Grounded (embodied) theories of cognition propose that memory, including knowledge and meaning, is grounded in sensorimotor and mental state processes. The main proposed mechanism for how memory is grounded is mental simulation. Simulation occurs when neural activity in modal association cortex triggers time-locked, recurrent and feedback activity across multiple lower-level modal processing areas from which the memory was initially constructed. Through this distributed multi-regional activity, seeing an object or reading its name (e.g., “dog”) re-enacts associated features that were stored during earlier learning experiences (e.g. its shape, color, motion, actions with it), thereby constructing cognition, memory, and meaning. This paper reviews convergent evidence from cognitive neuroscience of mental imagery, object cognition, and memory that supports a multi-state interactive (MUSI) account of automatic and strategic mental simulation mechanisms that can ground memory, including the meaning, of objects in modal processing of visual features.