探討潛在腦性視覺功能異常早產兒童與典型發展兒童在國小階段視覺擁擠效應及閱讀任務表現之差異：橫斷面研究

Abstract

腦性視覺損傷（cerebral visual impairment, CVI）是常見於早產兒，主要因大腦中後段視覺路徑損傷造成之視覺功能異常情形，為目前造成已開發與開發中國家兒童低視力的主因之一。腦性視覺功能異常（cerebral visual dysfunction, CVD）是CVI族群中視力沒有受到明顯損傷者，影響個案之視覺訊息處理及眼球動作功能。目前尚缺乏客觀且具有敏感性的評估工具，導致CVD診斷不易。過去研究指出，分析兒童閱讀表現與受其視覺擁擠效應影響之程度可能是有效辨識CVD的策略。本研究透過自行發展之繁體中文擁擠效應閱讀表現測驗(Crowding Effect Reading Performance Test, CERPT)，並結合視知覺能力、眼動與學業表現指標，評估多元指標對辨識CVD之可行性。 本研究招募8至13歲之國小階段之足月與早產兒童，並透過Flemish腦性視覺障礙問卷（The Flemish Cerebral Visual Impairment Questionnaire, FCVIQ）進行分組，分為典型發展組與疑似CVD組。進行正式評估時，使用CERPT、視知覺技巧測驗-第四版(Test of Visual Perceptual Skills 4th Edition, TVPS-4)、有效視野測驗(Useful field of view test, UFOV test) 、基本讀寫字綜合測驗、Higher Visual Function 58-Question Inventory for Parents (HVFQI-58 for Parents)等評估，以了解兒童之閱讀表現、視知覺表現、視覺反應速度、選擇性注意力、分散性注意力、學校與日常生活表現與困擾與視覺行為問題發生之情形，並量化兒童之閱讀表現。進行閱讀測驗時搭配兒童版見臻穿戴式眼動儀（Ganzin Sol eye tracker），實驗流程約1至1.5小時。 統計方面，使用描述性統計呈現參與者的人口學與測驗表現，以卡方檢定檢驗組間族群特質是否有顯著差異。組間之測驗結果比較因樣本數少採用Kruskal-Wallis檢定。CERPT部分先分析不同刺激出現先後順序之學習效應，接續比較組別與隨機組別之交互作用與組別主效果，並以事後分析檢驗差異發生之情境。相關性分析使用Spearman等級相關探討變項之間的關係。 本研究經排除不適用個案後，最終納入36名兒童，包含18名典型發展組，與18名疑似CVD組。兩組在性別、年齡、年級及瑞文氏智力測驗表現上均無顯著差異，惟Suspected CVD組在出生體重與妊娠週數顯著低於TD組。在標準化測驗方面，兩組在學業表現、有效視野測驗及視知覺測驗的所有分項與總分均無顯著統計差異。然而，家長填寫的兩份功能性視覺問卷顯示Suspected CVD組得分顯著高於TD組，指出其在日常生活中存在顯著的功能性視覺困難。在CERPT中，平均閱讀時間與擁擠句的跳視次數未見顯著組間差異。然而，針對完整句子閱讀的眼動分析呈現顯著的組別與刺激次序交互作用（p=0.036*）。事後檢定證實，在完整非擁擠句先出現的條件下，Suspected CVD 組的跳視次數顯著多於 TD 組（p=0.040*），顯示Suspected CVD兒童可能較難從簡單的前導情境中獲益並建立穩定的閱讀策略。相關性分析顯示，功能性視覺問卷得分與CERPT的閱讀時間呈顯著正相關，驗證了功能性視覺缺陷可能會反映在閱讀效率的降低上。值得注意的是，眼跳次數與閱讀時間及其他視知覺指標均無顯著相關，暗示眼動策略可能是一個獨立於視覺認知效率之外的鑑別指標。 本研究旨在比較典型發展兒童與疑似腦性視覺障礙早產兒組在各項視覺功能、視知覺與閱讀表現上的差異。研究結果清晰地劃分了兩個組別的特徵，其差異主要體現在日常生活中的行為觀察與複雜情境下的功能性視覺表現上。此外，標準化視知覺測驗未必能反映疑似 CVD 兒童的真實困難。本研究證實，結合功能性問卷與具擁擠效應之動態閱讀任務，更能有效捕捉 CVD 兒童在策略轉換與複雜情境下的功能性異常情形，為臨床評估提供具實證基礎的鑑別指標。

Cerebral visual impairment (CVI), prevalent among preterm infants, is characterized by visual dysfunction resulting primarily from damage to the posterior visual pathways. It stands as a leading cause of pediatric low vision in both developed and developing nations. Cerebral visual dysfunction (CVD) represents a specific group within the CVI spectrum in which individuals exhibit no significant impairment in visual acuity yet manifest deficits in visual information processing and oculomotor function. However, the current paucity of objective and sensitive assessment tools renders the clinical diagnosis of CVD challenging. Previous literature suggests that analyzing reading performance, specifically the susceptibility to the visual crowding effect, may serve as an effective strategy for identifying CVD. This study employed the self-developed traditional Chinese reading test, the Crowding Effect Reading Performance Test (CERPT), with metrics of visual perceptual skills, oculomotor behaviors, and academic achievement, to evaluate the feasibility of a multidimensional assessment approach for identifying CVD. This study recruited full-term and preterm elementary school-aged children. Participants were classified into a typically developing (TD) group and a suspected CVD group based on the results of the Flemish Cerebral Visual Impairment Questionnaire (FCVIQ). The comprehensive assessment battery included the CERPT, the Test of Visual Perceptual Skills–4th Edition (TVPS-4), the Useful Field of View (UFOV) test, the Battery of the Chinese of Pupils, and the Higher Visual Function 58-Question Inventory for Parents (HVFQI-58 for parents). These instruments were employed to evaluate reading performance, visual perceptual skills, visual processing speed, selective and divided attention, academic and daily functioning, and visual behavioral concerns, as well as to quantify specific reading metrics. During the reading tasks, eye movements were recorded using the Ganzin Sol wearable eye tracker (pediatric version). The entire experimental procedure lasted approximately 1 to 1.5 hours. Descriptive statistics were utilized to present demographic data and test performance. Chi-square tests were conducted to examine significant differences in group characteristics. Due to the limited sample size, the Kruskal-Wallis test was employed to compare assessment results between groups. For the CERPT analysis, the learning effects of stimulus presentation order were first evaluated. Subsequently, the main effects and interaction effects between group and random order were examined, followed by post-hoc analyses to identify specific conditions where differences occurred. Finally, Spearman’s rank correlation analysis was performed to explore the relationships among variables. After excluding ineligible cases, a final sample of 36 children was included in the analysis, comprising 18 in the TD group and 18 in the suspected CVD group. There were no significant differences between the two groups in gender, age, grade level, or intellectual level. However, the suspected CVD group had significantly lower birth weights and shorter gestational ages compared to the TD group. Regarding standardized assessments, no statistically significant differences were found between the groups in academic performance, the UFOV test, or the TVPS-4 across all subtests and total scores. Conversely, results from the two parent-reported visual-related function questionnaires indicated that the suspected CVD group scored significantly higher than the TD group, highlighting substantial functional visual difficulties in daily life. In the CERPT, no significant between-group differences were observed in mean reading time or saccade counts for crowded sentences. However, analysis of oculomotor behaviors during full-sentence reading revealed a significant interaction between Group and Stimulus Order (p=0.036*). Post-hoc tests confirmed that in the "Uncrowded-First" condition, the suspected CVD group exhibited significantly more saccades than the TD group (p=0.040*). This suggests that children with suspected CVD may struggle to benefit from simple priming contexts to establish reading strategies. Correlation analysis demonstrated a significant positive correlation between functional vision questionnaire scores and CERPT reading times, validating that functional visual deficits are reflected in reduced reading efficiency. Notably, saccade counts did not correlate significantly with reading time or other visual perceptual metrics, implying that oculomotor strategy may serve as a distinct discriminative indicator independent of visual cognitive efficiency. The primary objective of this study was to compare differences in visual function, visual perception, and reading performance between typically developing children and preterm children with suspected CVD. The results clearly delineate the distinct characteristics of the two groups, with disparities manifesting primarily in behavioral observations within daily life and functional visual performance under complex contexts. Furthermore, the findings suggest that standardized visual perceptual tests may not fully capture the actual difficulties experienced by children with suspected CVD. This study demonstrates that integrating functional questionnaires with dynamic reading tasks incorporating the crowding effect more effectively captures functional abnormalities related to strategy switching and complex environments. Consequently, this approach provides evidence-based discriminative indicators for clinical assessment.