Dynamic human anatomy rugg whiting 2012 pdf download

“The Scaffolding Theory of Aging and Cognition (STAC)”, proposed in 2009, is a conceptual model of cognitive aging that integrated evidence from structural and functional neuroimaging to explain how the combined effects of adverse and compensatory neural processes produce varying levels of cognitive function. The model made clear and testable predictions about how different brain variables, both structural and functional, were related to cognitive function, focusing on the core construct of compensatory scaffolding. The present paper provides a revised model that integrates new evidence about the aging brain that has emerged since STAC was published 5 years ago. Unlike the original STAC model, STAC-r incorporates life-course factors that serve to enhance or deplete neural resources, thereby influencing the developmental course of brain structure and function, as well as cognition, over time. Life-course factors also influence compensatory processes that are engaged to meet cognitive challenge, and to ameliorate the adverse effects of structural and functional decline. The revised model is discussed in relation to recent lifespan and longitudinal data as well as emerging evidence about the effects of training interventions. STAC-r goes beyond the previous model by combining a life-span approach with a life-course approach to understand and predict cognitive status and rate of cognitive change over time.

Decades of behavioral research in the latter part of the 20th century characterized a variety of age-related cognitive deficits including memory problems, executive processing dysfunction and declines in speed of processing that typify normal older adults (e.g., Craik and Salthouse 2000). Despite volumes of performance data and numerous theoretical advances (e.g., Schaie et al. 1996; Schaie and Willis 2011a, b; Birren and Schaie 2005), a coherent integrated account of cognitive aging based on behavioral data alone proved to be elusive. Fortunately, the end of the last century also brought major developments in in vivo human neuroscience methods, most critically, functional and structural imaging that permitted scientists to relate neural activity and structural brain measurements to specific cognitive processing abilities (Cabeza et al. 2005). Additional and more recent advances in imaging of white matter pathways, amyloid deposits, connectivity patterns, genetic, pharmacological and other biomarkers have provided a wealth of new indices of neurophysiological status that can be integrated with behavioral performance assessments to identify the neurocognitive underpinnings of typical age-related decline (Grady 2008; Buckner et al. 2009; Bäckman et al. 2006; Raz and Lustig 2014; Laukka et al. 2013).

In 2009 we published a model, which we referred to as the Scaffolding Theory of Aging and Cognition—“STAC” for short (Park and Reuter-Lorenz 2009). STAC aimed to explain age differences in cognitive function by incorporating the effects of a broad range of adverse biological and neurophysiological factors that had been associated with normal aging to date, and to delineate their dynamic interaction with protective factors and newly emerging, putative compensatory processes deemed to be at work in the older brain. While the model was originally developed in the context of cross-sectional studies comparing extreme groups of younger and older adults, it incorporated principles that were likely to be at play across the lifespan. The goal of the present review is to re-evaluate and revise STAC in view of new meta-analyses, lifespan (i.e., cross-sectional across adulthood) and longitudinal data that have been published since the model was conceived. We also consider new evidence about the effects of cognitive training and lifestyle factors in the context of STAC—these were identified as “future issues” in 2009, as relevant data were limited at that time.

The STAC model as originally conceived and depicted in Fig. 1, includes the following basic principles to explain an older individual’s level of cognitive function. First, relative to younger adults, healthy older adults are affected by varying degrees of neural degradation, which were categorized as “neural challenges” and “functional deterioration,” respectively. Neural challenges are primarily structural changes in the brain that occur with age, including cortical thinning and regional atrophy, loss of white matter integrity, and dopamine depletion. Functional deterioration refers to indicators of maladaptive, age-related brain activity that have been very well documented in the imaging literature including dedifferentiation (decreased specificity) of ventral-visual and motor areas (Park et al. 2004; Voss et al. 2008; Bernard and Seidler 2012), decreased memory-related recruitment of medial temporal lobe regions (Cabeza et al. 2004; Gutchess et al. 2005) and dysregulation of the default mode network (Lustig et al. 2003; Persson et al. 2007; for a review, see Park and Reuter-Lorenz 2009; Reuter-Lorenz and Park 2010).

Second, according to STAC, the level of cognitive function an individual displays is a consequence of these negative indices, combined with a beneficial process, which we term “compensatory scaffolding.” Compensatory scaffolding operates to ameliorate or counteract the adverse effects of neural and functional decline, and can be considered a form of “positive” plasticity that accompanies aging, whereas the adverse changes in brain structure that occur with age are negative forms of plasticity (Cramer et al. 2011; see also, Greenwood 2007). More specifically, scaffolds entail the engagement of supplementary neural circuitry that provides the additional computational support required by an aging brain to preserve cognitive function in the face of localized or global neurofunctional decline. Indications of compensatory scaffolding evident in the neuroimaging literature include greater activation or additional recruitment of prefrontal brain regions (Gutchess et al. 2005; Davis et al. 2008), compared to young adults, an effect now documented in parietal regions as well (e.g., Angel et al. 2011; Huang et al. 2012). Overactivation can also take the form of bilateral recruitment, where older adults activate left and right brain (sometimes homologous) regions on tasks for which younger adults show lateralized activity (Cabeza 2002; Reuter-Lorenz et al. 1999; Reuter-Lorenz et al. 2000; Tyler et al. 2010; Cappell et al. 2010; Schneider-Garces et al. 2010; for a review see, Cabeza and Dennis 2012). Meta-analytic evidence has now verified the pervasiveness and reliability of age-related overactivation in cross- sectional studies of younger and older adults, across a wide range of task domains including perceptual, memory and executive function tasks (Spreng et al. 2010). We note speculatively that neurogenesis, while more limited in older adults, is also a potential source of positive plasticity that may contribute to compensatory scaffolding (Fuchs and Flügge 2014; Lovden et al. 2013).

STAC represents the brain as a dynamically adaptive structure that changes in both positive and negative ways with age. Figure 1 indicates that both neural challenge and deterioration will stimulate scaffolding, which in turn moderates the effects of deleterious brain influences on cognitive performance. While typical age-related changes in brain structure and function can stimulate compensatory scaffolding, very severe deterioration can eventually undermine the brain’s ability to provide effective compensation. Finally, the model suggests that it is possible to enhance neural scaffolding activity by some explicit interventions that include various lifestyle activities including exercise, intellectual engagement and new learning, as well as more formal cognitive training interventions.

STAC was proposed as a neurofunctional account detailing the effects of age on cognition, most of which had been established from cross-sectional studies that compared extreme age groups of younger and older adults. However, continuous intra-individual lifespan principles related to compensatory scaffolding are inherent in the theory. In particular, the notion of scaffolding itself comes from cognitive development and skill acquisition research, which has demonstrated that existing mental abilities can be harnessed as support for the acquisition of new ones. The work of Petersen and colleagues (Petersen et al. 1998) was especially influential in the development of STAC. Most important was their observation that during early stages of skill acquisition, a network including prefrontal regions was very active, but activity in these regions decreased as performance became more skilled and activity increased in new skill-specific regions of the brain. They interpreted the initial but transient set of activations as providing scaffolding for the acquisition of novel skills, with the activation shifting elsewhere as skill increased. The ideas for STAC also drew upon evidence that with greater task demand, also considered a form of neural challenge, younger adults show increased activation of primary task regions, recruitment of additional brain regions or both, typically involving regions of prefrontal cortex that mediate executive functions (Reuter-Lorenz and Lustig 2005; Reuter-Lorenz and Cappell 2008). These lines of evidence suggested that the brain possesses particular adaptive neurocognitive “strategies” that are manifested under conditions of cognitive and behavioral challenge, and that similar mechanisms can be adopted with age to preserve established skills or to maintain optimal performance.

The STAC model as depicted in Fig. 1 accounts for individual differences in level of cognitive functioning at one specific time point, presumably in later adulthood. Over the past 5 years there has been an increase in new data sets concerning neurocognitive function in middle age, longitudinal age-related change, and intervention studies indicating later-life plasticity in response to experience. The goal of the present paper is to revisit STAC, and the scaffolding construct, and to re-evaluate the model in light of new developments in the field. We are particularly concerned with addressing emerging evidence about longitudinal influences on neural structure and function across the lifespan. We also consider new evidence about the effects of genetics, health, experiential and life-style variables on cognition, as these approaches were not integrated into the original STAC model. Based on these considerations, we propose a revised model, “STAC-r,” that integrates new data and knowledge about neurocognitive aging.

The concept of compensatory scaffolding is at the heart of the original STAC model. The notion that compensatory and supportive neural mechanisms enable maintenance of cognitive function with age has received considerable support since the model was developed (e.g., Berlingeri et al. 2010; Burzynska et al. 2013; Davis et al. 2012; Geerligs et al. 2012; Vallesi et al. 2011; Nyberg et al. 2014; Chanraud et al. 2013; Davis et al. 2008), although the indicators and mechanisms of compensation continue to be debated (Cabeza and Dennis 2012; Fabiani 2012). Because compensatory scaffolding is a key component retained in the revised model, we first review some of its properties in light of recent emerging evidence and other concepts that have come to the fore since 2009, and then introduce the revised model.

Brain Maintenance

According to STAC, neural challenge in the form of neurophysiological deterioration or neural insults that come with age, rather than age itself, are the impetus for compensatory scaffolding. Therefore and quite logically, older individuals who maintain a youthful neurobiological status, through favorable genetics, environmental factors, lifelong pro-health behaviors and beneficial lifestyle activities (Hillman et al. 2008; Josefsson et al. 2012; Vemuri et al. 2012) will need less compensatory scaffolding and reorganization. This axiom is captured by the term “brain maintenance” (Nyberg et al. 2012), an elegant and important construct that argues that a key characteristic of successful aging is simply the absence of age-related pathology. In terms of brain activation, one would expect that older adults who exhibited preservation of cognitive function in some domains would show more “youth-like” brain patterns, and minimal overactivation, and other age-specific neural indicators potentially characteristic of less compensatory scaffolding.

Several recent studies relating performance to fMRI activity and other measures of neurophysiological integrity bear this out (e.g., Nagel et al. 2011; Rosano et al. 2012). For example, using brain activation patterns obtained during a picture memory encoding task, Duzel and colleagues (2011) identified a subgroup of older adults whose activity profiles were virtually indistinguishable from those associated with encoding success in younger adults (Duzel et al. 2011). This “youth-like” older subgroup had recollection memory performance that was also indistinguishable from the younger group, while showing no evidence of prefrontal over-activity often taken to be compensatory. Thus, their preserved memory was more likely due to preserved neurobiology than to compensation. Consistent with this interpretation, the older subgroup whose encoding-related activation deviated most from the youth-like pattern had poorer memory, along with more default network dysregulation, and region-specific gray matter loss. Interestingly, Duzel et al. (2011), as the STAC model would predict, also identified a subgroup of older adults with preserved memory and overactivation of prefrontal and parietal regions relative to the younger group, which may have provided compensatory support for weaker engagement of other memory-dedicated circuitry. Their data, in accordance with STAC, suggest that older adults may achieve preserved cognition by means of preserved neurobiology, compensatory processes, or some combination of these factors.

If a hallmark of successful cognitive aging is maintenance of abilities and underlying neurobiology, then its assessment requires longitudinal measurements to evaluate the degree of change over time. Data of this type is just beginning to emerge. One important study (Pudas et al. 2013) classified individuals (ages 55–75) as having preserved memory versus average memory decline over a prior 15–20 year period, and used fMRI to assess their current brain activity profiles during a memory task. The comparison group with average memory decline had lower hippocampal activity than the successful agers, and a young comparison group. The successful group had greater hippocampal activity and more prefrontal activity (including left and right inferior frontal gyrus, IFG) than both young and average-old comparison groups (c.f., Persson et al. 2012). The two groups of older adults showed no differences in regional brain volumes or white matter integrity. So, as the authors point out, while overactivation in the successful agers is consistent with compensation, greater prefrontal and hippocampal engagement might have been characteristic of these people since an early age. More studies of this type are needed to clarify the contributions of brain maintenance and compensatory processes to sustained levels of high performance over time.

Brain Efficiency

Another notion entailed in the compensatory scaffolding component of STAC is efficiency of brain function (Duverne et al. 2008; Rypma et al. 2006; Rypma et al. 2007; Rypma and D’Esposito 1999; see also Neubauer and Fink 2009). Brain efficiency figures into the conceptualization of STAC in two important ways that have been advanced by research over the past 5 years (cf. Poldrack 2014). First, neurophysiological decline can lead to reduced efficiency, meaning that the rate and/or quality of neural processing (e.g., signal to noise ratio, fidelity of representations, speed of neural transmission) is reduced in association with aspects of perceptual and memory encoding, dedifferentiation, and poor default network regulation. As noted above, accumulating evidence continues to support these sources of dysfunction (Carp et al. 2011; Bernard and Seidler 2012; Garrett et al. 2013; Barulli and Stern 2013). In addition, structural decline, in the form of gray matter loss for example, could also be associated with reduced efficiency, resulting in a compensatory response in associated networks.

Recent work by Tyler and her colleagues (Tyler et al. 2010; Meunier et al. 2014; Shafto et al. 2012) is especially relevant for linking age-related neural decline to compensation and adaptivity of language processing networks. They report, for example, that older adults show marked gray matter loss in left lateralized regions specialized for syntactic processing in young adults, especially left IFG (Tyler et al. 2010). The greater the age-related volume loss in this region, the greater the recruitment of right IFG (and right temporal regions) and the more correlated its activity with left IFG. Critically, performance on syntactically demanding tasks was found to be age-equivalent, consistent with the idea that the recruitment of right hemisphere circuitry provides compensation for the declining left hemisphere regions specialized for language.

Gray matter reductions are not always detected in older adults however, due perhaps to methodological factors or subject variability, and atrophy does not always or fully correspond with regions that show activation changes measured with fMRI (e.g., Kalpouzos et al. 2012; Maillet and Rajah 2013; Chen et al. 2011). Furthermore, as Poldrack (2014) has pointed out, univariate fMRI indices of reduced or increased activation are ambiguous with respect to the “energy expenditure” of the neural system. While investigating the effect of parametric variations in task demand on activation levels can help to interpret group differences in activation levels (Reuter-Lorenz and Cappell 2008), multimodal imaging approaches and network analyses will become increasingly useful for clarifying how aging affects the efficiency of neural systems, and in turn drives compensation.

The second way that the idea of efficiency is relevant to STAC pertains to the efficiency of scaffolded networks. According to the model, while compensatory processes are proposed to assist (or attempt to assist) with computations mediated by the primary network, they are less efficient than primary networks in their youthful state. New evidence consistent with this proposition (Meunier et al. 2014) again comes from studies of language. Using fine-grained analyses of gray matter density Meunier et al. (2014) observed localized decreases that were interpreted to drive changes in network functional connectivity including the recruitment of additional right hemisphere circuitry during syntactic processing (Meunier et al. 2014). Moreover, using graph theoretic connectivity measures of network efficiency, Meunier and his colleagues showed that with these additional regions, network efficiency was lower in older adults. In this case, greater involvement of right hemisphere regions was also associated with poorer syntactic processing, raising questions about whether compensation is the most fitting interpretation of the function being served by these areas (Meunier et al. 2014) or whether dedifferentiation is more accurate.

What Neural Processes are Meditated by Compensation?

A fundamental question that arises in relation to the notion of scaffolding and compensation more generally is: What processes are being carried out by the additional regions or circuitry recruited to support the primary network? Are these additional regions assisting with the same neural computations conducted by the dedicated areas, and therefore the same cognitive strategies, do they provide alternative routes to achieve the same strategy, or do they mediate different strategies altogether, as has been suggested in some language domains (Shafto et al. 2012)? These interesting and fundamental questions are highly unlikely to have the same answer for every task circumstance or for each cognitive domain in which scaffolding may be evident, but the STAC-r model is sufficiently flexible to allow for such differences.

There are indications, for example, that some cognitive processes may be aided by recruiting domain-general executive control and working memory circuitry, providing a way to “off-load” high neural demands when a task requires high resource expenditure (e.g., Simmonds et al. 2008). A meta-analysis of response inhibition tasks in young adults confirmed a dominant role of right IFG in this process, but in addition found that for tasks with complex rules, right dorsolateral prefrontal cortex (PFC) regions were also active (Simmonds et al. 2008). Critically, when rule complexity has been varied, older adults have been found to recruit additional PFC regions during response inhibition tasks at lower levels of demand than younger adults (Vallesi et al. 2011). Related effects have been observed in the context of semantic processing, where additional domain-general executive control may be recruited by high-performing older adults (Peelle et al. 2013). These effects resemble the recruitment of additional prefrontal circuitry by older adults at lower levels of working memory task demand (compensation-related utilization of neural circuits hypothesis, CRUNCH, Reuter-Lorenz and Cappell 2008; see also Cappell et al 2010; Schneider-Garces et al. 2010). Multi-voxel pattern analyses have demonstrated that additional recruitment by older adults may indeed reflect greater reliance on domain-general resources at lower levels of demand than younger adults (Carp et al. 2010). In many instances however, there is insufficient information within a particular study to infer what functions are being served by additional regions of activity (or heightened connectivity) in older adults, so answers to these questions await future research.

Research in the future is also likely to focus further on individual differences in the use of scaffolding for particular cognitive functions, and to clarify further the conditions under which scaffolding is advantageous, especially given the likelihood that that the best cognition and healthiest brains will be associated with minimal structural degradation and little need for compensatory activity (de Chastelaine et al. 2011). It will also be important to determine whether there is selective vulnerability to aging of different parcellated brain systems (Wig et al. 2014) or large-scale brain networks (Bressler and Menon 2010), as patterns of compensatory scaffolding are likely to be systematically related to the magnitude and locus of neural degradation. Along this line, recent work from the Park lab investigating a large lifespan sample of adults from age 20 to 89 (Park et al. 2013) found evidence that task-activated fronto-parietal regions associated with successful subsequent memory, showed age differences earlier in the lifespan than task negative regions associated with the default network. Thus, different networks may have different trajectories of age-related decline (see also, Grady et al. 2010). Park et al. (2013) also found that low ability adults showed differences in task negative activity early in the lifespan whereas high ability adults maintained levels of neural activation until old age (see also, Daffner et al. 2011). These findings generally support the importance of individual difference variables, such as cognitive ability, in understanding the range of neural activity associated with compensatory activations, as well as the importance of lifespan studies to fully understand the developmental trajectory of neurocognitive aging (see also, Nyberg et al. 2010).

The longitudinal trajectories of neural and cognitive change and variables that promote brain maintenance and decline are beginning to figure more prominently in studies of neurocognitive aging. Moreover, there is increasing interest in neural and cognitive function in middle age, which likely sets the stage for the course of aging later in life (Karlamangla et al. 2014). For example, it is important to understand to what extent cognitive status in late adulthood is determined by neurofunctional status and reliance on compensatory processes in early and middle adulthood (e.g., Borghesani et al. 2012; Macpherson et al. 2014; Schaie and Willis 2011a, b; Willis et al. 2010; Thambisetty et al. 2013). Do middle-aged adults who rely on compensation earlier in life than their age-matched peers go on to age more poorly, given that they show older-age brain function at a young age?

The STAC model predicts cognitive function at a single time point during an individual’s lifespan with a focus on later-life cognition. This was partially because, with “aging” itself as the primary input to the model, it was not possible to afford a role for experience, genetics, and environment to influence the course of aging and, in turn, level of cognitive function. The increasing evidence that these broad factors are important determinants of the trajectories of neural and cognitive function (e.g., Agrigoroaei and Lachman 2011; Albert et al. 1995; Bender and Raz 2012; Anstey and Cherbuin 2012; Anstey 2008; Boron et al. 2012; de Frias et al. 2014; Stiehler et al. 2009; Zanjani et al. 2013) provides a sound basis for revising STAC to recognize the life-course influences on neurocognitive aging. Thus the revised model, which we refer to as “STAC-r”, now incorporates life-course variables that impact structure and function of the aging brain (see Fig. 2). We use the term "life course” to mean the accumulation of experiences and states an individual has experienced from birth to death (Mayer 2002). The model indicates that both life-span (aging) and life-course (experience) variables impact the structure and function of the brain and also directly affect the development of compensatory scaffolding, a construct that retains the core features from the original model that were described above. The next sections, consider the new components of the STAC-r model.