Regular articleAge-related differences in resting-state and task-based network characteristics and cognition: a lifespan sample
Introduction
Older adults experience decline in many cognitive functions, including working memory, general processing speed (Park et al., 2002; Park and Reuter-Lorenz, 2009), language production (Burke and Shafto, 2008; Shafto et al., 2007), and cognitive control (Paxton et al., 2008; Schaie, 1996). Concurrent with age-related cognitive decline, older adults often experience neural decline and many studies have focused on the patterns of correlated activity in the human brain (i.e., functional connectivity). However, previous studies have largely focused on either task-based data or resting-state data. The difference between task-based and resting-state data may be of particular importance in the study of aging as age-related cognitive and neural differences may be more pronounced during a task (e.g., Davis et al., 2014). Moreover, aging studies most commonly focus on younger and older adults, leaving out middle-aged participants who represent a significant portion of the adult lifespan. Finally, most previous studies have focused on brain-behavior relationships within a single cognitive domain (e.g., King et al., 2018) or a single network (e.g., Andrews-Hanna et al., 2007), instead of a whole-brain level across multiple cognitive domains. To address these issues, we examined task-based and resting-state data collected from a large sample of individuals from across the lifespan and relate these neural measures to a broad assessment of cognitive abilities including executive function, recall, working memory, and language.
Functional connectivity analyses directly measure the temporal correlation of functional activations across the brain, and one hypothesis is that this coordinated pattern of activations may reflect how functionally specialized brain regions work together and interact (Friston, 1994). When brain regions have coordinated patterns of activation, they are said to form functional networks such as the default mode network (DMN) or the executive control network (ECN, Power et al., 2011), among others. Studies of functional connectivity are often based on “resting-state” functional Magnetic Resonance Imaging (fMRI) data, which is collected while participants are not performing an explicit task (Biswal et al., 1995) and reflects a default state of overall brain organization which may be related to task performance (Raichle et al., 2001; Raichle and Gusnard, 2005). In such resting-state studies older adults often show lower within-network connectivity, and this has primarily been demonstrated in the DMN (Betzel et al., 2014; Cao et al., 2014; Geerligs et al., 2015; Onoda et al., 2012; Siman-Tov et al., 2017; Song et al., 2014; Tomasi and Volkow, 2012). These reduced connectivities in network characteristics have often been associated with worse behavioral performance across different cognitive domains (King et al., 2018; Onoda et al., 2012; Sala-Llonch et al., 2015; Varangis et al., 2019a; Wang et al., 2010).
Although there have been many functional connectivity studies during resting-state, these typically focus on connectivity within a specific functional network in isolation. However, networks often interact with one another, so examining the relationships between networks across the whole brain is important in understanding overall brain functioning. Yet, only a few studies have examined how functional networks work together, and how this relates to age and cognition (Chan et al., 2014, 2018; Varangis et al., 2019a). In looking at network segregation, a measure that combines both within- and between-network connectivities to provide a measure of the degree to which different networks share connections, older age was associated with lower network segregation during resting-state (Chan et al., 2014; Varangis et al., 2019a). In addition, less segregated networks, such as the DMN and dorsal attentional network, were associated with worse episodic memory scores (Chan et al., 2014) and with worse performance on speed, fluid intelligence, and memory tasks (Varangis et al., 2019a). Interestingly, these network segregation-cognition relationships were independent of age, indicating that less differentiated brain networks are associated with lower cognitive functioning across the lifespan.
In addition to examining whole-brain network characteristics, the type of data one examines is also relevant. Studies have demonstrated that brain connectivity patterns are different when engaged in a task, compared with connectivity patterns observed during resting-state (e.g., within- and between-network connectivity and reorganization of communication hubs; Cole et al., 2014; Gonzalez-Castillo and Bandettini, 2018). This may be particularly relevant for aging research, as older adults often show larger age-related differences during laboratory-based tasks than more naturalistic tasks (e.g., Davis et al., 2014). Several studies have looked at task-based connectivity and its relationship with age and cognition when participants were engaged in certain tasks. They found that increased age was associated with decreased task-based, within-network connections in the DMN, as well as other networks such as within supplementary motor regions, the dorsal attentional network, and the somatomotor network (Andrews-Hanna et al., 2007; Geerligs et al., 2014; Steffener et al., 2012). In addition, reduced task-based, within-network connectivity reported in these studies were associated with poorer cognitive performance in domains such as executive function, memory, and processing speed. In addition, studies also reported that older adults showed different between-network connectivity patterns compared with younger adults (Geerligs et al., 2014; Varangis et al., 2019b). For instance, Varangis et al. (2019b) found that age may particularly affect between-network connectivity with the DMN, memory, and salience networks, however, the age effects were variable, with older adults showing higher between network connectivities between the frontoparietal network and the salience and memory networks and younger adults showing higher between-network connectivities between the DMN and the salience and memory networks. Collectively, these studies suggest that the effects of age on functional connectivities were similar between resting-state and task performance, such that older adults typically show weaker within-network connectivity and sometimes weaker between-network connectivity.
In sum, many existing studies have examined age-related differences in functional connectivity during rest or a task, consistently finding that older adults showed lower within-network connectivity than younger adults. In addition, lower within-network connectivity has been associated with worse behavioral performance across several different cognitive domains. However, the findings for age-related differences in between-network connectivity are less consistent. To date, very few studies have investigated both resting-state and task-based connectivity and their relationship with age and cognition (but see a study comparing adults and children, Hutchison and Morton, 2015). Moreover, the previous literature examining the relationships between age, cognition, and the brain has primarily relied on the differences between younger and older adults, in which significant differences in cognitive and brain functions are typical. Few studies have included a middle-aged population and investigated age-related differences in cognitive and brain functions across the lifespan (Chan et al., 2014; Varangis et al., 2019a). Therefore, examining differences between resting-state and task-based functional connectivity within subjects across the lifespan is essential to further our understanding of how brain-behavior relations differ with age, as well as with task demand. In addition, individual differences in cognition may also play a role in these relationships, highlighting the importance of considering such factors. The current study used a whole-brain network approach to investigate the effect of age on resting-state and task-based functional connectivity and its relationship with cognition across the adult lifespan (i.e., 20–75 years). In addition, given the potential effect of education and socioeconomic status on cognition and neurodevelopment (Braveman and Gottlieb, 2014; Chan et al., 2018; Hackman et al., 2010; Hurst et al., 2013; Wang and Geng, 2019), education was controlled for in all analyses. We predicted that increased age would be associated with lower within-network connectivity. However, given the inconsistent literature on age-related differences in between-network connectivity, either a positive or negative relationship between age and between-network connectivity could be possible. We were also interested in the relationship between these network measures and cognition to better understand if age-related differences in these metrics are compensatory or reflect decreased neural efficiency. For instance, higher between-network connectivity in older adults that relates to worse cognitive performance would reflect dedifferentiation, in which increases in functional connectivity are interpreted as decreases in neural efficiency (Ghisletta and Lindenberger, 2003; Li et al., 2001). However, higher between-network connectivity or lower within-network connectivity in older adults that relates to enhanced behavioral performance could be interpreted as a potential compensatory mechanism for weakened efficiency (Cabeza and Dennis, 2012). Furthermore, given the increased cognitive demands of engaging in a task, we expected task-based connectivity to be more sensitive to age and cognitive performance compared with resting-state connectivity (i.e., Compensation-Related Utilization of Neural Circuits Hypothesis, Reuter-Lorenz and Cappell, 2008).
Section snippets
Participants
Ninety-one adults (ages: 20–75 years, mean age = 47.4 years, sd = 17.4 years, 54 women) participated in the experiment. All participants were community-dwelling, right-handed, native English speakers who were not fluent in a second language. All participants had normal or corrected-to-normal vision and reported no history of neurological, psychological, or major medical conditions (Christensen et al., 1992).
Before the MRI session, each participant completed a battery of psychometric and
Behavioral factor analysis and effect of age
We conducted an exploratory factor analysis to assess the data for latent factors. This gave us a 4-factor model that accounted for 54% of the variance in the data (TLI: 0.99; CFI: 1.00; RMSR: 0.04; and RMSEA: 0.04) (TLI: Tucker-Lewis Index, > 0.95 is considered to be excellent; CFI: Comparative Fix Index, > 0.95 is considered to be excellent; RMSR: Root Mean Square of the Residual, < 0.06 is considered to be excellent; RMSEA: Root Mean Square Error of Approximation, < 0.06 is considered to be
Discussion
Older adults often exhibit decline across cognitive domains, such as language production, memory, and executive function; however, how these age-related behavioral differences relate to whole-brain functional organization is not entirely understood. In addition, how brain-behavior relationships can be altered by task demands is not clear either. In the present study, we analyzed 333 brain regions organized into 10 functional networks and examined within- and between-network functional
Conclusion
To summarize, the present study investigated the effect of age on cognition, whole-brain functional connectivity during resting-state and during a task, and their relationships across the adult lifespan. We found that increased age was associated with worse recall, executive function, and verbal working memory, but better language ability, although these improvements lessened as age increased. We also found consistency across both resting-state and task-based data, in that increased age was
CRediT authorship contribution statement
Haoyun Zhang: Conceptualization, Formal analysis, Project administration, Writing - original draft, Visualization. Victoria H. Gertel: Project administration, Formal analysis, Writing - review & editing. Abigail L. Cosgrove: Project administration, Formal analysis, Writing - review & editing. Michele T. Diaz: Conceptualization, Writing - review & editing, Project administration, Supervision, Funding acquisition.
Acknowledgments
This publication was supported by funding from the National Institute on Aging NIH NIA R01 AG034138 (mtd). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies. The authors have no conflicts of interest. We thank the members in the Language and Aging Laboratory at Penn State for their help with data collection, in particular Hossein Karimi and Sara Troutman. We also thank the staff and scientists at the Social,
References (71)
- et al.
Disruption of large-scale brain systems in advanced aging
Neuron
(2007) - et al.
Unified segmentation
Neuroimage
(2005) - et al.
Changes in structural and functional connectivity among resting-state networks across the human lifespan
Neuroimage
(2014) - et al.
Topological organization of the human brain functional connectome across the lifespan
Dev. Cogn. Neurosci.
(2014) - et al.
Intrinsic and task-evoked network architectures of the human brain
Neuron
(2014) - et al.
Filtering induces correlation in fMRI resting state data
Neuroimage
(2013) - et al.
Age-related sensitivity to task-related modulation of language-processing networks
Neuropsychologia
(2014) - et al.
“Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician
J. Psychiatr. Res.
(1975) - et al.
Task-based dynamic functional connectivity: recent findings and open questions
Neuroimage
(2018) - et al.
Clinical utility of the 15-item geriatric depression scale (GDS-15) for use with young and middle-aged adults
J. Affect. Disord.
(2018)