In this dissertation I investigate the effects of effort in speech comprehension as it relates to hearing loss and adult aging. Chapter 1 provides a general introduction to the topics related to speech comprehension in difficult listening environments. In Chapter 2, I demonstrate the effects of aging on memory for spoken passages. In normal aging, as expected, older adults do not recall as much information as their younger adult counterparts, however, the pattern of recall is similar for both groups. Moreover, although all participants demonstrated that they could hear the stimuli at the presentation level, hearing acuity, age and verbal ability contributed to recall performance.
In Chapter 3, I explore speech comprehension and recall using masking to produce degraded, yet intelligible stimuli. Behaviorally, I show that masking of a word reduced associations both to the masked word and between the prior words in the original list, but did not reduce recall of words that followed it. In simulations of several biologically grounded models, the behavioral data matches only with a linking-buffer model, in which the masked word disrupts a short-term memory buffer, while links to the masked word are also weakened. These data provide support for an effortful hypothesis, where distorted input has a detrimental impact on prior information stored in memory.
In Chapter 4, I manipulated memory load and syntax in a recall task to examine the effectiveness of the pupillary response as a measure of cognitive load in younger and older adults. Both age groups’ pupil sizes were sensitive to the size of the memory set and sentence length, with the older adults showing a larger effect of the memory load on a normalized measure of pupil size relative to the younger adults. By contrast, only the younger adults showed a difference in the pupillary response to a change in syntactic complexity, even with an adjustment for the reduced reactivity of the older pupil.
In Chapter 5, I investigate the role of prosody during the course of probabilistic decision-making related to the compatibility between verb bias and sentence structure first behaviorally, then using functional imaging. Behavioral results show that the removal of ordinarily available prosody exacerbates the processing difficulty associated with less compatible sentences. Imaging results show significant right DLPFC involvement when listeners must resolve the structure of a sentence with a less compatible temporary syntactic ambiguity, and this activation is augmented in the absence of a normal prosodic pattern. Taken together these findings support the position that syntactically-tied prosody facilitates online syntactic parsing in the presence of temporary syntactic ambiguity in spoken material. Moreover, sentence processing without natural syntactic prosody recruits brain areas associated with executive resources that are not ordinarily activated during core-syntactic processing.
Overall, these studies supports the effortfulness hypothesis, which proposes that the extra cognitive effort that is required in a difficult listening environments comes at the cost of processing resources that might otherwise be available for encoding the speech content in memory.