Data from: Emerging Representational Geometries in the Visual System Predict Reaction Times for Object Categorization

Recognizing an object takes just a fraction of a second, less than the blink of an eye. Applying multivariate pattern analysis, or "brain decoding", methods to magnetoencephalography (MEG) data has allowed researchers to characterize, in high temporal resolution, the emerging representation of objects that underlie our capacity for rapid recognition. Shortly after stimulus onset, exemplar stimuli cluster by category in high-dimensional activation spaces. In these emerging activation spaces, the decodability of exemplar category varies over time, reflecting the brain's transformation of visual inputs into coherent categorical representations. How do these emerging representations relate to categorization behavior? Recently it has been proposed that the distance of an exemplar representation from a categorical boundary in an activation space is critical for perceptual decision-making, and that reaction times should therefore correlate with distance from the boundary. The predictions of this distance hypothesis have been born out in human inferior temporal cortex (IT), an area of the brain crucial for the representation of object categories. The time of peak decoding is the optimal time for category information to be "read out" from the brain's time varying representation of the stimuli. In this study, we tested the distance hypothesis, and specifically whether or not the brain reads out at the optimal time for choice behavior. Using MEG decoding methods, we show that the distance of a pattern of activity from a decision boundary through a high-dimensional activation space correlates with reaction times in a visual categorization task, but only during the period of peak decodability. Our results suggest the brain uses the optimal stimulus representation for choice behavior, and that neural representations for objects are partially constitutive of the decision process in visual perception.识别物体仅需短短一瞬，时长甚至不及一次眨眼。将多变量模式分析（multivariate pattern analysis）方法，亦即"脑解码"（brain decoding），应用于脑磁图（magnetoencephalography, MEG）数据后，研究者得以通过高时间分辨率刻画支撑快速识别能力的物体渐次形成的表征。刺激呈现后不久，样例刺激会在高维激活空间中按类别聚簇。在此类渐次形成的激活空间中，样例类别的可解码性随时间动态变化，这反映了大脑将视觉输入转化为连贯类别表征的神经加工过程。那么此类渐次形成的表征与分类行为之间存在何种关联？近期有研究提出，激活空间内样例表征与类别边界的距离对知觉决策至关重要，因此反应时应与该边界距离呈相关关系。该距离假说的预测已在人类下颞叶皮层（inferior temporal cortex, IT）中得到验证——这一脑区正是物体类别表征的关键脑区。解码峰值时刻，正是从大脑随时间动态变化的刺激表征中读取类别信息的最佳时机。本研究对该距离假说进行了验证，并专门检验了大脑是否会选取适配决策行为的最佳信息读取时机。借助MEG解码方法，我们发现：在视觉分类任务中，神经活动模式在高维激活空间内与决策边界的距离与反应时呈显著相关，但这一关联仅存在于解码峰值时段内。本研究结果表明，大脑会选取最优的刺激表征用于决策行为，且物体的神经表征是视觉知觉决策过程的核心组成部分之一。