The default mode network’s metabolic specialization

Core regions of the default mode network rank among the highest in aerobic glycolysis across the entire human brain (Vaishnavi et al., Reference Vaishnavi, Vlassenko, Rundle, Snyder, Mintun and Raichle2010). Despite producing less ATP per glucose than oxidative phosphorylation, aerobic glycolysis affords rapid and flexible energy availability (Magistretti & Allaman, Reference Magistretti and Allaman2015). Within the efficiency taxonomy that Haueis and Colaço outline (Section 2.1.2), this metabolic profile reflects a trade-off: the network sacrifices thermodynamic efficiency for temporal responsiveness. Notably, the same regions also show neotenous gene expressions linked to synaptic growth and plasticity (Goyal et al., Reference Goyal, Hawrylycz, Miller, Snyder and Raichle2014), implying a state kept metabolically pliable well beyond typical developmental windows. This pliability is well-suited to the dynamic recombination of information that characterizes spontaneous thought: autobiographical memory, future thinking, social cognition, and creative thought (Christoff et al., Reference Christoff, Irving, Fox, Spreng and Andrews-Hanna2016).

Evaluative application: Pruning models of mind wandering

The executive failure hypothesis (Kane & McVay, Reference Kane and McVay2012; McVay & Kane, Reference McVay and Kane2010), for instance, proposes that mind wandering (i.e., a member of a family of spontaneous-thought phenomena; Christoff et al., Reference Christoff, Irving, Fox, Spreng and Andrews-Hanna2016) arises primarily from failures of top-down control. Under this account, the energetic cost of thought content plays little explicit role. Yet empirical evidence suggests that patterns of default mode network activity during working‑memory maintenance track the phenomenological richness of ongoing experience (Sormaz et al., Reference Sormaz, Murphy, Wang, Hymers, Karapanagiotidis, Poerio, Margulies, Jefferies and Smallwood2018). In other words, this network continues to support spontaneous thought even when attentional resources are committed elsewhere. This finding is hard to square with an account that explains mind wandering solely in terms of failures of attentional control.

Metabolic data help explain this pattern. Using simultaneous PET-fMRI, Stiernman et al. (Reference Stiernman, Grill, Hahn, Rischka, Lanzenberger, Panes Lundmark, Riklund, Axelsson and Rieckmann2021) demonstrated that glucose metabolism in the default mode network remains elevated even when blood‑oxygen‑level-dependent signals decrease – a dissociation that conventional neuroimaging alone would miss. Godbersen et al. (Reference Godbersen, Klug, Wadsak, Pichler, Raitanen, Rieckmann, Stiernman, Cocchi, Breakspear, Hacker, Lanzenberger and Hahn2023) revealed further that the default mode network’s metabolic activity varies depending on which task-positive networks are active. Collectively, these results indicate that what determines the character of spontaneous thought is competition for bioenergetic resources across networks, not just the strength of executive suppression. In the terms of Haueis and Colaço, metabolism here functions as an explanatory constraint: it structures which cognitive forms can be sustained under a given task load, without itself initiating those processes. The executive failure account, which ignores this dimension, is thereby prunable.

Generative application: Predicting phenomenology

The default mode network’s metabolic profile also generates specific, testable predictions. First, as external task demands increase, spontaneous thought should become progressively less vivid, shifting from richly detailed, multisensory simulations toward more schematic mentation. This should correlate with reductions in aerobic glycolysis within core default mode network regions specifically, over and above changes in frontoparietal activation (Godbersen et al., Reference Godbersen, Klug, Wadsak, Pichler, Raitanen, Rieckmann, Stiernman, Cocchi, Breakspear, Hacker, Lanzenberger and Hahn2023; Sormaz et al., Reference Sormaz, Murphy, Wang, Hymers, Karapanagiotidis, Poerio, Margulies, Jefferies and Smallwood2018). Second, individuals with higher resting aerobic glycolysis in the default mode network should sustain more elaborate spontaneous thought under high cognitive load. More broadly, stable differences between people in the richness, sensory specificity, and temporal extension of their spontaneous thought should partly reflect metabolic architecture rather than attentional control alone. Recent work linking resting‑state thought contents to resting-state functional connectivity (Yu, Wong, & Billaud, Reference Yu, Wong and Billaud2025) offers early support for this point, suggesting that thought patterns may partly mirror underlying metabolic organization. Taken together, these predictions illustrate how metabolic knowledge, joined with network-level neuroscience, can populate the model space with novel, testable hypotheses.