A validation study published in The Journal of Medical Internet Research depicted that VStore (a new virtual reality shopping task) is a beneficial assessment that engages standard cognitive domains and has a sensitivity to the age-linked cognitive deficit. Investigators aimed to establish construct validity of VStore with respect to established computerized cognitive battery, Cogstate, and determine its sensitivity to age-linked cognitive impairment in 142 healthy subjects (age 20-79 years).

The major VStore endpoints incorporated time to collect items, total completion time, verbal recall of twelve grocery items, time to make payment, time to order coffee, and time to select items on a self-checkout machine. The construct validity was assessed using a series of backward elimination regression models for best anticipation of the performance of VStore.

Furthermore, two logistic regression models and two ridge regression models accompanied with receiver operating characteristic curves were created, with Cogstate outcomes in 2nd model and VStore outcomes in 1st model incorporated as predictors of age cohorts and age, respectively.

The overall performance of VStore, as recorded by overall time spent finishing the task, was better elucidated by Cogstate tasks determining attention, working memory, age, paired associate learning, and technological familiarity (accounted for 47% of variance).

Furthermore, with λ=5.16, the ridge regression model chose five parameters for VStore when assessing age (mean squared error 185.80, SE 19.34), and with λ=9.49 for Cogstate, the model chose all eight tasks (mean squared error 226.80, SE 23.48). The VStore displayed high sensitivity (87%) and specificity (91.7%) to the age cohorts, with 94.6% of area under receiver operating characteristic curve.