used as the gold standard in the study. Results: Findings of the cross validation with the complete dataset, the training dataset, and the test dataset showed that ResNet34 had the highest predictive accuracy among four DL models. ResNet34 DL model achieved an accuracy of 93.0%, sensitivity of 89.5%, and specificity of 89.6% in the grading of corneal staining. In the comparison of the grading accuracy of AI and two optometry students, AI showed better accuracy, with the respective grading accuracy being 87.0%, 78.0%, and 52.0% for AI, student 1, and student 2 ( P ACC=0.001). In addition, the average diagnostic time of AI was shorter than that of optometry students ( t AI=1.00 s, t S1=11.86 s, t S2=13.25 s, P t =0.001). In the comparative analysis of the intrarater agreement between the two assessments, AI ( kappa AI=0.658, P AI=0.001) achieved better consistency than the two optometry students did （ kappa S1=0.575, P S1=0.001; kappa S2=0.609, P S2=0.001). Conclusion: Applying deep learning algorithms in the grading assessment of corneal fluorescein staining has considerable feasibility and clinical value. In the performance comparison between AI and optometry students, AI achieved higher accuracy and better consistency, which indicates that AI has potential application value for assisting optometrists to make clinical decisions with speed and accuracy.

Altered cortical morphology of visual cortex in adults with monocular amblyopia.

BACKGROUND: The neural mechanism of amblyopia and its impact on the adult brain remain unclear. This hinders effective treatment for adults with this disease. PURPOSE: To investigate neuroanatomical differences in cortical morphometry between amblyopic adults and healthy controls, and to explore the structural covariance of abnormal morphometric changes. STUDY TYPE: Prospective. POPULATION: Twenty-one amblyopic adults and 34 healthy controls. FIELD STRENGTH/SEQUENCE: 3T MRI, T1 -weighted, MPRAGE sequence. ASSESSMENT: All participants completed ophthalmologic exams to confirm the diagnosis of amblyopia or its absence in the healthy controls, including tests of ocular motility and dilation, fundus exam, autorefraction and synoptophore tests. Cortical volume, thickness, and surface area measurements were obtained using FreeSurfer software. STATISTICAL TESTS: Statistical differences of MRI measures between amblyopic adults and healthy controls were identified using a general linear model with intracranial volume and age as covariates. Monte Carlo simulations were used to correct for multiple comparisons. The structural covariance of abnormal morphometric changes and the relationship between morphometric abnormalities and visual acuity of the amblyopic eye were examined. RESULTS: Compared with healthy controls, amblyopic adults showed reduced cortical volume in left lateral occipital cortex, and decreased cortical thickness in bilateral inferior temporal gyrus and left precentral gyrus (P < 0.05; Monte Carlo corrected). Structural covariance between cortical volume of left lateral occipital cortex and cortical thickness of right inferior temporal gyrus in amblyopic adults was significantly less than in healthy controls (z = 1.73; P < 0.05). DATA CONCLUSION: Our study identified morphological abnormalities in occipital cortex and in temporal and frontal cortex which are projection fields of visual cortex important for processing of visual form and object location information, and disrupted structural covariance of visual cortex with other brain regions in amblyopic patients. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2019;50:1405-1412.

Assessment of Cortical Dysfunction in Patients with Intermittent Exotropia: An fMRI Study.

Neural imaging studies have found the connection between strabismus and brain cortex. However, the pathological mechanisms of intermittent exotropia are still not fully understood. In the present study, changes of binocular fusion related cortices in intermittent exotropia were investigated with blood oxygen level dependent functional magnetic resonance imaging. Activated cortices induced by fusion stimulus were found to be distributed in several regions such as bilateral middle occipital gyrus, bilateral middle temporal gyrus, left superior parietal lobule and so on. Compared with normal subjects, the increased activation intensity was observed in bilateral superior parietal lobule and inferior parietal lobule in subjects with intermittent exotropia. These findings indicate that binocular fusion involves a complicated brain network including several regions. And cortical activities of bilateral superior parietal lobule and inferior parietal lobule compensate for the binocular fusion dysfunction in intermittent exotropia.

Assessment of cortical dysfunction in infantile esotropia using fMRI.

PURPOSE: Functional magnetic resonance imaging (fMRI) is popular in the field of neuroimaging. The aim of this study was to explore the neural basis of infantile esotropia and locate the cortical region of fusion function defects using blood oxygen level-dependent fMRI (BOLD-fMRI). METHODS: Subjects (5 with infantile esotropia and 8 normal) were presented with a visual stimulus (letter E). The BOLD-fMRI was used to measure the visual cortex response as the distance between the E and the eyes changing from 20 to 100 cm using a block experimental design. Data were processed offline using statistical parametric mapping and compared between the 2 groups. RESULTS: Areas activated in the visual cortex of normal subjects were located in the bilateral frontal gyrus and left lingual visual cortex (p<0.001). The BOLD signal in the left cingulate gyrus, bilateral precuneus, and left angular gyrus was higher in the infantile esotropia group than in the controls (p<0.001). None of the visual cortical areas showed higher signal in the control individuals than in those with infantile esotropia (p<0.001). CONCLUSIONS: The bilateral frontal gyrus and left lingual visual cortex regulate normal fusion function in human eyes. In infantile esotropia, the left cingulate gyrus, bilateral precuneus, and left angular gyrus visual cortex may compensate for the fusion dysfunction. These insights may help improve the diagnosis and treatment of strabismus.

[Study on crossed and uncrossed disparity evoked potential of dynamic random dot stereogram].

OBJECTIVE: To analyze the disparity evoked potentials (DEP) of fine disparities, coarse disparities, crossed disparities and uncrossed disparities and provide parameters of impersonal examining stereopsis. METHODS: A software package for generating dynamic random dot stereogram (DRDS) was developed as a visual stimulus to elicit DEP in 30 normal subjects. The DEP of every subject was recorded in different crossed and uncrossed disparity stimuli (4', 8', 15', 23', 30', 45', 53', 60', 72', 87', 102', 124', 150' of arc). RESULTS: (1) the constant negative-positive complex wave was observed in different disparity stimuli. (2) The N wave's latency of -45' was the longest, the N wave's latency of -150' was the shortest among crossed disparities. The N wave's latency of 4', 23' was the longest, the N wave's latency of 124' was the shortest among uncrossed disparities. (3) The amplitude peak of P wave occurred at -23', -60', -150' among crossed disparities. The amplitude peak of P wave occurred at 15', 45', 72' among uncrossed disparities. (4) All characteristic crossed disparities were smaller than the uncrossed disparities. The change orderliness of small crossed disparities was different from that of small uncrossed disparities. CONCLUSIONS: The N wave's latency and the P wave's amplitude can be used as parameters to impersonally examine stereopsis. These results suggest that stereopsis can be divided into fine crossed stereopsis, fine uncrossed stereopsis, coarse crossed stereopsis, and coarse uncrossed stereopsis.