School of Psychology, Washington Singer Laboratories, Exeter, UK.
Eur J Neurosci. 2011 Oct;34(8):1326-36. doi: 10.1111/j.1460-9568.2011.07847.x. Epub 2011 Oct 13.
An influential multi-process model of category learning, COmpetition between Verbal and Implicit Systems (COVIS), suggests that a verbal or a procedural category learning process is adopted, depending on the nature of the learning problem. While the architectural assumptions of COVIS have been widely supported, there is still uncertainty regarding the types of category structures that are likely to engage each of the COVIS systems. We examined COVIS in an fMRI study with two novel (in terms of COVIS research) categorizations. One of the categorizations could be described by a simple, unidimensional, rule that was expected to favor the verbal system. The other categorization possessed characteristics typically associated with the procedural system, but could also potentially be verbalized using a rule more complex than the ones previously associated with the verbal system. We found that both categorizations engaged regions associated with the verbal system. Additionally, for both categorizations, frontal lobe regions (including left ventrolateral frontal cortex) were more engaged in the first compared to the second session, possibly reflecting the greater use of hypothesis-testing processes in the initial stages of category acquisition. In sum, our results extend our knowledge of the conditions under which the verbal system will operate. These findings indicate that much remains to be understood concerning the precise interplay of the verbal and procedural categorization systems.
一个有影响力的多过程类别学习模型,即言语和内隐系统竞争模型(COVIS),表明采用言语或程序类别学习过程取决于学习问题的性质。虽然 COVIS 的架构假设得到了广泛支持,但对于可能涉及 COVIS 系统的每种类型的类别结构仍存在不确定性。我们在一项 fMRI 研究中检验了 COVIS,该研究采用了两种新的(就 COVIS 研究而言)分类方法。其中一种分类可以用一条简单的、单维的规则来描述,这条规则预计有利于言语系统。另一种分类具有与程序系统相关的特征,但也可以使用比之前与言语系统相关的规则更复杂的规则来言语化。我们发现,这两种分类都涉及到与言语系统相关的区域。此外,对于这两种分类,额叶区域(包括左侧腹外侧额皮质)在第一次比第二次会议中更活跃,这可能反映了在类别获取的初始阶段更频繁地使用假设检验过程。总之,我们的研究结果扩展了我们对言语系统将运作的条件的认识。这些发现表明,对于言语和程序分类系统的确切相互作用,仍有许多需要了解。