Scholarlink

Background <p>Migraine diagnosis remains largely symptom-based, and sleep, anxiety, and depressive symptoms may contribute to clinical heterogeneity. Whether oculomotor metrics provide complementary discriminative information beyond these self-reported symptom scales remains uncertain.</p> Methods <p>In this cross-sectional case–control study, 96 patients with migraine without aura and 96 group-level frequency-matched healthy controls completed four standardized saccadic tasks and questionnaires during the interictal period. Participants exceeding prespecified anxiety, depressive symptoms, or sleep-quality thresholds were excluded. Analyses were therefore conducted within a restricted symptom range. Questionnaire-only, oculomotor-only, and combined L2-regularized logistic regression models were evaluated. Incremental value was assessed using fully nested training-set out-of-fold predictions and paired stratified bootstrap ΔAUROC, with an independent test set reserved for final evaluation.</p> Results <p>In the fully nested out-of-fold analysis, AUROCs were 0.723 for the questionnaire-only model, 0.692 for the oculomotor-only model, and 0.796 for the combined model. The combined model showed a positive but non-significant ΔAUROC versus the questionnaire-only model (0.074; 95% CI − 0.002 to 0.147; <i>p</i> = 0.059). In the independent test set, AUROCs were 0.783, 0.692, and 0.803, respectively. The combined model showed descriptively higher specificity than the questionnaire-only model (85.0% vs 75.0%) but lower sensitivity (63.2% vs 73.7%), with unchanged accuracy (74.4%).</p> Conclusions <p>In this selected, symptom-restricted case–control cohort, oculomotor metrics showed possible complementary discriminative information beyond self-reported sleep, anxiety, and depressive symptom scales, but the primary ΔAUROC comparison did not reach conventional statistical significance. These findings support further evaluation of eye movement assessment as a potential corroborative measure rather than a stand-alone diagnostic test.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Oculomotor metrics and possible complementary discriminative information beyond patient-reported sleep, anxiety, and depressive symptom scales in migraine without aura: a cross-sectional study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月29日

查看全文
  • Yulong Li,
  • Hua Li,
  • Ruiqi Yang,
  • Ling Jin,
  • Xia Zhao,
  • Mingsha Zhang,
  • Lei Chen

摘要

Background

Migraine diagnosis remains largely symptom-based, and sleep, anxiety, and depressive symptoms may contribute to clinical heterogeneity. Whether oculomotor metrics provide complementary discriminative information beyond these self-reported symptom scales remains uncertain.

Methods

In this cross-sectional case–control study, 96 patients with migraine without aura and 96 group-level frequency-matched healthy controls completed four standardized saccadic tasks and questionnaires during the interictal period. Participants exceeding prespecified anxiety, depressive symptoms, or sleep-quality thresholds were excluded. Analyses were therefore conducted within a restricted symptom range. Questionnaire-only, oculomotor-only, and combined L2-regularized logistic regression models were evaluated. Incremental value was assessed using fully nested training-set out-of-fold predictions and paired stratified bootstrap ΔAUROC, with an independent test set reserved for final evaluation.

Results

In the fully nested out-of-fold analysis, AUROCs were 0.723 for the questionnaire-only model, 0.692 for the oculomotor-only model, and 0.796 for the combined model. The combined model showed a positive but non-significant ΔAUROC versus the questionnaire-only model (0.074; 95% CI − 0.002 to 0.147; p = 0.059). In the independent test set, AUROCs were 0.783, 0.692, and 0.803, respectively. The combined model showed descriptively higher specificity than the questionnaire-only model (85.0% vs 75.0%) but lower sensitivity (63.2% vs 73.7%), with unchanged accuracy (74.4%).

Conclusions

In this selected, symptom-restricted case–control cohort, oculomotor metrics showed possible complementary discriminative information beyond self-reported sleep, anxiety, and depressive symptom scales, but the primary ΔAUROC comparison did not reach conventional statistical significance. These findings support further evaluation of eye movement assessment as a potential corroborative measure rather than a stand-alone diagnostic test.