Frontal cortex, laterality, and memory: encoding versus retrieval

The cerebral hemispheres differ in their capabilities and response to verbal versus nonverbal visual material. A priori, it might thus be expected that the right hemisphere would be best activated during a mnemonic task with fMRI when using nonverbalizable images, and the left hemisphere with verbal material. However, previous psychological tests had shown a high degree of similarity in measures of memory for these disparate items. It was thus hypothesized that extensive commonality in the areas activated would prevail when this previously tested material was employed with fMRI. Six subjects underwent fMRI with four types of trials in blocks: fixating; passively viewing 12 words and 12 nonverbalizable images; endeavoring to remember (encoding) another set of 12 words and images; endeavoring to recognize (retrieve) previously viewed words or images. Passive viewing produced small islands of activation in left versus right frontal cortex for words and images, respectively. Endeavoring to remember enlarged the areas of activation and produced some bilaterality. Retrieval greatly augmented activation as well as bilaterality, and some 20% of the activated frontal volume was shared by words and images. Thus, on the one hand, the distribution of activation upon retrieval differed substantially for words versus images, but on the other, as predicted, there was considerable commonality. Predominant laterality of activation in some areas shifted between encoding and retrieval (HERA), importantly involving different regions for words versus images. Of course, processes other than memory per se are undoubtedly involved in these distributions of fMRI activation in frontal cortex, yet the nature of the to-be-remembered items is clearly a major factor, in accord with the asymmetric lateralization in their basic representation.