fMRI time series in orbitofrontal cortex during face-house state-space task.

This dataset contains the preprocessed fMRI time series within orbitofrontal cortex (OFC), the design matrix and event timing information of the task performed in the work of Schuck et al., Neuron 2016: Schuck, Nicolas W., Ming Bo Cai, Robert C. Wilson, and Yael Niv. "Human orbitofrontal cortex represents a cognitive map of state space." Neuron 91, no. 6 (2016): 1402-1412. In the task, participants watched overlapping face and house images and judged the age of one category to be young or old (meaning modern or ancient for houses). The specific rule of which category to judge required participants to keep track of which state out of 16 in total they are in. The result support the theory that orbitofrontal cortex encodes task-relevant hidden states of environments. The data of 24 healthy human participants used in the trial-wise analysis in the paper are included. Except for the first participant who had four successful runs of data, all participants had five runs of data. The fMRI time series were slice-timing corrected and spatially realigned to correct for head motion within and across experiment runs. An evaluation of an algorithm of Bayesian representational similarity analysis (BRSA) against other approaches of representational similarity analysis by Cai, et al. PLOS Computational Biology, 2019[3] was partly based on this dataset. The work suggested that performing RSA based on activity pattern estimated from the same runs suffer from statistical bias in the similarity matrix, and recommended either BRSA or cross-run RSA. Paper referenced: [3] Cai, Ming Bo, Nicolas W. Schuck, Jonathan W. Pillow, and Yael Niv. "Representational structure or task structure? Bias in neural representational similarity analysis and a Bayesian method for reducing bias." bioRxiv (2018): 347260.

本数据集包含Schuck等人2016年发表于《神经元》（Neuron）的研究中所使用的预处理后眶额皮层（orbitofrontal cortex, OFC）功能磁共振成像（functional magnetic resonance imaging, fMRI）时间序列、任务设计矩阵以及事件计时信息。相关原始研究为：Schuck NW, Cai MB, Wilson RC, Niv Y. 人类眶额皮层表征状态空间的认知地图. 神经元（Neuron）, 2016, 91(6): 1402-1412。实验任务中，受试者将观看重叠呈现的人脸与房屋图像，并对其中一类刺激的年龄进行判断：人脸需判断为年轻或年老，房屋则对应判断为现代或古老。本次任务需判断的类别规则，要求受试者追踪自身当前所处的共16种潜在状态中的某一种。本数据集对应的研究结果支持了"眶额皮层会编码环境中与任务相关的隐藏状态"这一理论。本数据集包含了该论文中逐试次分析所使用的24名健康人类受试者的fMRI数据。除第一名受试者拥有4组有效扫描数据外，其余所有受试者均拥有5组有效扫描数据。所有fMRI时间序列均经过了切片时间校正，并针对扫描批次内及跨扫描批次的头部运动进行了空间重对齐处理。Cai等人2019年发表于《公共科学图书馆·计算生物学》（PLOS Computational Biology）的一项研究[3]，对贝叶斯表征相似性分析（Bayesian Representational Similarity Analysis, BRSA）算法与其他表征相似性分析（Representational Similarity Analysis, RSA）方法进行了对比评估，该研究部分基于本数据集完成。该研究指出，基于同一扫描批次估算的神经活动模式进行RSA分析时，相似度矩阵会存在统计偏差，并建议采用BRSA方法或跨扫描批次的RSA方法。引用文献[3]：Cai MB, Schuck NW, Pillow JW, Niv Y. 表征结构还是任务结构？神经表征相似性分析中的统计偏差及一种可降低偏差的贝叶斯方法. 预印本平台bioRxiv, 2018: 347260。