Perceptual Learning with Complex Objects: A Comparison between Full-Practice Training and Memory Reactivation.

Perception improves with repeated exposure. Evidence has shown object recognition can be improved by training for multiple days in adults. In particular, a study of Amar-Halpert et al. (2017) has compared the learning effect of repetitive and brief, at-threshold training on a discrimination task and reported similar improvement in both groups. The finding is interpreted as evidence that memory reactivation benefits discrimination learning. This raises the question how this process might influence different perceptual tasks, including tasks with more complex visual stimuli. Here, this preregistered study investigates whether reactivation induces improvements in a visual object learning task that includes more complex visual stimuli. Participants were trained to recognize a set of objects during five days of training. After the initial training, a group was trained with repeated practice, the other with brief, near-threshold reactivation trials. In both groups we found improved object recognition at brief exposure durations. Traditional intense training shows a daily improvement; however, the group with reactivation does not reach the same level of improvement. Our findings show that reactivation has a smaller effect relative to large amounts of practice.Significance Statement Perceptual learning helps to explore adult plasticity in visual processing. Gradual improvements in the perception of complex objects have been demonstrated across multiple daily training sessions of hundreds of trials. These improvements in the trained objects and the transfer to new objects, in that sense, support "practice makes perfect." Recent research challenges this idea, and suggests that a few critical reactivation trials can boost the learning processes. Here, we extend this idea to other learning tasks and investigate the extent to which short reactivation with a small number of trials can replace extensive training with complex visual objects. In our paradigm, we found larger training effects with extensive training.