Dissociating early and late error signals in perceptual recognition.
Dissociating early and late error signals in perceptual recognition.
复制标题
分离感知识别中的早期和晚期错误信号。
DOI：
10.1162/jocn.2008.20155
复制
发表时间：
2008
影响因子：
3.2
通讯作者：
Velanova,Katerina
中科院分区：
文献类型：
--
作者：
Wheeler,MarkE;Petersen,StevenE;Nelson,StevenM;Ploran,ElisabethJ;Velanova,Katerina
关于对象身份的决定是在收集和分析证据的一段时间之后做出的。证据可以包括与任务相关的外部刺激和内部产生的目标和期望。神经系统如何收集各种信息并将其过滤成有意义的证据，在很大程度上是未知的。虽然目标识别通常是高效和准确的，但错误是常见的。错误可能与早期对传入刺激信息的误解导致的证据收集错误有关。此外，众所周知，任务绩效中的错误会导致延迟纠正绩效监控机制，从而优化或以其他方式调整未来的行为。在这项研究中，我们在物体识别的扩展试验范式中使用功能磁共振成像(FMRI)来研究我们是否可以识别识别时刻之前和之后的基于性能的信号调制。目前这份报告的基本原理是，fMRI活动的早期调制可能反映了证据收集的缺陷，而晚期调制可能反映了性能监测机制的存在。我们通过比较事先选择的感兴趣区(ROI)正确试验和错误试验(ROI)上的fMRI活动来测试这种可能性。我们在不同的先验ROI集合中发现了识别前和识别后的精确度相关的调制，这表明存在可分离的误差信号。
Decisions about object identity follow a period in which evidence is gathered and analyzed. Evidence can consist of both task-relevant external stimuli and internally generated goals and expectations. How the various pieces of information are gathered and filtered into meaningful evidence by the nervous system is largely unknown. Although object recognition is often highly efficient and accurate, errors are common. Errors may be related to faulty evidence gathering arising from early misinterpretations of incoming stimulus information. In addition, errors in task performance are known to elicit late corrective performance monitoring mechanisms that can optimize or otherwise adjust future behavior. In this study, we used functional magnetic resonance imaging (fMRI) in an extended trial paradigm of object recognition to study whether we could identify performance-based signal modulations prior to and following the moment of recognition. The rationale driving the current report is that early modulations in fMRI activity may reflect faulty evidence gathering, whereas late modulations may reflect the presence of performance monitoring mechanisms. We tested this possibility by comparing fMRI activity on correct and error trials in regions of interest (ROIs) that were selected a priori. We found pre- and postrecognition accuracy-dependent modulation in different sets of a priori ROIs, suggesting the presence of dissociable error signals.