Snap Back to Reality: Comparing Real-World Spatial Memory in the "Wild" vs. in the Lab

Abstract

We thank the Department of Psychology at Colby College for supporting this work. Derek J. Huffman played a primary role in conceptualization, funding acquisition, project administration, methodology, resources, supervision, writing-initial draft, and writing-review and editing; and a supporting role in formal analysis, software, validation, and visualization. Paisley J. Annes played a lead role in validation and visualization; and a supporting role in conceptualization, data curation, formal analysis, methodology, software, writing-initial draft, and writing-review and editing. Chandrachud Gowda played a primary role in Investigation; and a supporting role in conceptualization, data curation, formal analysis, methodology, project administration, visualization, and software. Lior Colina played a supporting role in conceptualization, methodology, and software. Memory is most commonly studied using lab-based measures; however, spatial memory for large-scale environments, such as cities, typically takes extended time and motivation to learn, thus raising questions about the ecological validity of findings from traditional lab-based paradigms (e.g., information that is learned over a short timescale and with no real-world relevance). We leveraged technological breakthroughs to create a custom smartphone application to compare situated spatial memory with lab-based measures. Specifically, we compared performance in healthy young adults across two established lab-based tasks, judgments of relative direction (JRD) and map drawing, and our novel app-based, in situ pointing task for the same familiar, large-scale, real-world environment (the participants college campus). We observed a strong correlation between JRD performance and map-drawing performance. App-based pointing showed lower error and reduced inter-individual variability relative to JRD performance, but weak correlations with lab-based measures. We also developed a novel analytical technique in which we transformed the app-based pointing into a relational, JRD-like metric, and we observed strong correlations with the lab-based tasks. Moreover, we observed correlated patterns of errors across all tasks, thus suggesting that these tasks tap into partially overlapping cognitive representations. Our results provide a pivotal advancement to our understanding of both shared and unique variance across spatial memory paradigms and support the use and further development of mobile navigation tools as complements to lab-based assessments for studying spatial cognition and differences across populations, including patients with Alzheimers disease who often report getting disoriented even within familiar spatial environments.