Functional magnetic resonance imaging of source versus item memory Jin Fan, CA Joan Gay Snodgrass 1 and Robert M. Bilder 2 Sackler Institute for Developmental Psychobiology, Department of Psychiatry, Weill Medical College of Cornell University,1300 York Avenue, Box 140, New York, NY 10021; 1 Department of Psychology, New York University; 2 Center for Advanced Brain Imaging, Nathan S. Kline Institute for Psychiatric Research, USA CA Corresponding Author: jif2004@med.cornell.edu Received 2 June 2003; accepted 9 June 2003 DOI: 10.1097/01.wnr.0000090582.35425.e1 Both the frontal lobes and the medial temporal lobe (particularly the hippocampus) have been implicated in encoding and retrieval of episodic memory information. We report an experiment that ma- nipulates whether source information, item information, or both are required at retrieval. Two sources were used in a factorial de- sign in which the main e¡ect of source and item retrieval, along with their interaction, could be measured by fMRI activations. When source information was required at retrieval the left frontal lobe showed signi¢cant activation but not when item retrieval was required. Hippocampal activation showed no di¡erence between source and item retrieval. This pattern of results supports a model proposing a larger role for the frontal lobes in encoding and retrie- val of source information. NeuroReport 14:2275^2281 c 2003 Lippincott Williams & Wilkins. Key words: fMRI; Item memory; Source memory INTRODUCTION Episodic memory is memory for events that occurred in a particular place at a particular time [1]. Episodic memory contains information about the content of an experience (item), as well as its context (source) [2]. Some researchers have argued for a dual-process model of source and item memory whereas others have argued for a single process model, in which source memory does not differ in kind from item memory but simply reflects a more elaborated memory trace. Dual-process model: The dual-process model of source vs item memory for the brain holds that the prefrontal regions subserve source memory and the hippocampal regions subserve item memory [3]. Evidence for the dual-process model initially came from studies of patients with frontal lobe dysfunction, who often show normal item memory but deficient source memory. For example, patients with frontal damage who recall or recognize information at normal levels may show disrupted memory for temporal order [4]. While lesion evidence suggests that memory for source is subserved by the frontal lobes [5,6], similar clinical data suggest that the recognition of an item is subserved by medial temporal and/or diencephalic structures [7], sup- ported by a double dissociation between item memory and source memory in elderly subjects [8]. A PET study has supported the dual brain process model [9]. It is found that item retrieval is related to increased activity in medial temporal and basal forebrain regions, whereas temporal order (source) retrieval is related to activations in dorsal prefrontal, cuneus/precuneus, and right posterior parietal regions. Single-process model: The single-process model proposes that both frontal and hippocampal areas are more involved in source than in item memory retrieval because retrieval of source information is more difficult than retrieval of item information. Some support for the single process model is provided by studies demonstrating that in addition to the frontal lobes, the medial temporal system (hippocampus and adjacent medial temporal lobe) is important for encoding and retrieving source information [10]. Contra- dicting the dual-process model, source accuracy was related to scores on the medial temporal measure suggesting that medial temporal structures are involved in the binding of contextual features (source information) to item information [11]. In an event-related fMRI study, prefrontal activity was related more to source than to item retrieval [12]. In this study, four participants performed acquisition tasks with picture and word stimuli by rating the difficulty of drawing each item or by rating the number of possible uses for each item. During a subsequent surprise memory test, partici- pants performed an old/new recognition test or a source (picture vs word) test. The left prefrontal region was more activated in the source than the item test, suggesting that the 0959- 4965 c Lippincott Williams & Wilkins Vol 14 No 17 2 December 2003 2275 BRAIN IMAGING NEUROREPORT Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.