基于贝叶斯概率估计算法的多维度周边对比敏感度函数检测方法的建立及验证

Establishment and verification of the multi-dimensional peripheral contrast sensitivity function measurement based on Bayesian probability estimation algorithm

目的建立基于贝叶斯概率估计算法的多维度周边快速对比敏感度函数(pqCSF)检测方法,并评估此检测方法的可行性和准确性。方法采用横断面研究设计,2017年9月至2018年3月于中山大学中山眼科中心非连续纳入12名健康正视眼受试者19眼,平均年龄(22.92±2.91)岁,平均球镜度和柱镜度分别为(-0.34±0.52)D和(-0.30±0.42)D,裸眼视力均≥1.0。基于贝叶斯概率估计算法,通过对比敏感度峰值γ_(max)、空间频率峰值f_(max)、带宽β和低对比度截距δ来快速描绘全空间频率的对比敏感度函数(CSF)曲线,建立了(pqCSF)检测方法。所有受试者均在上方、下方、颞侧和鼻侧4个方向6°、12°、18°和24°离心率的16个周边视野位置(其中周边18°的颞侧视野接近生理盲点,不纳入研究)进行pqCSF检测。比较各个视野位置的CSF曲线下面积(AULCSF)以及19个空间频率(以对数单位等间隔分布)所对应的对比敏感度。结果各方向内随着离心率不断增加,AULCSF逐渐降低,不同离心率间AULCSF两两比较,差异均有统计学差异(均P<0.05)。周边6°、12°和24°水平方向(鼻侧+颞侧)AULCSF均较垂直方向(上方+下方)AULCSF大,差异均有统计学意义(均P<0.05)。CSF曲线显示,随着与中心凹距离的增加,pqCSF、AULCSF及高频截止频率都有不同程度的降低,并且数据标准差也逐渐增大。结论pqCSF检测可以描绘出周边大范围视野较完善的周边CSF曲线图,较好地反映周边视功能质量。

Objective To evaluate the feasibility and accuracy of a multi-dimensional peripheral quick contrast sensitivity function(pqCSF)measurement established based on Bayesian probability estimation algorithm.Methods A cross-sectional study was conducted.Nineteen eyes of 12 healthy emmetropic subjects in Zhongshan Ophthalmic Center of Sun Yat-sen University from September 2017 to March 2018 were included,with an average age of(22.92±2.91)years.The average spherical power and cylindrical power were(-0.34±0.52)D and(-0.30±0.42)D,respectively,and the average uncorrected vision acuity was≥1.0.Based on the Bayesian probability algorithm,the peak contrast sensitivity γ_(max),the peak spatial frequency f_(max),the bandwidth β and the low contrast intercept δ were used to quickly describe the contrast sensitivity function(CSF)curve of the full spatial frequency through multi-dimensional pqCSF method.The 16 peripheral visual field positions of all subjects were tested at 6°,12°,18°and 24°eccentricity of the superior,inferior,the temporal and nasal visual field by the pqCSF method,but the 18° eccentricity of temporal field,which was near the physiological blind spot,was excluded.The area under Log CSF(AULCSF)of different peripheral visual fields and the Log CSF of 19 spatial frequencies(distributed at equal intervals in logarithmic units)were compared.This study followed the Declaration of Helsinki,and the study protocol was approved by an Ethics Committee of Zhongshan Ophthalmic Center of Sun Yat-sen University(No.2018KYPJ017).Written informed consent was obtained from each subject prior to any examination.Results With the increase of eccentricity in different visual fields,the AULCSF decreased gradually,and there were significant differences in AULCSF between different eccentricities(all at P<0.05).The AULCSF of the nasal and temporal visual field at 6°,12° and 24° eccentricity was significantly larger than that of the superior and inferior visual field(all at P<0.05).As the distance from the fovea was increased,the pqCSF,the AULCSF,and the high-frequency cutoff were all decreased,and the standard deviation of AULCSF was increased gradually.Conclusions The pqCSF method can depict a relatively complete peripheral CSF curve of a wide peripheral visual field,and reflect the function quality of the peripheral vision comprehensively and accurately.