Integrative neurocomputational perspectives on cognitive aging, neuromodulation, and representation

Besides neuroanatomical changes, neuromodulatory mechanisms are also compromised during aging. Neural network models are suitable tools for exploring the relatively broad and homogenous neuromodulatory influences on cortical function. Computational approaches for understanding neuromodulation of the dynamic properties of cortical function and recent neurocomputational theories relating different aspects of cognitive aging with declines in neuromodulation are reviewed. Considered within an integrative cross-level neurocomputational framework, aging-related decline in dopaminergic neurornodulation reduces the fidelity of neural information and gives rise to less distinctive neural pattern representations that may underlie various facets of aging cognitive and, possibly also, sensorimotor phenomena. (C) 2002 Elsevier Science Ltd. All rights reserved.