Skip to main content

Aging-related changes in brain activation and deactivation during cognition: novel insights into the physiology of the human mind from simultaneous PET-fMRI imaging

Periodic Reporting for period 3 - SIMULTAN (Aging-related changes in brain activation and deactivation during cognition: novel insights into the physiology of the human mind from simultaneous PET-fMRI imaging)

Reporting period: 2020-06-01 to 2020-12-31

"Functional magnetic resonance imaging (fMRI) is an imaging method that is sensitive to transient increases in neural activity during cognitive operations. There is no doubt that fMRI has led to a breakthrough in our ability to measure how the complexities of the mind are rooted in biology. However, fMRI measurements are only an indirect measure of neural activity and reflect a complex interplay between hemodynamic and metabolic demands. Indeed, there are highly reliable patterns of fMRI activation that have had a large impact on the study of the human mind but that cannot be readily linked to concomitant changes in neural activity because invasive animal studies are not applicable. The SIMULTAN project pioneers a new imaging technique in healthy humans to address the neural basis of two such patterns of fMRI activity. The first work package investigates the neural basis of fMRI “deactivations”, i.e. fMRI responses that decrease during cognitive control relative to a passive resting state and asks whether deactivations are reflective of decreases in neural, or synaptic, activity. Future work in SIMULTAN will explore the intriguing observation that older adults (65+) often display higher levels of activation of the prefrontal cortex during performance of a cognitive task when compared to young adults. The main question here is whether over-activation of the frontal cortex in older adults can be explained by increased neural activity that implies a compensatory response or whether age-related over-activations are merely reflective of cardiovascular changes in aging.
The results impact a basic understanding of neurocognitive architecture. Knowing whether deactivations reflect an active ""suppression"" of task-irrelevant areas would suggest a structure in which the human mind is divided into separable cognitive systems supporting externally and internally directed thought processes that are not independent of each other. The results would further guide the development of clinical interventions. Knowing whether over-activations reflect neural activations of additional brain areas or whether they may simply reflect altered blood flow properties in an aging brain could decide whether it is worthwhile to pursue interventions that target the vascular system or interventions that target a specific neurotransmitter system for example. Finally, it is hoped that the methodological advances made by SIMULTAN can be adapted by other researchers to evaluate the degree to which fmri reflects the underlying biology and further the development of fmri as a biomarker of human brain function."
Over the first year and a half, imaging data were collected for 30 healthy young adults and 40 healthy older adults. The project uses a novel imaging technique that combines traditional fMRI with a simultaneous measurement of glucose consumption via a PET scan. In SIMULTAN, we assess transient changes in glucose consumption as a surrogate marker of neural activity that is independent of vascular coupling (blood flow). Our assumption that transcends through the project is that signal changes in fMRI are reflective of neural activity if they are paralleled in space and time by transient changes in glucose metabolism. If they are not, they are likely reflective of cardiovascular events only (i.e. blood flow).
For the first work package, the SIMULTAN team analysed data from the 30 young adults. Confirming our assumption and providing proof of concept for the novel hybrid imaging technique, we find good correspondence between fMRI signal increases and increased glucose metabolism in frontal-parietal areas that are active during cognitive control versus rest (Figure 1). Intriguingly, we find that the task-induced fMRI deactivations (shown in blue in figure 2), which have to date been difficult to fully understand, are not paralleled by overlapping task-induced changes in glucose metabolism. This shows that task-induced deactivations are not antagonistic to fMRI activations, an important discovery on its own. Most interestingly, we also discovered a set of regions that are metabolically active during task (versus rest) but show no spatially congruent fMRI signal change during the task. These regions are adjacent to task-induced deactivations, i.e. at the borders of the functional activations and deactivations. Further inspection of these areas shows that they form a functional network that aligns with activations at the beginning of a task but with the nearby deactivations later on in the task. We think that this points toward a “switch” area, situated along a cortical functional gradient, that plays a role in suppressing activity of areas involved in the passive state (mind wandering) during cognitive control. The results were presented in oral presentations at smaller meetings and two different international conferences and are currently being prepared for journal publication.
In SIMULTAN, a novel cutting-edge human imaging technique is used that integrates a fully functional positron emission tomography (PET) system inside a traditional MRI magnet. This technique allows imaging of molecular targets and processes, including measures of metabolic activity, concurrent with any type of MRI imaging sequence. This technique has almost exclusively been used in cancer research to gain complementary information from a tumor location; in SIMULTAN we successfully adapted the methods to assess and compare change in fMRI signal and a change in glucose consumption (as a measure of neuronal activity) within an individual as he or she transitions from one cognitive state to another.

State of the art research on fMRI deactivations of the so called “default mode” network (i.e. the areas active during passive rest and deactivated during cognitive tasks) has begun to realize that the default network is not a homogene network but instead is comprised of several “interwoven” networks and that cortical networks are better understood in terms of a gradient-like organization from sensory-motor cortex to attention and higher order association networks with the default mode network at its apex (Buckner 2019 Nature Neuroscience Reviews,for review). Our results from period 1 align well with these emerging ideas and go beyond the state of the art by providing empirical evidence for a new understanding of default mode deactivations during cognitive control. Specifically, we discovered a set of regions that are positioned at the borders of the traditional boundaries of the default mode network but that are functionally unique in that they flexibly align with either of the adjacent networks depending on task demand. We believe that these areas are implicated in “switching” between different cognitive sets (externally and internally directed modes of thought) by suppressing the brain activity present in passive resting states (i.e. internal mode) during cognitive control. Importantly, we do not find evidence that neural activity in the networks implicated in externally directed thought appears suppressed by the same mechanism during passive rest. The latter finding challenges views of the default mode network as having an antagonistic relationship with the network implicated in externally directed or “active” cognition. In further work, we will use the same novel imaging technique established in work package 1 to delineate the neural basis of over-activations in aging.
SIMULTAN results 1
SIMULTAN results 2