Todd Winn, Leila LeSueur, and Dr. Brock Kirwan, Psychology Department
Introduction
In current literature, researchers have proposed that the relationshipbetween objects and the context they are found in is integral to long-term declarative memory, and converges at hippocampal processes. In order to explore the effect of visual context on memory, our experiment was designed to evaluate the behavioral performance of subjects completing a contextual memory task intended to tax the pattern separation process in the hippocampus. Pattern separation is the computational process performed in the hippocampus with the goal of making similar patterns of neural activity as dissimilar as possible, thereby allowing us to distinguish between two objects that appear similar. Intentionally taxing this process with a task involving sets of similar objects paired with background images that are either contextually congruent or incongruent with the foreground image, will aid in determining the role visual context plays in relation to how well subjects remember distinct objects.
Methodology
This study utilizes a two-phase study-test paradigm. During phase one, subjects are randomly presented with 260 composite images consisting of a foreground object overlaid on either a contextually congruent or incongruent background. Subjects are directed to focus on the foreground images as there will be a memory test related to the object in the foreground in the next phase. Participants are prompted to respond whether the foreground object would be larger or smaller than a shoebox in order to gauge participation in the first phase. Phase two testing consists of a pattern separation task where 320 composite images with either the same object seen in the study phase or a similar foreground object superimposed on either the same background as before, or a similar one. The images are presented for 2.5 seconds, and subjects are asked to respond whether the foreground object is either old (repeat of an object seen in study phase) or similar. Demographics for subjects was n = 75, with 43 female participants.
Results
The main effects of background (Figure 1) were that background changes facilitated distinguishing between similar stimuli, hindered recognition of repeat stimuli, and incongruent backgrounds had a significant effect on both discrimination and recognition. The interaction between background and congruency (Figure 2) were only significant when the background changed between study and test phase. Specifically, discrimination was hindered by incongruent backgrounds that changed, and recognition was facilitated by incongruent backgrounds that changed. Finally, the effects of gender (Figure 3) only showed significance when the background changed. Females showed improved recognition when the background was incongruent and background changed. Males showed improved discrimination when the background was congruent with the object and the background changed.
Discussion
In studying the relationship between visual context and declarative memory, we found in the current study that the congruency of an object to the background did not influence discrimination or recognition of that object unless the background changed. Interference in the form of backgrounds changing between the study and test phases appeared to be the primary determinant of significant results. Background changes facilitated identification of similar stimuli, but hindered the ability to recognize repeat stimuli. When an object and background were incongruent, both discrimination and recognition were facilitated, but only if the background had changed. One major limitation in the design of this study is the large number of trial type permutations. Although there appears to be an interaction with interference, there are too many combinations to narrow down the type of interference or what is interacting with it. An experimental designed explicitly with only one type of interference would help resolve this issue. The differences in sex interactions may be related to the known differences between males and females when it comes to solving spatial tasks, and would likely be more fully elucidated by use of fMRI to image differences in activation during the two specific conditions in which males and females differed in performance (Figure 3).
Conclusion
A possible explanation for the improvement in performance for discrimination is that when the background changes and is incongruent, individuals pay more attention to the object. Further investigation using a paradigm that removes the background overlay from the test phase may prove useful in decreasing a potentially confounding mechanism. Investigation utilizing fMRI may also assist in identifying disparate brain sub-region activation mnemonic discrimination behavior and also help identify causes for the difference in performance between sexes on repeat stimuli.