We investigate verbal learning and cognitive map development of simulated layouts using a non-visual interface called a virtual verbal display (VVD). Previous studies have questioned the efficacy of VVDs in supporting cognitive mapping (Giudice, Bakdash, Legge, & Roy, in revision). Two factors of interface fidelity are investigated which could account for this deficit, spatial language vs. spatialized audio and physical vs. imagined rotation. During training, participants used the VVD (Experiments 1 and 2) or a visual display (Experiment 3) to explore unfamiliar computer-based layouts and seek-out target locations. At test, participants performed a wayfinding task between targets in the corresponding real environment. Results demonstrated that only spatialized audio in the VVD improved wayfinding behavior, yielding almost identical performance as was found in the visual condition. These findings suggest that learning with both modalities led to comparable cognitive maps and demonstrate the importance of incorporating spatial cues in verbal displays.