知觉学习效果的影响因素、模型与应用研究

知觉学习反应了大脑可塑性，在视觉功能损害恢复方面具有潜在重要价值，是认知神经科学研究前沿和热点之一。过往研究集中探讨了知觉学习的位点和机制并取得了丰硕成果，但对如何理解和预测学习效果这一兼具理论意义和应用价值的基本问题还欠缺深刻认识。本研究首次系统考察了重要实验设定（难易、反馈属性等）和调节因素（奖赏、电刺激等）对典型视觉任务学习效果（速度、幅度、传递性和保持性）的影响特性，一方面深入分析不同因素及其对应的监督、无监督和强化学习原则在学习效果中的贡献，另一方面建立针对不同视觉功能的综合优化训练方案，并以此为基础收集较大规模样本的心理、行为、生理等多维度指标数据，建构个体学习效果预测模型，最后在弱视病人开展训练方案和学习效果预测模型的验证研究。相关数据和模型不但可加深我们对知觉学习规律、机制和大脑可塑性等理论问题的理解，对如何更有效恢复视功能受损病人的视觉功能也具有借鉴意义和应用前景。

Perceptual learning, reflecting brain plasticity and possessing great potential in in the recovery of visual impairment, has long been one of the frontiers and hotspots in cognitive neuroscience. Past research focused on the loci and mechanism of perceptual learning and achieved fruitful results, but there is still a lack of in-depth comprehension of how to predict and understand learning outcomes, a fundamental question of both theoretical significance and practical value. For the first time, the present study systematically investigated the impact of important experimental settings (e,g. task difficulty, with and without feedback, authenticity of feedback) and modulating factors (e.g. rewards, electrical stimulation) on the practicing outcomes (speed, magnitude, transferability and retention) in typical visual tasks, established a comprehensive optimization training program for different visual functions, and dissected the contribution of supervised, unsupervised and reinforcement principles in perceptual learning. By collecting psychological, behavioral, and physiological data and training a large sample of subjects with the optimized paradigm, we constructed a learning effect prediction model at individual level. We also verified the validity of the program and learning effect prediction model in amblyopic patients. Relevant data and models can not only deepen our understanding of perceptual learning laws, mechanisms, and brain plasticity, but also have practical significance on how to more effectively restore the deficient visual functions in patients with visual impairment.