Gonsalves, B., Reber, P.J., Gitelman, D.R., Parrish, T.B., Mesulam, M.-M., & Paller, K.A. (2004). Neural evidence that vivid imagining can lead to false remembering. Psychological Science, 15, 655-660.
Introduction
Subsequent-memory, or Dm, analyses have been performed by cognitive neuroscientists in which they examine brain activity during encoding as a function of later remembering or forgetting on an item-by-item basis. The evidence yielded by these studies has shown that a more positive ERP and greater left inferior prefrontal cortex (using fMRI) activation at encoding occur for words that are later remembered. Reality-monitoring errors are errors at recall where an imagined event is believed to have been perceived. The authors hypothesized that greater activation in brain regions associated with generating visual imagery would occur for words that are later misremembered as perceived in a paradigm designed to elicit reality-monitoring errors.
Method
Stimuli consisted of 350 visually displayed words, 350 color photographs, and 525 spoken words (test). All words were object names.
During study, subjects were shown a word and asked to generate a visual image corresponding to the object name. For half of the words, a photograph of the object was shown 2,000 ms after the word and for the other half a blank rectangle was shown. There were seven separate functional scans during study, and a structural scan was performed after the end of the study phase.
Testing began 20 min after study and was conducted outside the scanner. Ss (8 male, 3 female volunteers) listened to digitally recorded spoken words and indicated if they had seen a picture of the object. There were 175 object names that had been presented with picture, 175 object names that had been presented with word only, and 175 new objects. Responses were made on a keyboard.
Estimates of differences in neural activity between trial types were obtained for each participant and analyzed to identify differences that were consistent across subjects.
Results
Endorsement for the test resulted in: 74% of photo-and-word items (hits), 27% of word-only items (false memories), and 6% for new items (false alarms). False memories occurred significantly more often than false alarms. Words leading to false memories were on average more concrete but did not differ in frequency from words that were correctly rejected.
Three areas showed larger responses to words that were later falsely recognized as presented in picture: anterior cingulated, precuneus region in the medial parietal lobes, and right inferior parietal area.
Two areas showed larger responses for words correctly recognized at test: left inferior prefrontal cortex, left anterior medial temporal activation including the hippocampus.
Precuneus and right inferior parietal cortex showed greater activation for words that were later forgotten compared to words remembered.
Discussion
Previous work using ERP recordings was insufficient for localizing the positive neural activity associated with later false remembering. The current study, however, shows that greater activation in the precuneus, right inferior parietal cortex, and anterior cingulate is associated with encoding of words that are later falsely remembered to have been shown as a photograph. These regions have also been shown to activate during visual imagery tasks. Accurate memory was predicted by activity in the left prefrontal cortex and left hippocampus. These areas have also been show to be active in studies of episodic memory formation with left prefrontal activation being associated with increased phonological and semantic processing, and hippocampal activation establishing links among multiple cortical regions involved in representing an episode.
The authors conjecture that the vivid visual imagery produced by some words parallels the activity produced during actual perception, and subsequently this activity leads to the formation of memory traces akin to those established by actual experience. False memories are produced when a word alone produces such imagery, but also pictures are forgotten when this system produces and internal image that interferes with the encoding of the actual presented image. Most research on memory formation using ERP and fMRI has focused on accurate formation and retrieval. This article demonstrates the activity predictive of false memory expression. The type of false memories examined are those most likely to occur outside of the laboratory with imagery distorting memories that lacked the visual component (or parts of the visual component, such as eyewitnesses being vulnerable to misinformation) at encoding.