Verbal memory in brain damaged patients under different conditions of retrieval aids: a study of frontal, temporal, and diencephalic damaged subjects.

The performance of 36 patients, divided into six groups, and 13 control subjects was investigated in paired-associate learning. The patients had right or left prefrontal, right or left anterior lateral temporopolar or medial temporal lobe damage, or lesions restricted to diencephalic areas. As tasks, two lists of paired words had to be learned, with the first list presenting only the word pairs, and the second one embedding the word pairs in sentences of a highly imaginable content. Recall consisted of immediate or delayed (48 hrs) free recall of the first list (condition I), of immediate or delayed recall of the second list (with visual imagery as a learning aid; condition II), and of cued recall (in which the sentence form was presented with a blank space where the word to be recalled had been previously; condition III). Control subjects clearly performed best under all conditions, manifesting a ceiling effect for the second and third ones under immediate recall. Among the brain-damaged groups the diencephalic subjects were poorest and gained only little from the aids given for learning and recall. Of the four patients with medial temporal lobe damage, those two with bilateral lesions were nearly as bad as the diencephalic lesioned subjects. The other patients were markedly inferior to the control subjects, but gained considerably under conditions II and III. These statements hold for immediate and delayed recall, though for the delayed recall conditions all groups showed a reduction in performance which amounted to roughly half of the values they had had under immediate recall. It is concluded that increasing the possibilities for depth of information processing assists brain damaged (as well as normal) subjects in verbal learning, but that the advantage of aiding them at the moment of encoding and retrieval is highest for patients with restricted lesions and/or with lesions not invading the two regions most regularly implicated in long-term information processing (the diencephalon and the medial temporal lobe area).