Abstract: Background Item memory and source memory are two important aspects of episodic memory. Recent studies about the behavioral characteristics and neural basis of item memory and source memory indicate that they are two different psychological processions. Two models have been proposed to describe the relation between these two kinds of memories: single-process model and dual-process model. Each model is supported by substantial evidence. Up to now, three kinds of paradigms have been formed to discriminate the relation between item memory and source memory: sequential task, three-button task and exclusion task. Reviews of these paradigms indicate that the results based on sequential task and three-button task are consistent whereas the results with exclusion task are not. Based on many event-related potential (ERP) studies, Cycowicz, et al. (2003) argued that the inconsistent results of exclusion task were due to the sensory-specific of source but not the test paradigm. To validate the results by Cycowicz, et al. and to extend previous study on item memory and source memory, the present ERP study was conducted to investigate the temporal and spatial distributions of the old/new effects for item recognition and location source retrieval on line drawing.
Method Sixteen healthy undergraduates participated in this study. The stimuli were 400 drawings which were divided into 8 blocks. In each block, 30 drawings were presented either on left or on right screen during study phase, and then two tests were performed with 20 other drawings added. The first task was to decide whether each drawing presented at the central of the screen was learned or not, and the second task was to judge drawings learned from one location (left or right) as targets and other drawings as non-targets (called non-target-old or non-target-new) (exclusion task). Each drawing was displayed 500ms with the ISI of 1300ms and 1800ms during study phase and test phase respectively. EEG was recorded continuously using an Electro-cap with 62 locations (extended 10-20 system placements).
Results The results were analyzed using a three-way repeated-measures ANOVA for three windows: 200~600ms, 600~1000ms,1000~1400ms. The three factors were memory judgment, left-right electrode sites and anterior-posterior sites. The ANOVA results indicated that the temporal distribution of the old/new effect for location source retrieval was longer than that of the old/new effect for item memory. Moreover, the spatial distribution of the former effect was wider than that of the latter. In other words, the old/new effect during 1000~1400ms for location source retrieval was not significant in item memory. The spatial and temporal distributive characteristics of the old/new effect for source memory were different from those of the previous studies. In addition, the old/new effects for target and non-target-old drawings were different.
Conclusions First, location source retrieval is more difficult than item memory for drawing, which supports dual-process model. Second, Experimental paradigm and source characteristics regulate the spatial and temporal distributions of source retrieval effect together. Third, the level of consciousness during retrieval is also an important factor affecting source retrieval

Key words: item memory, source memory, old/new effect