The success of episodic memory retrieval is reliant on the degree of similarity between encoding conditions and cognitive operations at retrieval (‘encoding specificity’) and thereby the ability to flexibly adopt a retrieval orientation which biases retrieval cue processing in line with the specific goal of retrieval is associated with better memory performance. The present study adapted Jacoby et al.’s (2005) memory-for-foils paradigm to investigate whether agreement between retrieval orientation at encoding (picture-target vs. spoken word-target) and stimulus material at test (pictures vs. spoken words) produced superior memory for foils in young subjects. A series of neuropsychological tests assessing frontal function, long-term memory and intelligence were also administered to subjects. No difference in memory for foils between conditions was evident, a result attributed to a lack of retrieval selectivity demanded by the experimental tasks. Memory for foils performance was however correlated with level of frontal function in subjects. Poorer frontal function was associated with higher recognition errors and lower successful recognition of mismatched items in which demand for retrieval constraint is presumed to be greater. The use of the memory-for-foils paradigm for future research into source-constrained retrieval cue-processing is discussed.