Tests of pattern separation and pattern completion in humans-A systematic review.

Abstract

To systematically review the characteristics, validity and outcome measures of tasks that have been described in the literature as assessing pattern separation and pattern completion in humans. Electronic databases were searched for articles. Parameters for task validity were obtained from two reviews that described optimal task design factors to evaluate pattern separation and pattern completion processes. These were that pattern separation should be tested during an encoding task using abstract, never-before-seen visual stimuli, and pattern completion during a retrieval task using partial cues; parametric alteration of the degree of interference of stimuli or degradation of cues should be used to generate a corresponding gradient in behavioral output; studies should explicitly identify the specific memory domain under investigation (sensory/perceptual, temporal, spatial, affect, response, or language) and account for the contribution of other potential attributes involved in performance of the task. A systematic, qualitative assessment of validity in relation to these parameters was performed, along with a review of general validity and task outcome measures. Sixty-two studies were included. The majority of studies investigated pattern separation and most tasks were performed on young, healthy adults. Pattern separation and pattern completion were most frequently tested during a retrieval task using familiar or recognizable visual stimuli and cues. Not all studies parametrically altered the degree of stimulus interference or cue degradation, or controlled for potential confounding factors. This review found evidence that some of the parameters for task validity have been followed in some human studies of pattern separation and pattern completion, but no study was judged to have adequately met all the parameters for task validity. The contribution of these parameters and other task design factors towards an optimal behavioral paradigm is discussed and recommendations for future research are made. © 2015 Wiley Periodicals, Inc.