FMRI studies of autobiographical memory (AM) retrieval typically ask subjects to retrieve memories silently to avoid speech-related motion artifacts. Recently, some fMRI studies have started to use overt (spoken) retrieval to probe moment-to-moment retrieved content. However, the extent to which the overt retrieval method alters fMRI activations during retrieval is unknown. Here we examined this question by eliciting unrehearsed AMs during fMRI scanning either overtly or silently, in the same subjects, in different runs. Differences between retrieval modality (silent vs. narrated) included greater activation for silent retrieval in the anterior hippocampus, left angular gyrus, PCC, and superior PFC, and greater activation for narrated retrieval in speech production regions, posterior hippocampus, and the DLPFC. To probe temporal dynamics, we divided each retrieval period into an initial search phase and a later elaboration phase. The activations during the search and elaboration phases were broadly similar regardless of modality, and these activations were in line with previous fMRI studies of AM temporal dynamics employing silent retrieval. For both retrieval modalities, search activated the hippocampus, mPFC, ACC, and PCC, and elaboration activated the left DLPFC and middle temporal gyri. To examine content-specific reactivation during retrieval, the timecourse of narrated memory content was transcribed and modeled. We observed dynamic activation associated with object content in the lateral occipital complex, and activation associated with scene content in the retrosplenial cortex. The current findings show that both silent and narrated AMs activate a broadly similar memory network, with some key differences, and add to current knowledge regarding the content-specific dynamics of AM retrieval. However, these observed differences between retrieval modality suggest that studies using overt retrieval should carefully consider this method's possible effects on cognitive and neural processing.

中文翻译：

---

FMRI 与自传体记忆的关联：无声检索与叙述检索的比较

自传体记忆 (AM) 检索的 FMRI 研究通常要求受试者安静地检索记忆，以避免与语音相关的运动伪影。最近，一些功能磁共振成像研究已经开始使用公开（口头）检索来探究即时检索的内容。然而，显性检索方法在检索过程中改变 fMRI 激活的程度尚不清楚。在这里，我们通过在功能磁共振成像扫描过程中，在相同的受试者中，在不同的运行中，公开或无声地引出未经排练的 AM 来研究这个问题。检索方式（无声与有叙述）之间的差异包括前海马、左角回、PCC和上前额叶皮层中无声检索的更大激活，以及言语产生区域、后海马和DLPFC中有叙述检索的更大激活。为了探究时间动态，我们将每个检索周期分为初始搜索阶段和后来的细化阶段。无论模态如何，搜索和细化阶段的激活大致相似，并且这些激活与之前采用静默检索的 AM 时间动态的 fMRI 研究一致。对于这两种检索方式，搜索激活海马体、mPFC、ACC 和 PCC，而精化则激活左侧 DLPFC 和颞中回。为了检查检索过程中特定内容的重新激活，对叙述记忆内容的时间过程进行了转录和建模。我们观察到与枕外侧复合体中的物体内容相关的动态激活，以及与压后皮质中的场景内容相关的激活。目前的研究结果表明，无声 AM 和有叙述 AM 都会激活大致相似的记忆网络，但存在一些关键差异，并增加了有关 AM 检索的内容特定动态的当前知识。然而，这些观察到的检索方式之间的差异表明，使用公开检索的研究应该仔细考虑该方法对认知和神经处理的可能影响。