Emily White with Dr. Brock Kirwan, Psychology Department
Introduction
Memory discrimination requires forming distinct, separate memories and depends on the process of pattern separation. Memory generalization allows you to retrieve previously-stored memories that may be grouped in memory among other, similar events and depends on the process of pattern completion. Interestingly, both of these complementary processes are mediated by the hippocampus. This means that the hippocampus must switch between pattern separation and pattern completion, a process which may be slow enough for us to measure experimentally.
Duncan, Sadanand, and Davachi (2012) found an interesting behavioral effect related to pattern separation and pattern completion. In a visual recognition test, participants saw a series of images, some of which were repeated and some of which were test images similar to ones previously shown. Participants had to identify each image as “old”, “new”, or “similar”. Identifying a picture as “old” requires a pattern completion process and identifying a picture as “new” requires a pattern separation process. When participants identified the image before the similar test image as “new”, they were more likely to correctly identify the test stimulus as “similar.” On the other hand, when participants identified the image before the similar test image as “old,” they were more likely to incorrectly identify the test image as “old.” Interestingly, the benefit of seeing a new image before the test stimulus was greater the more similar the test image was to the old one it resembled. The implication of this is that if the hippocampus is in a pattern separation or pattern completion “mode,” that bias lasts long enough to influence memory for events immediately thereafter. Duncan and colleagues tested this idea and demonstrated that the effect goes away after a few seconds. We took advantage of this time effect to study the differences in brain function between pattern completion and pattern separation.
Electroencephalography (EEG) is one of the few neuroscience techniques that allow us to see brain activity in real time. The electrical neural responses for each trial are averaged to provide a picture of brain activity. Theta waves (4-8 Hz) have been shown to occur just before a participant correctly remembers where he or she has seen something before (Addante et al., 2011). These theta waves are therefore associated with retrieving memories.
Funded by ORCA, we designed an experiment to (1) replicate the behavioral effect observed by Duncan and colleagues and (2) use EEG to determine if there is a difference in brain activity between pattern completion bias and pattern separation bias during the pre-stimulus interval. Specifically, we predicted that there would be an increase in power in the theta band (4-8 Hz) in the interval between “new” and “similar” stimuli when compared to the time interval between “old” and “similar” stimuli.
Methodology
For our experiment, participants recruited through the BYU Psychology Department’s SONA database were fitted with an EEG electrode net. Modeled on the standard pattern separation task used by Duncan and colleagues, they were given a visual memory task consisting of 8 blocks with 150 images of everyday objects. Each image appeared on the screen for 1.5 seconds and was separated from the next by a fixation cross. For each image, the participant determined if he or she had seen it before, if it was similar to one seen before, or if the image was new. The participant pressed one of three buttons to respond “old,” “similar,” or “new.” Images were shown and responses recorded through E-Prime software and brain activity was recorded through the Net Station program. The data was averaged across trial types and compared to other participants. We compared data from between “old” and “similar” and between “new” and “similar” stimuli.
Results
Behavioral data mimics the pattern demonstrated by Duncan and colleagues in which “similar” stimuli preceded by “old” stimuli are more likely to be incorrectly identified as “old” while those preceded by “new” stimuli are more likely to be correctly identified as “similar”. Preliminary EEG data suggest there may be greater power in the theta band for the latter condition. However, the filtering and screening processes have eliminated too much data for statistical significance to be determined.
Discussion
We continue to investigate the neural mechanisms behind memory processes. More participants will need to be added to our data pool before any conclusions can be drawn about the nature of the electrical signals produced in successful memory retrieval. We hope that a combination of research techniques can be used to shed further light on the relationship between memory generalization and memory discrimination as well as other neural processes.
Scholarly Sources
- Addante R. J., Watrous, A. J., Yonelinas, A. P., Ekstrom, A. D., & Ranganath, C. (2011). Prestimulus theta activity predicts correct source memory retrieval. Proceedings of the National Academy of Sciences of the United States of America, 108 (26), 10702-10707.
- Duncan, K., Sadanand, A., &Davachi, L. (2012). Memory’s Penumbra: Episodic memory decisions induce lingering mnemonic biases. Science, 337(6093), 485-487.