Psychiatry: Memory

Memory

Memory functioning – broadly defined as the storage and recall of past experience – is of major interest and importance in cognitive neuroscience. The intensity of research in the area of memory has been especially great in the past decade, when the fields of cogni-tive psychology and neuropsychology converged to offer unified models of memory functioning on the basis of experimentation in normal and brain-damaged humans, as well as lesion studies in animals. Functional neuroimaging has also begun to yield valu-able insights into the patterns of neuroanatomical connectivity involved in aspects of memory. Because this field is relativelmassive, the reader is referred to comprehensive sources for fur-ther details (Christianson, 1992; Squire and Butters, 1992).

Historically, the debate between the localizationist and ho-listic views of neural processing has been nowhere sharper than in the search for regionally specific entities involved in memory. The search for the “engram” or memory trace led Lashley (1950) to conclude that memories are distributed throughout the brain, not localized within a discrete or unitary structure. The debate has been given a more sophisticated focus with the realization that although there is regional specificity in aspects of memory, the complexity of the processes requires multiple brain regions, probably operating in a parallel fashion.

The initial contributions to the present knowledge of mem-ory were from two methodologies. These are the lesion analytical method and cognitive studies of normal individuals. The lesion analytical method has contributed the most to an understandingof both the cognitive and neuroanatomical aspects of memory functioning.

In the last 10 years, functional neuroimaging techniques (particularly positron emission tomography [PET] and functional magnetic resonance imaging [fMRI]) have become increasingly popular and provide an alternative method of studying human cognition. These techniques have the dual potential to validate existing models of memory and to suggest newer theories based on patterns of anatomical connectivity observed during the performance of memory tasks. Furthermore, unlike the lesion analytical method, they afford the ability to observe cognitive processes in healthy human brains. Neither PET nor fMRI di-rectly measures neural activity; rather, they rely on changes in regional cerebral blood flow (rCBF) that result from the meta-bolic demands of neural activity. As neural activity occurs, the demand for oxygen causes an increase in the relative concentra-tion of deoxyhemoglobin in the surrounding region.

There are several things to keep in mind when consider-ing findings from neuroimaging studies. First, it is important to remember that PET and fMRI are based on relative blood flow. Therefore, the baseline task subtracted from the events of interest must be tightly controlled to ensure that the finding is as specific as possible. Furthermore, although functional neuroimaging can suggest brain regions functionally involved in performing a cog-nitive operation, blood flow studies cannot suggest what is unique about that brain region. Also, when distributed brain regions are simultaneously activated, the technique itself does not immedi-ately reveal the functional connectivity between those regions. Unlike the electrophysiological studies of working memory, im-aging cannot suggest whether a brain region is in a sense storing information or transmitting information or is in some other way involved in the performance of an operation.