Virtual reality navigation deficits predict future brain volume loss

Alzheimer's disease (AD) often begins long before it is clinically recognized, with subtle brain changes emerging years before noticeable memory loss or cognitive decline. Among the earliest regions affected are the hippocampus and entorhinal cortex, areas essential for spatial navigation. This has led researchers to look beyond memory and explore navigation ability as a potential early indicator of the disease. One key component of navigation is path integration (PI), the brain's ability to track position and direction using internal cues such as movement and balance. As these systems begin to deteriorate, disruptions in PI may appear early, making it a promising behavioral marker of preclinical AD. Despite growing interest in this idea, a key question is whether PI deficits can predict structural brain changes.

To address this, a team of researchers led by Senior Assistant Professor Kazuya Kawabata, Dr. Sayuri Shima, and Prof. Hirohisa Watanabe from the Department of Neurology, Fujita Health University, Japan, conducted a study to examine whether subtle impairments in virtual-reality PI (VR-PI) could signal future neurodegeneration in cognitively healthy individuals. Their findings were published in the journal Alzheimer's Research & Therapy on April 20, 2026.

The study followed 71 cognitively unimpaired adults over approximately one year. At baseline, participants completed an immersive VR navigation task designed to assess PI ability. In this task, individuals navigated a circular virtual environment, visited two checkpoints, and were then asked to return to their starting point without visual cues. Two primary measures were derived: PI error (distance from the true starting point) and angular error (directional deviation). In addition, high-resolution magnetic resonance imaging (MRI) scans were analyzed to assess structural changes such as longitudinal cortical thickness and volume. Moreover, Alzheimer's-related plasma biomarkers, including p-tau181 and glial fibrillary acidic protein (GFAP), were assessed. Longitudinal brain changes were analyzed using linear mixed-effects models to evaluate whether baseline PI performance predicted structural decline.

The results revealed a clear pattern. Individuals with higher PI error at baseline showed greater cortical thinning and volume loss over the follow-up period. These changes were observed in brain regions known to be vulnerable in early AD, including the parahippocampal gyrus, middle temporal gyrus, posterior cingulate cortex, and caudal middle frontal gyrus.