Research with the mismatch negativity (MMN) has revealed an informationally encapsulated system, presumably dependent upon primary as well as auditory association cortex, that detects change in the recent acoustic past on an automatic, preattentive basis. By contrast, early and late ERP signs of controlled, selective attention (processing negativities; PN), receive inputs from a widespread attentional network dependent upon auditory temporo-parietal and frontal cortical association areas. It has been hypothesized that the MMN system operates on two types of memories: 1) raw input stored in sensory memory (e.g., pitch and intensity), and 2) representations of invariances in the recent past, which entail memories based on transformations of the raw sensory input. The system's ability to learn, and to form traces based on abstract or conceptual stimulus features, has promoted investigators to suggest that the memory on which the MMN is based has more in common with """"""""implicit memory"""""""" than it does with sensory memory. Therefore, this area of research is a natural extension of the investigator's previous indirect memory studies with Alzheimer's disease patients. The requirement to selectively attend gives rise to early and late ERP components that reflect, respectively, selection of stimuli based on attributes that characterize stimuli in the attended channel, and selective rehearsal of these attributes. The major focus of this proposal is on a better understanding of auditory attentional mechanisms in probable Alzheimer's disease (PAD). Neuropathological data suggest that primary auditory cortex, upon which the generation of MMNs to simple physical differences depends, is intact in PAD, whereas temporo-parietal and frontal association areas, upon which early and late PNs depend, are damaged in PAD. These data lead to the expectation that ERP signs of pre-attentive auditory mechanisms based on raw sensory input may be relatively intact in mild and moderately demented patients, whereas those based on transformations of the raw input may not. Similarly, ERP signs of selective attention for a single, location, cue may be relatively intact in PAD, whereas those based on conjoint features will not. ERPs will be recorded from 62 scalp placements in order to determine whether MMN and PN component scalp distributions and by implication, their underlying neural sources, differ systematically between the patient and control groups and/or conditions of the experiments. The data will be relevant to preattentive processing, selective attention based on complex features, age-related pathological changes in these mechanisms, and their physiological underpinnings.
Showing the most recent 10 out of 25 publications
