Functional neuroimaging studies have associated episodic memory with activations in the prefrontal cortex (PFC), the medial temporal lobes (MTL), and other brain regions (for reviews, see Cabeza, 1999; Cabeza & Nyberg, 2000b; Nyberg & Cabeza, 2000). However, the specific contributions of each of these regions and their subregions to various episodic memory processes remain unclear. Currently, we are trying to clarify the neural correlates of two sets of episodic memory processes: relational memory and recollection, and true vs. false memories.

• Dennis, N. A., & Cabeza, R. (2008). Neuroimaging of healthy cognitive aging. In F. I. M. Craik & T. A. Salthouse (Eds.), Handbook of aging and cognition: Third edition. Mahwah, NJ: Erlbaum.
• Nyberg, L., & Cabeza, R. (2000). Brain imaging of memory. In E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 501-519). New York: Oxford University Press.
• Cabeza, R., & Nyberg, L. (2000). Neural basis of learning and memory: Functional neuroimaging evidence. Current Opinion in Neurology, 13, 415-421.
• Cabeza, R. (1999). Functional neuroimaging of episodic memory retrieval. In E. Tulving (Ed.), Memory, consciousness, and the brain: The Tallinn Conference (pp. 76-90). Philadelphia: The Psychology Press.

Relational Memory and Recollection

When we remember past events, we typically remember not only the various components of the event, or item memory, but also the relationships among these components, or relational memory. Consistent with current models of MTL function, our fMRI data has linked item memory to the rhinal cortex, and relational memory, to the hippocampus (Daselaar et al., in preparation).

Relational memory involves different kinds of associations. For example, when we remember a statement we heard at a party, we may also recall the topic of the conversation (semantic associations), the voice of the speaker (perceptual associations), her location (spatial associations), and the events that preceded and followed the statement (temporal-order associations).

We have compared the neural correlates of relational memory for semantic vs. perceptual associations and found differences in domain-specific regions, as well as overlapping activity in the hippocampus (Prince et al., 2005). We have also compared item memory for faces and scenes and relational memory for face-scene associations (Prince et al., in prep.). Whereas face and scene memory engaged stimuli-specific regions in ventral temporal cortex (i.e., fusiform face area vs. parahippocampal place area), face-scene associations differentially recruited the hippocampus.

The distinction between item and relational memory is closely related to the distinction between familiarity (item memory without relational memory) and recollection (item plus relational memory). In recognition tests with confidence ratings, familiarity is assumed to increase gradually from "definitely new" to "definitely old" (a linear function), whereas recollection is assumed to occur mainly for "definitely old" responses (an exponential function). By comparing linear vs. exponential brain activity functions, we found that, consistent with our relational memory data, familiarity differentially engaged rhinal cortex, whereas recollection differentially engaged the hippocampus (Daselaar et al., 2006). Additionally, we found familiarity/recollection dissociations between subregions of left parietal cortex and posterior midline cortex, brain areas which are now intensely studied in relation to episodic memory.

• Daselaar, S. M., & Cabeza, R. (in preparation). Neural correlates of relational memory encoding: Role of preexisting associations.
• Prince, S. E., & Cabeza, R. (in preparation). Functional neuroanatomy of relational memory encoding and retrieval: Associating faces and scenes.
• Daselaar, S. M., Fleck, M. S., & Cabeza, R. (2006). Triple dissociation within the medial temporal lobes: Recollection, familiarity, and novelty. Journal of Neurophysiology, 96, 1902-1911. (Must Read label in Faculty of 1000)
  
• Prince, S. E., Daselaar, S. M., & Cabeza, R. (2005). Neural correlates of relational memory: Successful encoding and retrieval of semantic and perceptual associations. Journal of Neuroscience, 25, 1203-1210. (Recommended label in Faculty of 1000)
  

True vs. False Memories

One of the most fundamental questions in memory research is how our memory can sometimes be highly accurate and other times grossly inaccurate. In a recent fMRI study (Daselaar et al., 2006), we found evidence that activity in a more posterior MTL region leads to accurate recognition responses whereas activity in a more anterior MTL region leads to memory errors. Logistic regression analyses demonstrated that both regions make a significant and independent contribution to recognition memory decisions, and functional connectivity analyses revealed an opposing relationship between the two activations. Thus, our results suggest that our memory responses reflect a compromise between accurate and inaccurate memory signals in our brain.

To amplify the difference between these two signals, we have been using the Deese-Roediger-McDermott (DRM) false memory paradigm, in which participants show a strong tendency to false alarm to nonstudied words that are semantically related to several studied words. In a retrieval fMRI study (Cabeza et al., 2001), we found a dissociation between a hippocampal region, which was activated during both true and false recognition, and a parahippocampal region, which was associated with true but not with false recognition. In another retrieval fMRI study (Kim & Cabeza, in preparation), we found a dissociation between confidence in true recognition responses, which primarily reflected MTL activity, and confidence in false recognition responses, which primarily reflected PFC activity. Finally, in an encoding fMRI study, we found a clear dissociation between MTL activity, which predicted mainly subsequent true memories, and left ventrolateral PFC activity, with predicted both true and false memories. Thus, in several studies we have found that some MTL regions are differentially more involved in the encoding and retrieval of accurate memories, whereas other MTL regions and some PFC regions do not distinguish between true and false memories or are more associated with false than with true memories.

• Kim, H., & Cabeza, R. (2007). Trusting our memories: Dissociating the neural correlates of confidence in veridical vs. illusory memories. Journal of Neuroscience, 27, 12190-12197.

• Kim, H., & Cabeza, R. (2007). Differential contributions of prefrontal, medial temporal, and sensory-perceptual regions to true and false memory formation. Cerebral Cortex, 17, 2143-2150.
• Daselaar, S. M., Fleck, M. S., Prince, S. E., & Cabeza, R. (2006). The medial temporal lobe distinguishes old from new independently of consciousness. Journal of Neuroscience, 26, 5835-5839.
• Cabeza, R., Rao, S. M., Wagner, A. D., Mayer, A., & Schacter, D. L., (2001). Can medial temporal lobe regions distinguish true from false? An event-related fMRI study of veridical and illusory recognition memory. Proceedings of the National Academy of Sciences, USA, 98, 4805-4810.