摘要：Autobiographical memory (AM), the recollection of personally-experienced events, has several adaptive functions and has been studied across numerous dimensions. We previously introduced two methods to quantify across the life span AM content (the amount and types of retrieved details) and the everyday occurrence of its recollection. The CRAM (cue-recalled autobiographical memory) test used naturalistic word prompts to elicit AMs. Subjects dated the memories to life periods and reported the numbers of details recalled across eight features (e.g., spatial detail, temporal detail, people, and emotions). In separate subjects, an experience sampling method quantified in everyday settings the frequency of AM retrieval and of mental representation of future personal events or actions (termed prospective memory: PM); these data permit evaluation of the temporal orientation of episodic recollection. We describe these datasets now publicly released in open access (CRAM: doi.org/10.6084/m9.figshare.10246958 ; AM-PM experience-sampling: doi.org/10.6084/m9.figshare.10246940 ). We also present examples of data mining, using cluster analyses of CRAM (14,242 AMs scored for content from 4,244 subjects). Analysis of raw feature scores yielded three AM clusters separated by total recalled content. Normalizing for total content revealed three classes of AM based on the relative contributions of each feature: AMs containing a relatively large number of details related to people, AMs containing a high degree of spatial information, and AMs with details equally distributed across features. Differences in subject age, memory age, and total content were detected across feature clusters. These findings highlight the value in additional mining of these datasets to further our understanding of autobiographical recollection.
