健康与遗传心理学研究室知识库

双稳态知觉是一种人类在观察存在歧义的外界刺激时主观知觉经验在两种相互排斥的状态中不断切换的现象。双稳态知觉过程被用来作为研究意识的产生以及动态变化过程的一种重要手段已经有上百年的历史, 科学研究发现与双稳态知觉转换相关的大脑区域广泛地分布于额顶叶皮层以及低级视觉皮层，已有的很多研究提出人类大脑的额顶网络在调控双稳态知觉的维持和切换中发挥着主要的作用, 另外一种观点则认为与知觉切换相关联的额叶活动可能只是反应了知觉变化的结果或是执行功能相关的神经活动，但是对于负责物体识别的颞下皮层在双稳态知觉中发挥的作用却鲜有报告。这些严谨的研究提供了很多有用的信息, 但是对于调控双稳态知觉的潜在的神经机制仍然没有一个确切而明晰的结论。研究一中我们采用运动产生的结构(structure from motion, SFM) 双稳态刺激, 使用连续报告范式区分了知觉转换和知觉维持两种条件，结合电技术从神经振荡角度更精确地探究影响知觉切换过程的大脑区域以及其对应的神经振荡特征, 我们也采用了单稳态的刺激作为对照，即通过物理刺激的变化引起被试知觉状态的变化, 全面地探讨双稳态知觉中的自发切换过程相关的神经活动。我们发现双稳态知觉过程相关的神经振荡主要集中在低频的theta(2-8Hz)和alpha(8-13Hz) 频段, 涉及的大脑区域主要包括视觉通路的颞中皮层(middle temporal, MT/V5）, 颞下皮层（inferior temporal cortex, ITC）以及前额叶的额下回(inferior frontal gyrus,IFG), 在alpha 频段上我们观察到有差异的脑区在V5，表现为双稳态相比于单稳态知觉切换减弱的活动, 可能反应的是两种条件的刺激属性之间的差异。在theta 频段上，IFG，ITC 和V5 都表现出双稳态知觉切换特异于知觉维持的活动，由此我们推测这三个区域是双稳态知觉切换相关的区域。通颅内脑过计算这些脑区之间的相位连接关系我们发现, 腹侧视觉通路V5-ITC-IFG 的连接在双稳态知觉切换条件下出现显著的增强, 并且呈现出时序增强的现象，结合更精确的有向性相位梯度指数分析，我们确认了不同脑区之间的信号的流向表现为先从V5 到ITC 然后再从IFG 到ITC，提示双稳态知觉是一个包含了自下而上和自上而下的过程。这个结果证明了腹侧视觉区域在双稳态知觉转换过程中发挥的重要作用, 其中V5 区域的活动反应了一种对歧义刺激的表面重构的过程, 这种对物体或者表面的重构传到ITC 要求对物体进行重新识别，这时候需要IFG 的调控来完成整个知觉变化过程。研究二中我们改用了传统的只有在知觉变化时报告的反应方式，用同样的刺激探索不同的反应方式对双稳态知觉过程的影响，结果发现跟研究一的结果高度相似，同样地在IFG，ITC 和MT/V5 区域表现出知觉转换特异的theta 活动和连接关系。我们也使用了另外一种类型的的双稳态刺激纳克方块(necker cube), 尝试探究不同类型的双稳态刺激是否受到类似脑区的神经活动的调控, 结果发现IFG 和ITC 之间的神经活动以及连接关系的强弱在不同类型的双稳态刺激之间存在较高的一致性，提示不同类型的双稳态知觉过程也存在十分类似的神经加工过程。研究三我们从多变量模式分析(multivariable pattern analysis, MVPA) 的角度，使用常用的线性分类器，探索双稳态知觉转换过程中存在的更一般性的神经活动模式，从另一个角度验证了我们在前面研究中发现的IFG，ITC 和V5 在theta 频段呈现的结果，同时还把知觉转换相关的theta 活动模式扩展到了更广泛的背外侧前额叶，顶上小叶和颞枕交界处等区域，提示了更复杂的潜在的神经机制。预测编码的理论认为双稳态知觉转换的推动因素来源于我们根据先验的预测和歧义的外界刺激之间做出知觉决策时残存的预测误差，我们通过构建这样一个模型能够比较好地拟合了不同类型的双稳态知觉过程，结合记录到的电生理数据我们发现这种预测误差跟视觉区域的theta 功率之间存在显著的相关。综上所述我们发现双稳态知觉的转换过程是由于额叶皮层的预测同步到腹侧视觉区域跟低级视觉皮层的歧义输入之间相互作用产生的残差引起的，这种误差累积到一定程度会引起视觉皮层的表征发生变化，这种新的表征以前馈的方式往上传递引起对物体的知觉的变化并需要额叶的反馈以形成新的知觉状态。

Bistable perception is a phenomenon that human perception continuously alternate between two exclusive status when viewing ambiguous stimulus from the outside world. The process of bistable perception has been used as an important tool to study the generation and variation of consciousness for over a hundred years, scientific research has discovered that bistable perception related brain regions are widely spread over the frontal-parietal cortex and low level visual cortex, and emphasized the critical role of frontal-parietal network in modulating the perceptual alternation, another point of view regards the perceptual switch related activation among frontal areas as a consequence of perception or executive control function, but the role of inferior temporal cortex which is in charge of object recognition has seldom been reported. All these rigorous studies provided us a lot of useful information, but the underlying mechanism of bistable perception is still disputable. In the first study, we utilized a structure from motion (SFM) bistable stimulus and a continuous report paradigm in which we can distinguish between perceptual switch and maintenance condition. Combined with intracranial electroencephalogram (iEEG) we were able to investigate bistable switch related brain regions and neural oscillatory characters with better temporal and spatial precision. We also made a disambiguous control stimulus whose perception status are controled through change of physical property, to achieve a comprehensive understanding of the spontaneous activity during bistable perception. Our results found prevailing theta (2-8Hz) and alpha (8-13Hz) neural oscillatory activities during bistable perception. We discovered IFG, ITC and MT/V5 displayed switch specific theta activities in ambiguous condition but not in unambiguous condition. The phase connectivity showed strengthened connection along the ventral pathway and frontal area in a sequential manner. The results were further validated using a more precise directional phase slope index method that the information flow was first delivered from V5 to ITC, and then IFG to ITC, which indicates both bottom-up and top-down process during perception alternation. Our results proved the importance of ventral visual areas in bistable perception, the activity in V5 may reflect a surface deconstruction process, the deconstructed information is then transmitted to ITC for a re-recognition process which is accompanied by a feedback control of IFG to form the new perceptual status. In the second study, we used a traditional report paradigm that subjects only respond to perception change and a similar SFM stimulus, we replicated most of the foundlings in the first study that IFG, ITC and V5 also revealed bistable switch specific theta oscillation and connectivity regardless of report paradigm in experiment. We further studied the neural representation of another bistable stimuli necker cube, in order to uncovered a more general mechanism of the perception alternation. We also replicated the findings in the previous study in IFG and ITC which both showed similar bistable switch related activation and connectivity change. Our results indicate that different bistable perception may share a common neural mechanism. In the third study, we utilized search light multivariate pattern analysis (MVPA) to explore the possible neural features related to bistable perception, and validate our results of the switch related theta oscillation in the IFG,ITC and V5,and also extended our findings towards wider cortical regions such as dorsal lateral prefrontal cortex,superior parietal lobe and temporal parietal junction,which suggest a more complicated underlying mechanism.According to the existing predictive coding model, the driving force of perception alternation comes from the prediction error calculated from perceptual prediction and sensory evidence. We fitted our behavioral data with this model and correlated possible neural activities with the fitted prediction error signal and found a significant correlation between prediction error and the theta power in V5 and ITC. To sum up, we discovered the cause of bistable perception alternation is the error derived from the prediction of frontal cortex and the ambiguous input of visual cortex, the accumulation of this error will finally change the representation of visual cortex, the new representation will be transmitted up to higher level cortex in a feedforward manner and lead to a re-recognition of object which need the feed-back control of frontal area to form a new percept.