Retention of memory for large-scale spaces

Abstract

This study empirically examined the retention of large-scale spatial memory, taking different types of spatial knowledge and levels of sense of direction into consideration. A total of 38 participants learned a route from a video and conducted spatial tasks immediately after learning the route and after 2 weeks or 3 months had passed. Results showed that spatial memory decayed over time, at a faster rate for the first 2-week period than for the subsequent period of up to 3 months, although it was not completely forgotten even after 3 months. The rate of forgetting differed depending on the type of knowledge, with landmark and route knowledge deteriorating at a much faster rate than survey knowledge. Sense of direction affected both the acquisition and the retention of survey knowledge. Survey knowledge by people with a good sense of direction was more accurate and decayed much less than that by people with a poor sense of direction.

Keywords:

• Forgetting
• Spatial knowledge
• Cognitive maps
• Internal representations
• Sense of direction

I thank W. Wen for her help with the route video and Y. Nagai for his assistance in data processing. Thanks are also due to the participants in the experiment.

I thank W. Wen for her help with the route video and Y. Nagai for his assistance in data processing. Thanks are also due to the participants in the experiment.

Notes

¹ Another possible reason for the inconsistency among past studies concerns differences in the instructions given to participants. For example, in some studies participants were told beforehand that they would be tested repeatedly or instructed explicitly to learn the layout of an experimental space, but not in other studies. Also, in some studies participants were exposed to the experimental space multiple times for repeated test sessions, which might have contributed to memory retention.

² Montello and Xiao (2011) reported a mean SBSOD score of 4.2 and a standard deviation of 1.1 for 550 respondents in five different cities in the US, Europe, and Asia (computed from the data shown in their Tables 1 and 2). Similar values for our participants indicate the validity of classifying and labelling the groups as the good- and poor-SOD groups.

³ Participants’ performance was compared to a chance value of 33% correct for route choice, an absolute error of 90° for direction estimation, and a bidimensional correlation of .44 for map sketching. The chance value for map sketching was obtained from a Monte Carlo simulation in which 1000 pairs of x- and y-coordinates were generated randomly for anchor points used in the bidimensional regression analysis.

⁴ Differences in forgetting rates between the 2-week and 3-month conditions were examined, in terms of the ratio of performances in the second and first test sessions, as done by Corazzini et al. (2008). The results were the same as those obtained from the mixed ANOVAs for each measure (i.e., significant differences for landmark knowledge, route knowledge in terms of topological accuracy, and survey knowledge in terms of directions), confirming the consistency between the ANOVAs and effect-size analyses.