Carnegie Mellon University, Dept. of Psychology, Pittsburgh, USA; University of Groningen, Dept. of Artificial Intelligence, Groningen, The Netherlands.
Carnegie Mellon University, Dept. of Psychology, Pittsburgh, USA. Electronic address: ja+@cmu.edu.
Neuroimage. 2015 Mar;108:60-73. doi: 10.1016/j.neuroimage.2014.12.029. Epub 2014 Dec 19.
In this paper we propose a new method for identifying processing stages in human information processing. Since the 1860s scientists have used different methods to identify processing stages, usually based on reaction time (RT) differences between conditions. To overcome the limitations of RT-based methods we used hidden semi-Markov models (HSMMs) to analyze EEG data. This HSMM-EEG methodology can identify stages of processing and how they vary with experimental condition. By combining this information with the brain signatures of the identified stages one can infer their function, and deduce underlying cognitive processes. To demonstrate the method we applied it to an associative recognition task. The stage-discovery method indicated that three major processes play a role in associative recognition: a familiarity process, an associative retrieval process, and a decision process. We conclude that the new stage-discovery method can provide valuable insight into human information processing.
在本文中，我们提出了一种新的方法来识别人类信息处理中的加工阶段。自 19 世纪 60 年代以来，科学家们已经使用不同的方法来识别加工阶段，通常基于条件之间的反应时间（RT）差异。为了克服基于 RT 的方法的局限性，我们使用隐半马尔可夫模型（HSMM）来分析 EEG 数据。这种 HSMM-EEG 方法可以识别处理阶段以及它们如何随实验条件而变化。通过将这些信息与识别阶段的大脑特征相结合，可以推断它们的功能，并推导出潜在的认知过程。为了演示该方法，我们将其应用于联想识别任务。阶段发现方法表明，联想识别中涉及三个主要过程：熟悉过程、联想检索过程和决策过程。我们得出结论，新的阶段发现方法可以为人类信息处理提供有价值的见解。
