Identifying retrieval problems using the California Verbal Learning Test.

Wilde, Boake, and Sherer (1995) examined the discrepancy between Long Delayed Free Recall (LDFR) and Recognition Discriminability (RD)--the California Verbal Learning Test (CVLT) index thought to indicate the presence of memory retrieval problems--and found little evidence to recommend its use in traumatic brain injury (TBI). The present investigation re-examined this index from the perspective of a continuum of retrieval deficit severity. CVLT performance was examined in 122 TBI patients, and 2 retrieval deficit indicators of varying severity were evaluated. Memory-impaired control groups were matched with retrieval deficit groups on initial acquisition and demographic characteristics. Individuals with a LDFR/RD discrepancy did not show predicted differences on other CVLT indices of retrieval problems, similar to the findings of Wilde et al. (1995). In contrast, individuals with a consistent discrepancy between free recall and semantic cued recall (Short and Long Delay) had greater improvement with recognition cueing and made fewer intrusive errors than controls. Individuals who benefited from semantic cues (where retrieval of the target word is still required) also benefited from recognition cues (where retrieval demands are minimal). Evidence supported the existence of a continuum of retrieval deficit severity. An LDFR/RD discrepancy without performance improvement from semantic cueing appears to indicate a more severe retrieval deficit, whereas performance improvement from both recognition and semantic cueing indicates less severe retrieval deficits.