Mnemonic Discrimination Language Evinces Recollection Rejection of Similar Lures

Mnemonic discrimination of visual objects entails differentiating among repetitions of target objects, unstudied foil objects, and, critically, lure objects that are similar, but not identical to, studied objects (e.g., a different coffee mug than what was studied). Correctly rejecting lures may involve hippocampal pattern separation, a process that orthogonalizes representations of similar experiences. However, lures can also be rejected when recollection of studied objects enables detection of changed lure features. The present study examined whether verbal justifications of recognition decisions in an object-based mnemonic discrimination task could reveal recollection-rejection as the primary basis for lure rejections. Across multiple study-test trials, participants studied everyday objects and, at test, attempted to classify similar lures, studied targets, and novel foils. Participants sometimes verbally justified their decisions. Machine learning classifiers showed that verbal justifications discriminated among different classifications given to the same item types for both in- and out-of-sample data. Lure rejection language often expressed the use of recollection of studied objects to detect changes in perceived objects (viz., recollection-rejection strategy). Verbal justifications also discriminated correct from incorrect responses better than numeric confidence, which could not be explained by a model assuming a one-dimensional memory strength signal. Finally, verbal justifications best predicted accurate recognition decisions for all item types at the highest level of subjective confidence, which further implicated the use of recollection. The present findings verify that lure rejections in mnemonic discrimination tasks do not only reflect hippocampal pattern separation and instead suggest that rejections also reflect recollection of studied targets.