Please enable JavaScript.

Coggle requires JavaScript to display documents.

Cognitive Psychology and Aging - Question 3 Discuss two basic mechanisms…

- - - - Working memory effects can be attenuated by providing cuing and minimizing cognitive load
      - Evidence suggests that working memory can predict a great deal of age group differences in processing across cognitive tasks.
    - - Defined loosely, involves components of executive attention that is mediated by the anterior attention system, inhibition, task management, and set shifting.
      - Proposed the selective compensation method. OA are able to compensate in declines in maintenance through the utilization of executive resources. However, maintenance mechanisms cannot be utilized to compensate for deficits in EF.
      - Age differences are morel likely to emerge in nonverbal working memory tasks than verbal working memory tasks.
      - The maintenance of dissociation of subsystems appears to be maintained with age. Under dual-task presentation in both YA and OA, verbal WM memory tasks interfere more with the process of other verbal WM tasks relative to visuospatial WM tasks (and vice versa)
        
        Reuter-Lorenz and Sylvester (2005): Some neuroimaging data does suggest that with aging, the modularity of these systems is challenged as it becomes more likely that left and right prefrontal areas associated with these different types of functioning become less differentiated. However this could also be the result of hippocampus under-activation and relate to binding issues
    - - Support for the distinct process hypothesis primarily derives from behavior evidence. Related to aging, results generally show maintenance of STM with moderate declines in WM.
      - Reuter-Lorenz and Sylvester (2005): Individuals taking a neural network approach tend to lie more on the side of the single continuum hypothesis. This perspective believe that all tasks requiring the on-line short-term storage of limited amounts of information are measures of working memory, with the differences in performance being a reflection of the degree to which the tasks places demand on the executive processing operations.
        
        Support for the single continuum hypothesis lies in patterns of neural activation across the lifespan. OA tend to recruit areas thought to be representative of executive functioning in YA during difficult STM memory tasks.
        
        Across a variety of WM task, trends of neural activation as they relate to aging are highly variable. However, two common trends are that OA tend to activate PFC areas (especially the DLPFC) under lower task demands relative to YA. Additionally, this recruitment is more likely to be lateralized to the left hemisphere
  - - - Cherry, Park, Frieske, and Smith (1996): OA benefited more than YA in a task in which they were to process and later recall critical adjective with or without the presence of picturing portraying the context of the sentence.
    - - OA who spent less time allocating resources to text-level demands and relied more on top-down mechanisms tend to show better memory and comprehension of material
      - Federmeier et al. (2003): When processing discourse, OA generally exhibited delays in sensory and attentional neurophysiology relative to YA but neurophysiological processing of semantic information was not delayed.
- - - - Has been thought to potentially be the one factor that can account for nearly all variance in age-related differences in cognitive performance
        
        However, substantial evidences suggests that a single cause model cannot account for age differences across domains of cognition. These conclusions have been supported by work that suggests analytical bias towards a generalized slowing hypothesis (e.g. Cerella, 1985; Allen et al., 2001) and that more precise statistical techniques (i.e. SEM and variance partitioning methods vs. mediation models and Brinley plots).
      - Can result in OA achieving poorer performance due to not being able to complete the relevant operation in time allotted (i.e. the limited-time mechanisms) or because of earlier stages of processing failing to be completed by the time that later processing stages have commenced (i.e. simultaneity mechanism)
      - Frequently, research is conducted that finds cognitive slowing to play a role in individual domains being assessed.
        
        While this information is relevant and valuable to the individual domain, it presents only a piece of the picture.
        
        In order to form a more generalized slowing perspective, meta-analyses have been performed to unite the findings of many of these domain-specific studies
        
        However, they are limited in their insight due to these analyses having to piece together domain-specific data from different experimental methods in order to form an across-domain model.
        
        A preferred method utilizes several tasks, representing different cognitive functions, from the same participants (i.e. Hale et al., 1995; Salthouse, 2000; Salthouse, 2009) in order to provide a better picture for across domain cognitive slowing
        
        By measuring individual performance across an array of tasks, we are able to better address the role that cognitive slowing plays at both the individual and group level while removing potential confounds that may appear with varying experimental methods. This method frequently concludes there to be evidence for across-domain cognitive slowing.
        
        However, it should be noted that in these studies, processing speed has not been found to have an equal impact across all cognitive tasks per se, but rather show to have a mediating influence on these tasks that exceeds other factors.
      - Measured on simple tasks in which accuracy is guaranteed given enough time (e.g. digit symbol substitution, digit parity)
      - Woodruff-Pak (1997); Salthouse (2011): Many mechanisms have been proposed including decreases in brain volume, synaptic delay, reduction in frequency of alpha rhythms, and more recently, WMH and FA
    - - Dictates that most of age-related differences in reaction time across cognitive domains is due to a single causal factor that is thought to be the fundamental speed at which individuals are able to process informations
      - Cerella (1985): This hypothesis has received much support in that laboratory measures of processing speed have been found to predict up to 95% of the variance in cognitive performance in higher level tasks
        
        These relations could be spurious due to the strong correlations between processing speed and other fundamental variables such as sensory acuity, working memory, and response inhibition.
    - - Dictates that differences in reaction time performance are the result of orthogonal processes that do not have a singular underlying causal mechanism.
      - Traditionally, age group x task complexity trends have been interpreted as supporting the domain-specific hypothesis. However, a more loose definition of general slowing that allows for a differentiated pattern of associated between a singular slowing factor has reduced the value of these arguments
  - - - In the study, measures of comparison speed (letter and line pattern comparison tasks) along with measures of working memory (Computation Span and Listening Span), were given to all participants and these scores were then paired with scores on a multitude of verbal and visuo-spatial reasoning tasks.
      - Hierarchical regression were performed to analyze direct and mediating influences of age, perceptual comparison speed, and working memory on performance in the reasoning tasks.
        
        Suggested that performance in reasoning tasks has no direct relationship to age, only one that represents age related changes in either processing speed or working memory.
        
        Regression analysis, along with a subsequent path analysis, confirmed a large proportion of the age-related differences in working memory to be mediated by processing speed.
    - - Mediation models
        
        Several criterion variables are assumed to be mediated by a more fundamental variable (e.g. speed of processing, reasoning ability, working memory). The effect of these fundamental mediators are examined by controlling for them and examining the age relationships to various criterion variables.
        
        many variables can function as substantial mediators if they have a relatively large relationship to both age and criterion variables (e.g. sensory acuity, postural balance, working memory, inhibitory functioning, attentional capacity, hang grip strength, lower limb strength) (Salthouse, 2001; Hartley, 2006).
      - Shared influence models
        
        Often referred to as common cause factor (CCF) models, appear similar to mediation models. Where CCF models differ is in that unlike mediation models, all criterion variables are thought to be related to a single factor and all age-related effects are hypothesized to function only on what is common to all variables.
      - Hierarchical models
        
        Aims to address a hypothesis that agerelated influences operate at many levels.
        
        In hierarchical models, multiple levels of common cause and/or mediational variables create levels of impact with more fundamental variables having a stronger impact as all variables beneath them with either a direct relationship or indirect relationship through another variable are influenced.
    - - [p(Re) = c(1 – e[(a-t)/b]) for t > a; and p = 0 for t > or = a.] In this equation, p(Re) represents the proportion of recognized or recalled items, t represents presentation time, a represents the onset where performance begins to rise above 0, b represents the rate of approach, or the rate at which performance goes to the asymptote, and c represents the asymptote, or level of performance reached when given an unlimited amount of time (Verhaeghen et al., 1998)
      - The older adult group both took longer to reach optimal accuracy (represented by a more shallow slope) and was unable to reach the same accuracy rate as the younger group independent of time allotted for response (represented by the lower asymptote of the older adult equation). This trend was seen for both recall and recognition experiments.
        
        Suggests that age-related decline in processing speed can be measured with both time-limited and the simultaneity mechanism, but researchers should be wary of conclusions drawn from these paradigms as they seemingly represent different constructs. In addition, researchers need to be wary in using time-limited designs because as demonstrated, it is impossible to disentangle the mixture of time-limited and simultaneity mechanism effects when time to reach optimal accuracy is restricted
    - - Also demonstrates that an independent factors model, though not as commonly explored, is even more parsimonious as a hierarchical CCF model and for some datasets, explains variance just as well.
      - Structural modeling goes beyond what regression models are capable of in regard to interpreting the across-domain influence of cognitive slowing. What is gained in structural modeling is the ability to assess levels of influence.
      - Discusses the limitations of establishing structure amidst highly colinear variables. Extant methodologies are likely not sensitive enough to parse this shared variance into more than one factor, even if a multi-factor solution is the reality
    - - Salthouse (2001), Hartley (2006), Salthouse (2009): Similar trends are seen in other domains as cognitive slowing is purported to be responsible for or exacerbate deficits in levels of functioning in cognitive tasks as individuals age. Support for these trends can be found in such other domains as visualization, psychomotor speed, & verbal fluency, inhibitory function and attentional capacity, and reasoning and spatial orientation among others.
    - - Limited-time designs involve tasks with preset time limits. Participants are evaluated by their reaction times and correct responses. However, because time limits are occasionally reached, an incorrect or incomplete response may be due to either external time pressure or concurrent demands on processing.
      - Researchers may design experiments in which tasks are presented without time restriction. Experiments employing this design operate under the simultaneity mechanism. That is, in the information processing procedure, the products of earlier processing are lost before the time later processing can be completed.
        
        Failure to create a correct response in a task like this would point to the inability of an internal mechanism, such as cognitive slowing, rather than the possibility that an individual simply did not have enough time to process and determine a correct response.
    - - When beginning to interpret the Brinley plot, one can start by understanding that it represents a ratio in the performance of two groups.
        
        When performance is equivalent across age groups, the plotted function would have a slope of one
      - When a plotted function favors one axis (in this case older group mean reaction time), it represents a ratio that favors the axis.
        
        In Hale et al. (1995), the Brinley plot represented reaction times that are slower for the older group relative to the younger group
        
        Because these summary data points, which represent a wide variety of information processing tasks, lie so tightly upon these lines of best fit, it can also be concluded that performance in these different tasks are being mediated by the same construct, cognitive slowing (Lima et al., 1991).
        
        While polynomial regression of the same data yielded a significant term, suggesting the appropriateness of a power function over a linear equation, both models clearly support the data (this supports the idea another factor is at play such as working memory or response disinhibition)
        
        Perfect (1994): Argues that Brinley plots cannot differentiate the validity of a single vs. a multiple-factor model of the slowing of processing speed.
        
        When simple, single factor models account for large proportions of the total criterion variance, then small increase in predictive value by adding a second predictor seems redundant.
        
        Range restriction of reaction time values minimize the likelihood of a well-fitting multi-factor model. Conversely Analyzing trends across tasks that vary widely and have minimal overlap in their reaction time indices are more likely to support a single factor slowing model.
        
        Studies that fail to fit a single factor function (> 2 SD from y predicted) are exceedingly rare from a mechanical perspective, biasing conclusions towards a single factor slowing model.
        
        Arguing for a general slowing hypothesis is problematic if multiple predictor models are allowed to support a uniprocess slowing phenomenon. This is of even greater concern provided that several of these multiple predictor variables fit the data equally well.
    - - Data suggested that differences in slowing at the perceptual level are greater than the differences in slowing at the cognitive level
      - Suggests that a two-factor model, rather than a uniprocess generalized slowing model best fit the existing data