Previous studies suggest that motor inhibition training (MIT) – i.e., associating appetitive stimuli with behaviour inhibition - can help change unhealthy eating behaviours (e.g., Jones et al., 2016). Yet, most studies use a control condition that associates appetitive foods with motor response. In theory, this motor response training (MRT) can promote unhealthy eating, precluding any conclusion about what is driving the effects of training. We aimed to disentangle the effects of MIT from MRT by including a control training condition without presenting foods (no-food training). The study included 180 female adults randomly assigned to one of three training conditions: MIT, MRT or no-food training. After training, we measured participants’ food choices, desire to eat and proposed moderators of training effects (e.g., hunger, dietary restraint). Linear models with training conditions as a between-subjects factor did not show an effect of training on trained food choices, F(2, 176) = .668, p = .514, and desire to eat, F(2, 176) = 1.391, p = .251. The average number of trained food choices was slightly lower for MIT (M = 1.69, SD = 1.15) than MRT (M = 1.87, SD = .91), but similar to the no-food training (M = 1.67, SD = 1.00). Moreover, the proposed moderators did not influence the effects of training. These results suggest that the effects of previous studies might be driven by MRT and the effects of MIT can be overestimated. We conclude that to assess the effectiveness of MIT it is crucial to include an appropriate control condition.

--