Would you give us a brief summary of your paper and describe its significance for the field?

A major goal of the article was to attempt to specify what might be responsible for age-related differences in many different types of cognitive variables. Two simple mechanisms were proposed: limited time, referring to the idea that when the rate of processing was slow some relevant operations would not be able to be completed in the available time; and simultaneity, referring to the idea that slow processing could lead to a decrease in synchronization of processing such that intermediate products of processing might no longer be available by the time that later processing was completed. I hypothesized that these two mechanisms might account for age-related human processing limitations in a manner somewhat analogous to the limitations of a slow versus a fast computer in a real-time environment. Although I was unable to assess the mechanisms directly, I reasoned that estimates of an individual’s speed of processing might be obtained from his or her performance on a number of relatively simple speeded tasks. Consistent with the hypothesis, application of several different types of statistical control procedures revealed that the age-related differences in many cognitive variables were substantially reduced after eliminating the variation in these measures of speed.

  What are the circumstances which led you to your work?

I began thinking about this issue when I realized that moderate to large age differences had been reported in many different types of cognitive variables, but that most of the attempts at explaining those differences relied on quite different types of mechanisms that were assumed to be largely independent of one another. For example, age differences in certain memory variables might be explained by less effective rehearsal, and age differences in reasoning measures might be explained by failure to identify and aggregate relevant information. Although it was clearly possible that different age-related influences operated on different cognitive variables, and that they were independent of one another, I felt that this was an assumption that should be investigated. Because variables assumed to reflect aspects of processing speed had some of the strongest relations to age, I focused on processing speed as a potential mediator of at least some of the adult age differences in cognitive functioning.

Although articles are still being published related to the finding that statistical control of measures of speed generally attenuate the relations between age and other cognitive variables, much of the contemporary research in my laboratory and in other laboratories has relied on a broader multivariate perspective in which several theoretical constructs are examined simultaneously. The goal of this more recent research is not to investigate the plausibility of a single theoretical mechanism or explanatory construct, but rather to determine how several constructs jointly contribute to the age differences on a variety of cognitive variables. I believe that the research summarized in the 1996 article was important for demonstrating that age-related influences on different cognitive variables are unlikely to be completely independent of one another because eliminating the variation in one type of variable (in this case measures of speed) substantially reduced the age-related effects on other types of variables. However, processing speed is only one of many distinct cognitive abilities, and much of the current research in this field considers speed influences in the context of influences of other cognitive abilities on adult age differences in measures of cognitive functioning.

  Are you still involved with this research? Where do you see it going 10 years from now?

I am still very much involved in research concerned with the fundamental question of what is responsible for adult age differences in cognitive functioning. As noted above, the specific nature of the research has changed somewhat, but I still believe that a productive approach to understanding the phenomenon of cognitive aging is to consider the age-related influences on the variable or construct of primary interest in the context of age-related influences on other variables and constructs. In the future I expect that neurobiological variables such as estimates of the volumes of brain regions from MRI, of regional brain activation patterns from functional neuroimaging, of neurotransmitter quantity, and of myelin integrity will be combined with cognitive variables to allow more comprehensive analyses of the interrelations of age-related influences on cognitive functioning.