Lexical decision in aging: Stimulus variables and response time components

A combination of a few lines of work. Not only does this combine three lines of research topics (word retrieval, aging and diffusion modeling) but it is also a melding of two researcher's experiences. I bring experience in word retrieval and diffusion modeling while Dr. Gordon brings her experience with word retrieval and aging. This collaboration is the result of these combinations. The lexical decision task investigates what is important in a word which makes it easier or harder to recognize. In aging research, it is known that reaction times slow down with age, but what specific variables have the largest impact on reaction times and accuracy. The diffusion model on the other hand parses apart the reaction time into separate cognitive processes which give a better description of what is occurring in this decision making process. The following is an abstract that is being submitted for a conference in aging and speech communication.

It is known that different types of variables affect word retrieval, including lexical (e.g. frequency, length), subject (e.g. age, education) and task factors (e.g. mode of retrieval, composition of the stimulus set). Moreover, recent studies have demonstrated, through diffusion modeling, that reaction time can be decomposed into several processes, those related to the decision (stimulus information accumulation, response bias) and non-decision processes (encoding, response execution), which may be differentially affected. We tested older (>50 yrs.) and younger adults in an auditory lexical decision task to investigate how aging affects these different aspects of word retrieval. For both word and non-word responses, older subjects were both slower (Word: 71ms, p=0.016; Non-word: 159ms, p=0.003) and less accurate (Word: 0.8%, p=0.038; Non-word: 4.4%, p=0.003) than younger subjects. Both older and younger subjects were affected by the length and phonological neighborhood density of the word and non-word targets, and by the frequency of the word targets. Measures of phonotactic probability tended to influence the non-word responses of younger subjects, but the word responses of older subjects. This stronger effect of sub-lexical factors on word recognition is likely due to declines in peripheral processes (hearing acuity, speed of processing). In the diffusion model, older subjects had a significantly lower rate of information accumulation for both word (p=0.005) and non-word (p=0.0003) responses. For non-word responses, older subjects also took longer for non-decision processes (p=0.007). These results help clarify the cognitive and non-cognitive factors that contribute to word recognition, and how these change with age.

Following analyses will investigate how subject variables such as hearing acuity affect the behavioral variables. This investigation will also be broadened to include aphasic patients. This is the second project that I have submitted to a conference. I should find out in early August whether I am able to attend the conferences and more importantly, whether I am awarded money to travel to the respective conferences. More to come...