【目的】开展功能性磁共振成像适应(fMRIa)实验,验证立体曲率后效的多级适应机制在人类大脑皮层的16个ROI上的表现。【方法】通过使用水平双目视差定义的随机点立体图作为视觉刺激,设计了6种适应条件以最大限度地分离出不同的适应源。通过测量适应期间和测试期间的BOLD信号变化来评估适应强度。【结果】ANOVA分析结果表示,皮层的背侧视路和腹外侧视路均表现出对形状属性的高级适应。此外,还发现在背侧视路的V7和IPS区存在对平均视差信息的适应,说明这些区域中的低阶视差计算用于精确编码3D形状的深度结构,而腹侧视路则用于形状分类。【结论】立体曲率后效确实是由多级适应所导致。本研究成果可在虚拟现实中立体场景构建时提供对用户体验的优化。

【Purposes】 In this study, a functional resonance imaging adaptation (fMRIa) ex- periment was conducted to examine the multi-level adaptation mechanism in ROIs of the hu- man brain cortex. 【Methods】 Random dot stereograms were defined by horizontal binocular dis- parity as visual stimuli. And six different adaptation conditions were designed to separate differ- ent adaptation sources. Adaptation strengthes were measured as changes in the BOLD signal be- tween adaptation and test periods. 【Findings】 ANOVA results show that both the dorsal and ventral pathways exhibit advanced adaptation effects to shape attributes. Additionally, adapta- tions to average disparity information in the V7 and IPS regions along the dorsal visual pathway exist, indicating that low-level disparity computations in these regions are used to precisely en- code the depth structure of 3D shapes, while the ventral pathway is used for shape classification. 【Conclusions】 These evidences validate the multi-evel adaptation for stereo-curvature aftereffects. The findings can provide guidance for optimizing the construction of stereoscopic scenes in virtual reality.