Nathaniel-James D A, Frith C D
Wellcome Department of Cognitive Neurology, Institute of Neurology, University College London, London, WC1N 3BG, United Kingdom.
Neuroimage. 2002 Aug;16(4):1094-1102. doi: 10.1006/nimg.2002.1167.
Although the prefrontal cortices, in particular the dorsolateral prefrontal cortex (DLPFC), have been examined in numerous imaging and neuropsychological studies, it has proved difficult to assign a specific function to this brain region. The aim of this study was therefore to delineate the function of the DLPFC first by using positron emission tomography with a word generation task and second by comparing these findings with a series of different circumstances in which the DLPFC is activated in association with response selection. Six healthy volunteers were presented with a sentence completion task based on the Hayling test. In each of two conditions, (A) response initiation and (B) response suppression, volunteers saw a sentence with the final word omitted. In condition A they had to provide a word that fitted at the end of the sentence and in condition B they had to provide a word that did not fit. The corpus of sentences used varied systematically in their level of contextual constraint, ranging from low to high. With all levels of constraint combined, significant greater activation was observed in the left DLPFC (BA46/9) under the suppression condition and in the medial orbitofrontal cortex (BA11) under the initiation condition. Under the high-constraint condition with both tasks combined, significant right middle temporal activity (BA21) was observed, whereas under low constraint, the left DLPFC was significantly activated. An interaction of task and constraint revealed that the left DLPFC was significantly more active in the suppression task at all levels of constraint, but only under the low-constraint conditions in the initiation task. The reaction time analyses mirrored the pattern of activity observed, with slower reaction times under conditions of suppression and under conditions of initiation with low constraint. By comparing these findings with other studies of response selection, we argue that the most likely single cognitive function of the DLPFC is to specify a set of responses suitable for a given task and to bias these for selection (sculpting the response space). This function resembles the biasing of competition between stimuli in the model of Desimone and Duncan and is analogous to the component of the Supervisory Attentional System that modulates the contention scheduling system.
尽管前额叶皮质，尤其是背外侧前额叶皮质（DLPFC），已经在众多成像和神经心理学研究中得到了检验，但事实证明，要为这个脑区赋予特定功能是很困难的。因此，本研究的目的首先是通过使用正电子发射断层扫描技术和单词生成任务来描绘DLPFC的功能，其次是将这些发现与一系列不同的情况进行比较，在这些情况下，DLPFC与反应选择相关联而被激活。六名健康志愿者参与了基于海林测试的句子完成任务。在两种条件下，（A）反应启动和（B）反应抑制，志愿者会看到一个省略了最后一个单词的句子。在条件A下，他们必须提供一个适合放在句子末尾的单词，而在条件B下，他们必须提供一个不适合的单词。所使用的句子语料库在上下文约束水平上系统地变化，范围从低到高。综合所有约束水平，在抑制条件下，左DLPFC（BA46/9）观察到显著更大的激活，而在启动条件下，内侧眶额皮质（BA11）观察到显著更大的激活。在两种任务结合的高约束条件下，观察到显著的右侧颞中回活动（BA21），而在低约束条件下，左DLPFC被显著激活。任务和约束的交互作用表明，在所有约束水平下，左DLPFC在抑制任务中显著更活跃，但仅在启动任务的低约束条件下如此。反应时间分析反映了观察到的活动模式，在抑制条件下和低约束启动条件下反应时间较慢。通过将这些发现与其他反应选择研究进行比较，我们认为DLPFC最可能的单一认知功能是指定一组适合给定任务的反应，并对这些反应进行偏向选择（塑造反应空间）。这个功能类似于德西蒙和邓肯模型中刺激之间竞争的偏向，并且类似于调节竞争调度系统的监督注意系统的组成部分。
