Multivariate pattern analysis of the human pSTS: A comparison of three prototypical localizers
Multivariate pattern analysis of the human pSTS: A comparison of three prototypical localizers
复制标题
DOI：
10.1016/j.neuropsychologia.2018.10.004
复制
发表时间：
2018-11-01
期刊：
影响因子：
2.6
通讯作者：
Grossman, Emily D.
中科院分区：
文献类型：
--
作者：
Dasgupta, Samhita;Srinivasan, Ramesh;Grossman, Emily D.
人类上级颞沟（pSTS）的后部范围是检测生命力、将能动性归因于他人以及解码目标导向行为的重要皮层区域。理论上的解释将这些认知技能归因于独特的神经群，这些神经群很难凭经验识别（Hein和Knight，2008）。本研究的目的是评估pSTS激活模式的多元统计结构时，查看不同的社会线索。我们确定了一个核心的连接区域pSTS单变量响应与偏好点光生物运动，面孔和简单的动画形状的社会概念的属性。在一个多变量分析中，我们的特点是相似性结构的激活模式后，控制方差的激活配置文件引起的形式和运动功能。我们发现强烈的拮抗激活配置文件之间的社会条件和他们的定位控制，一个预兆，为什么这些典型的定位是如此有效，即使在个别科目。我们还发现了三种核心社会条件之间的独特相似模式。我们的研究结果与Shultz等人（2015）的pSTS功能模型一致，其中存在独立的神经群体用于从身体部位检测生命，而不是从运动中提取有意的线索。
The posterior extent of the human superior temporal sulcus (pSTS) is an important cortical region for detecting animacy, attributing agency to others, and decoding goal-directed behavior. Theoretical accounts attribute these cognitive skills to unique neural populations that have been difficult to identify empirically (Hein and Knight, 2008). The aim of this study is to evaluate the multivariate statistical structure of pSTS activation patterns when viewing different social cues. We identified a core conjunction region on pSTS from univariate responses with preference for point-light biological motion, faces and the attribution of social concepts to simple animated shapes. In a multivariate analysis, we characterized the similarity structure of the resulting activation patterns after controlling for variance in the activation profile elicited by form and motion features. We found strong antagonistic activation profiles between the social conditions and their localizer controls, a harbinger of why these canonical localizers are so effective, even in individual subjects. We also found unique patterns of similarity between the three core social conditions. Our findings are consistent with the Shultz et al. (2015) model of pSTS function in which separate neural populations exist for animacy detection from body parts versus for extracting intentional cues from movement.