Objectives: Withania somnifera Dunal (WS, ashwagandha) has been traditionally used as an adaptogen in Ayurveda.Administered over a long-term, it has demonstrated its potential as a remedy against age-and stress-related cognitive dysfunction. Aims of the Study:In this long-term study, the effects of WS alone and WS combined with whole food diet and voluntary wheel-running on age-related alterations in acetylcholine esterase (AchE) levels and in the density of the cholinergic muscarinic receptor subtype 1(M₁GPCR), combined with the downstream transcription factor phospho-CREB (P-CREB) were evaluated. Additionally, index of memory function of the animals (measured by Novel object recognition test, NOR) and spatial learning ability (Barnes Maze) was correlated to the aforementioned molecular outcomes.Methods: WS root powder was tested in the live rats by oral administration of 100mg/kg for 19 months, followed by novel object recognition (NOR) and Barnes Maze tests at 20 months of age. AchE activity in frontal cortex was measured via colorimetric assay; M₁-GPCR and P-CREB/total CREB density via Western blot (WB).Results: WS herbal supplement alone or in combination with additional lifelong lifestyle interventions did not lead to an increase in the levels of muscarinic acetylcholine receptor in the frontal cortex of the aged rats, did not affect AchE activity significantly, nor did it alter expression of P-CREB in the frontal cortex of the aged rat. However, aerobic exercise and healthy diet treatments led to significant decrease in AchE activity combined with an increase in P-CREB and M₁GPCR immuno reactivity. In addition, via NOR Discrimination Index (DI), WS-treated rats have the best object memory, have above average hippocampus-dependent spatial memory and frontal-cortex mediated learning flexibility; combined intervention did not improve object discrimination while it preserved comparable spatial memory. Additionally, exercised rats showed adequate mental flexibility in special tasks, but their object memory was below average. Linear regression analysis revealed a significant correlation between M1 receptor density and memory index; between M1 receptor density and AchE activity in WS and Combo-treated rats, but failed to reach significance for inverse correlation with latency to complete Barnes Maze spatial navigation task.