Behavioral evidence indicates that false memory, like true memory, can be rich in sensory detail. By contrast, there is fMRI evidence that true memory for visual information produces greater activity in earlier visual regions than false memory, which suggests true memory is associated with greater sensory detail. However, false memory in previous fMRI paradigms may have lacked sufficient sensory detail to recruit earlier visual processing regions. To investigate this possibility in the present fMRI study, we employed a paradigm that produced feature-specific false memory with a high degree of visual detail. During the encoding phase, moving or stationary abstract shapes were presented to the left or right of fixation. During the retrieval phase, shapes from encoding were presented at fixation and participants classified each item as previously “moving” or “stationary” within each visual field. Consistent with previous fMRI findings, true memory but not false memory for motion activated motion processing region MT+, while both true memory and false memory activated later cortical processing regions. In addition, false memory but not true memory for motion activated language processing regions. The present findings indicate that true memory activates earlier visual regions to a greater degree than false memory, even under conditions of detailed retrieval. Thus, the dissociation between previous behavioral findings and fMRI findings do not appear to be task dependent. Future work will be needed to assess whether the same pattern of true memory and false memory activity is observed for different sensory modalities.
False memory can be rich in sensory detail under certain task conditions (Lampinen et al., 1998, Payne et al., 1997, Porter et al., 1999 and Roediger and McDermott, 1995). For instance, Roediger and McDermott (1995) implemented a behavioral paradigm adapted from Deese (1959) in which participants heard lists of words (e.g., table, sit, legs, etc.) that were each related to a single critical nonpresented word (e.g., chair). The critical words were falsely recalled with high confidence, and “remember” judgment rates were not significantly different for true recognition of studied words and false recognition of critical nonpresented words.
By contrast, there is fMRI evidence that suggests true memories are associated with a greater degree of sensory detail as compared to false memories, as true memories have been associated with greater sensory cortical activity than false memories. In an fMRI study of memory for abstract shapes (Slotnick & Schacter, 2004), true memory activated earlier visual processing regions (BA 17, BA 18) to a greater degree than false memory, while both true memory and false memory activated later visual processing regions (BA 19, BA 37) to a similar degree. In a more recent fMRI study (Stark, Okado, & Loftus, 2010), participants were presented with visual vignettes and, the following day, heard conflicting auditory misinformation (participants were under the impression that the auditory information was truthful). During the retrieval phase, participants were asked whether sentences accurately or inaccurately described previously presented vignettes. Consistent with previous findings (Slotnick & Schacter, 2004; see also, Garoff-Eaton, Slotnick, & Schacter, 2006), true memory produced greater activity in earlier visual processing regions (BA 17, BA 18) than false memory, while false memory for auditory misinformation produced activity in auditory/language processing regions (BA 22, BA 42). These fMRI results suggest true memory and false memory are distinct in that true memory produced greater activity than false memory in earlier visual processing regions, while true memory and false memory activated later processing regions to a similar degree.
The aim of the present study was to address the apparent discontinuity between previous behavioral results that suggest true memory and false memory can be associated with similar levels of subjective sensory detail and the fMRI results that suggest true memory is associated with a greater degree of sensory detail than false memory. Specifically, previous fMRI paradigms may have produced false memory for modality-specific (e.g., visual) information without sufficient sensory detail to activate earlier visual processing regions (see Slotnick, 2004). To investigate this possibility, in the present fMRI study we employed a paradigm to produce false memory for feature-specific (motion) information with a higher degree of visual detail than in previous fMRI studies (although these false memories may not be as detailed as true memories, which is addressed below). During the encoding phase, participants were presented with moving or stationary abstract shapes to the left or right of fixation (Fig. 1, left). During the retrieval phase, shapes from encoding were presented at fixation and participants classified each item as previously “moving” or “stationary” within the “left” or “right” visual field (Fig. 1, right). True memory corresponded to a “moving” response to a previously moving item, whereas false memory corresponded to a “moving” response to a previously stationary item.
