The Effect Of Visual Long-term Memory On Visual Short-term Memory

It is a fundamental question that whether long-term memory (LTM) and short-term memory (STM) are two different and separate stores or two different states of the same representation. However, existing studies on this topic yielded inconsistent results. So far, though the view considering LTM and STM as two separate and different stores (multi-stores model) was widely accepted by most of the researchers, as the development of cognitive neuropsychology and neuroimage techonolgy, more and more studies find that, LTM and STM shared the same neural pathways or owned the same activation pattern. Therefore, the unitary-store model which has being ignored for years now attracts our attention again. According to the unitary store model, STM consists of temporary activated LTM, they are two different states of the same representation. Therefore, some associations exist between LTM and STM, and the activation level of LTM per se could affect the performance of STM. Conversely, if LTM and STM are two disassociated stores, STM would not be affected by LTM.The present study first compiled and analyzed related studies that focused on visual long-term memory (VLTM), the result implied it ought to be noted that in studies that found no facilitation effect, the exposure of the experimental materials was not sufficiently long to activate VLTM. Therefore, it was hypothesized in this study that only a highly activated VLTM could facilitate visua short-term memory (VSTM). So, in the present studies, we'd like to explore more evidences from different angles.Three experiment were adopted in chapter3to systematically manipulate the actiation levels of VLTM so that to explore possible effects of VLTM on VSTM. Within change-detection paradigm, participants were first instructed to memorize a study image including6geometric figures (6random shape-color bindings from a pool of eight shapes and eight colors), and then after a random inter-stimulus interval (1,1.5, or3s), either the same image or an image with one shape or color changed was presented, and participants were asked to judge whether they detected a change. Because the stimuli figures were random bindings of color and shape, it was considered as no VLTM condition and the base line of the following experiments. In experiment2, a pool of8geometric figures with fixed combinations of color and shape were used to substitute all the random combinations of experiment1. The experimental procedure and design was identical to experiment1. One hour after the experiment, the activation level of VLTM was examined. In Experiment3, participants were asked to visually study the8figures in experiment2with high frequence to form highly activated VLTM. At last, the experimental procedure as described in Experiment2was performed. The results showed that although VLTM was formed in experiment2, there was no significant increase in d'or K value of VSTM. However, in experiment3, VLTM was highly activated, and a better performance was observed as compared with experiment1and2. In addition, the effect size of ISI decreased as the VLTM activation level increased from experiment1to experiment3. The results of the present study suggest that only a highly activated VLTM can facilitate VSTM, and, it can also inhibit the rapid fading of VSTM traces.In chapter4, the highly activated VLTM was adopted to explore its effect on feature binding in VSTM, the dependent on location was considered as an indicator that VLTM facilitated VSTM. The stimuli in experiment4was identical to experiment1, but, when the probe appeared, participants had to judge whether there were two items that exchanged their color or shape. The same pre-experiment as as to experiment3was adopted to form highly activiated VLTM in experiment5and experiment6. The design, experimental procedure and task of experiment5were identical to experiment4. In experiment6, the probe image could be identical to the study image, or, two of them were substituted by another two items that also pertained to the VLTM, and the task was to judge whether two of the study image was subsitituted by two new items, by which, the strategical processing could be excluded. To test and exclude VLTM component that formed based on study image, variable presentation time of study image and fixed ISI was used, the experimental procedure was similar to experiment4. The result showed that, the dependent on location was found in experiment4, experiment5and experiment6. In addition, compared to experiment4, the performance of experiment5and6was improved as well as the effect size of location decreased. In experiment7, the dependent on location was also found except for the longest presentation of study image (1min). These results suggest that, the highly activated VLTM could facilitate VSTM, maybe partly by strengthen the encoding of VSTM.The serial item-recognition task was used in chapter5to confirm the'activation level'and'focus of attention', which were the two core concepts of unitary store model. In experiment8,6items (which were identical to experiment1) were first presented to participants in sequence, which was separated by a short ISI. The recognition image include only one item, which maybe one of the6items or not. The task of participant was to judge if the recognition item was one of the6items they memorized before. The same pre-experiment and stimuli as in experiment3were adopted in experiment9to form highly activated VLTM, the experimental procedure and task was identical to experiment8. The result showed that, as the item was presented later, the recognition performance was better, the last presented item own the highest performance. In addition, compared to experiment8, the the first presented5items was better recognized in experiment9, however, the recognition performance of the last item was not improved. These results suggest that, the distinction between focus of attention and out of focus is true, and the highly activated VLTM could facilitate the performance of these items which are out of focus of attention.In sum, the present study suggest that, highly activated VLTM can facilitate VSTM, and this facilitation is widely existed in VSTM processing, such as encoding and retention. Therefore, LTM and STM are more likely two states of the same representation.