This study used functional MRI data from the StudyForrest dataset to investigate the role of subnetworks of the default mode network (DMN) during naturalistic stimulus transitions of different types and magnitudes. We found distinct activation profiles within the DMN: the dmPFC subnetwork was specifically associated with character and location transitions, the MTL subnetwork preferred location and temporal transitions, while the Core DMN subnetwork responded to all three transition types. The multiple-demand network instead responded selectively to temporal transitions. These distinct response profiles appeared largely invariant to the semantic distance implied by the transitions. All subnetworks also responded significantly, and in a graded manner, to subjective event boundaries. Results suggest specific roles of the DMN subnetworks in perceiving and segmenting naturalistic events, supporting the view that DMN subnetworks cooperate in interpreting continuous external events and maintaining an updated contextual model of the world. ### Competing Interest Statement The authors have declared no competing interest.

This study used functional MRI data from the StudyForrest dataset to investigate the role of subnetworks of the default mode network (DMN) during naturalistic stimulus transitions of different types and magnitudes. We found distinct activation profiles within the DMN: the dmPFC subnetwork was specifically associated with social and location transitions, the MTL subnetwork preferred location and temporal transitions, while the Core DMN subnetwork responded to all three transition types. The multiple-demand network instead responded selectively to temporal transitions. These distinct response profiles appeared largely invariant to the semantic distance implied by the transitions. All subnetworks also responded significantly, and in a graded manner, to subjective event boundaries. Results suggest specific roles of the DMN subnetworks in perceiving and segmenting naturalistic events, supporting the view that DMN subnetworks cooperate in interpreting continuous external events and maintaining an updated contextual model of the world.

Task context affects stimulus representations in human visual cortex, suggesting that visual representations are flexible. However, this interpretation is at odds with a major computational goal of the human visual system: creating a perceptually stable representation of the external visual environment. How does the visual system balance stability and flexibility? Here, human participants (71 percent females) categorized object images and written words according to different task rules, while brain responses were measured with fMRI. Using an ANOVA-based modeling strategy, we precisely quantified the relative contributions of stimulus, task, and their interaction in explaining representational variance across the cortical hierarchy. Our results show that stimulus effects account for the overwhelming majority of explainable representational variance across the ventral visual system: > 95 percent in V1 and V2, and > 90 percent in higher-level visual cortex. In prefrontal cortex, the relative contributions reverse: task effects dominate stimulus effects, accounting for 80 percent of explainable representational variance. In parietal cortex, contributions of stimulus and task are approximately equal. Our findings suggest that population coding in sensory cortex is optimized for representational stability to allow a consistent interpretation of the external environment. Population coding in parietal and frontal multiple-demand cortex, by contrast, is optimized for representational flexibility to accommodate changing behavioral goals and support flexible cognition and action. Significance statement Stimulus representations in human visual cortex are affected by behavioral goals and are therefore thought to be flexible. However, this view is inconsistent with a major computational goal of the human visual system: creating a perceptually stable representation of the external environment. Here, we show that modulatory effects of behavioral goals on stimulus representations in visual cortex are surprisingly small. In contrast, behavioral goals strongly affect representations in parietal and frontal multiple-demand cortex. Our findings suggest that population coding in sensory cortex is optimized for stable perception, while population coding in parietal and frontal multiple-demand cortex is optimized for flexible cognition. ### Competing Interest Statement The authors have declared no competing interest. Medical Research Council, https://ror.org/03x94j517, SUAG/002/RG91365 British Academy, https://ror.org/0302b4677, PS140117 Dutch Research Council, https://ror.org/04jsz6e67, Rubicon Postdoctoral Fellowship

Recent findings challenge traditional views of the Default Mode Network (DMN) as purely task-negative or self-oriented, showing increased DMN activity during demanding switches between externally-focused tasks (Crittenden et al., 2015;Smith et al., 2018;A. X. Zhou et al., 2024). However, it is unclear what modulates the DMN at switches, with transitions within a stimulus domain activating DMN regions in some studies but not others. Differences in the number of tasks suggest that complexity or structure of the set of tasks may be important. In this fMRI study, we examined whether the DMN's response to task switches depended on the number of tasks that could be encountered in a run, or on abstract task groupings defined by the temporal order in which they were learnt at instruction. Core DMN activation at task switches was unaffected by the number of currently relevant tasks. Instead, it depended on the order in which groups of tasks had been learnt. Multivariate decoding revealed that Core DMN hierarchically represented individual tasks, task domains, and higher-order task groupings based on instruction order. We suggest that, as the complexity of instructions increases, rules are increasingly organised into higher-level chunks, and Core DMN activity is the highest at switches between chunks.

Functional compensation is a common notion in the neuroscience of healthy ageing, whereby older adults are proposed to recruit additional brain activity to compensate for reduced cognitive function. However, whether this additional brain activity in older participants actually helps their cognitive performance remains debated. We examined brain activity and cognitive performance in a human lifespan sample ( N = 223) while they performed a problem-solving task (based on Cattell’s test of fluid intelligence) during functional magnetic resonance imaging. Whole-brain univariate analysis revealed that activity in bilateral cuneal cortex for hard vs. easy problems increased both with age and with performance, even when adjusting for an estimate of age-related differences in cerebrovascular reactivity. Multivariate Bayesian decoding further demonstrated that age increased the likelihood that activation patterns in this cuneal region provided non-redundant information about the two task conditions, beyond that of the multiple demand network generally activated in this task. This constitutes some of the strongest evidence yet for functional compensation in healthy ageing, at least in this brain region during visual problem-solving.

The Default Mode Network (DMN), traditionally associated with internally oriented cognition, has been increasingly implicated in processing external stimuli that signal the need to update an internal mental model, defined as neural representations that simulate and predict relationships between elements that make up events. This review synthesizes recent findings on the DMN’s role in diverse cognitive transitions, including feature changes and event boundaries during movie watching, and externally cued task switches during complex behavioural paradigms. Studies using naturalistic stimuli reveal the functional specialization of DMN subnetworks: the medial temporal lobe subnetwork responds to location and temporal transitions, the dorsomedial prefrontal cortex subnetwork prefers social transitions, and the Core DMN exhibits generalised responses across all these transition types. Meanwhile, in task-based paradigms, the Core DMN is also activated during external task switches, compared to task repeats, both representing and dependent upon a hierarchical task structure, further underscoring DMN involvement in maintaining and updating mental models. We propose that the Core DMN binds information across DMN subnetworks into a hierarchically chunked internal model, where elements are organized into nested levels of abstraction, to provide context for evolving naturalistic experiences and a substrate to guide behaviour. This unifying framework highlights the DMN's integral role in context-dependent cognitive transitions, dynamically integrating internal and external representations during both active and passive states.

Default mode network (DMN) activity, measured with fMRI, typically increases during internally directed thought, and decreases during tasks that demand externally focused attention. However, Crittenden et al. (2015) and Smith et al. (2018) reported increased DMN activity during demanding external task switches between different cognitive domains, compared to within-domain switches and task repeats. This finding is hard to reconcile with many dominant views of DMN function. Here, we aimed to replicate this DMN task-switch effect in a similar paradigm and test whether it reflects increased representation of broader context, specifically of a scene presented behind the focal task. In the Core DMN, we found significant activity for all task switches, compared to task repeats, and stronger activity for switches between rest and task. Although the content of the background scene was attended, recalled, and neurally decodable, there was no evidence that this differed by switch type. Therefore, external task switches activated DMN without enhanced processing of the surrounding background. Surprisingly, DMN activity at within-domain switches was no less than at between-domain switches. We suggest that modulation of DMN activity by task switches reflects a shift in the current cognitive model and depends on the overall complexity of that model.

Cognitive task performance can be supported through multiple neural pathways, a concept referred to as brain degeneracy. We used a novel approach to consider brain degeneracy during a visual short-term memory task (VSTM) across the adult lifespan in the Cam-CAN study (n = 113, 23-87 years old). Our main goal was to identify subgroups of participants whose VSTM performance was characterized by distinct brain activation patterns. First, we identified seven brain modules that responded similarly to the VSTM task and resembled previously identified functional networks (adjusted mutual information [aMI] = 0.45). Subsequently, latent profile analysis revealed four distinct subgroups of participants. Each subgroup was characterized by different recruitment patterns of these brain modules, predominantly in the frontal control module (FCM), visual module (VM), and default mode module (DMM). Subgroups did not differ in demographics or task performance. However, associations between brain activity and performance varied across subgroups, particularly in the FCM, suggesting that individuals may use different cognitive operations to perform the VSTM task. Further analyses revealed group differences in white matter integrity, mostly in the uncinate fasciculus, suggested that individual differences in structural brain properties may shape the different brain activation patterns. Altogether, our study contributes to our understanding of how multiple neural pathways could underlie cognitive performance. ### Competing Interest Statement The authors have declared no competing interest.

Working memory is critical to higher-order executive processes and declines throughout the adult lifespan. However, our understanding of the neural mechanisms underlying this decline is limited. Recent work suggests that functional connectivity between frontal control and posterior visual regions may be critical, but examinations of age differences therein have been limited to a small set of brain regions and extreme group designs (i.e., comparing young and older adults). In this study, we build on previous research by using a lifespan cohort and a whole-brain approach to investigate working memory load-modulated functional connectivity in relation to age and performance. The article reports on analysis of the Cambridge center for Ageing and Neuroscience (Cam-CAN) data. Participants from a population-based lifespan cohort (N = 101, age 23–86) performed a visual short-term memory task during functional magnetic resonance imaging. Visual short-term memory was measured with a delayed recall task for visual motion with three different loads. Whole-brain load-modulated functional connectivity was estimated using psychophysiological interactions in a hundred regions of interest, sorted into seven networks (Schaefer et al., 2018, Yeo et al., 2011). Results showed that load-modulated functional connectivity was strongest within the dorsal attention and visual networks during encoding and maintenance. With increasing age, load-modulated functional connectivity strength decreased throughout the cortex. Whole-brain analyses for the relation between connectivity and behavior were non-significant. Our results give additional support to the sensory recruitment model of working memory. We also demonstrate the widespread negative impact of age on the modulation of functional connectivity by working memory load. Older adults might already be close to ceiling in terms of their neural resources at the lowest load and therefore less able to further increase connectivity with increasing task demands.

The ability to solve novel complex problems predicts success in a wide range of areas. Recent research suggests that the ability to cognitively segment complex problems into smaller parts constrains nonverbal reasoning in adults. This study aimed to test whether cognitively segmenting problems improves nonverbal reasoning performance for children as it does for adults. 80 children aged 7-10 years completed two versions of a modified traditional matrix reasoning task in which demands on working memory, integration, and processing speed were minimised, such that the only significant requirement was to break each problem into its constituent parts. In one version of the task, participants were presented with a traditional 2x2 matrix and asked to draw the missing matrix item into a response box below. In a second version, the problem was broken down into its component features across three separate cells, reducing the need for participants to segment the problem. As with adults, performance was better in the condition in which the problems were separated into component parts. Children with lower fluid intelligence did not benefit more in the separated condition than children with higher fluid intelligence, and there was no evidence that segmenting problems was more beneficial for younger than older children. This study demonstrates that cognitive segmentation is a critical component of complex problem-solving for children, as it is for adults. By forcing children to focus their attention on separate parts of a complex visual problem, their performance can be dramatically improved.

Fluid intelligence, the ability to solve novel, complex problems, declines steeply during healthy human aging. Using fMRI, fluid intelligence has been repeatedly associated with activation of a frontoparietal brain network, and impairment following focal damage to these regions suggests that fluid intelligence depends on their integrity. It is therefore possible that age-related functional differences in frontoparietal activity contribute to the reduction in fluid intelligence. This paper reports on analysis of the Cambridge Center for Ageing and Neuroscience data, a large, population-based cohort of healthy males and females across the adult lifespan. The data support a model in which age-related differences in fluid intelligence are partially mediated by the responsiveness of frontoparietal regions to novel problem-solving. We first replicate a prior finding of such mediation using an independent sample. We then precisely localize the mediating brain regions, and show that mediation is specifically associated with voxels most activated by cognitive demand, but not with voxels suppressed by cognitive demand. We quantify the robustness of this result to potential unmodeled confounders, and estimate the causal direction of the effects. Finally, exploratory analyses suggest that neural mediation of age-related differences in fluid intelligence is moderated by the variety of regular physical activities, more reliably than by their frequency or duration. An additional moderating role of the variety of nonphysical activities emerged when controlling for head motion. A better understanding of the mechanisms that link healthy aging with lower fluid intelligence may suggest strategies for mitigating such decline.SIGNIFICANCE STATEMENTGlobal populations are living longer, driving urgency to understand age-related cognitive declines. Fluid intelligence is of prime importance because it reflects performance across many domains, and declines especially steeply during healthy aging. Despite consensus that fluid intelligence is associated with particular frontoparietal brain regions, little research has investigated suggestions that under-responsiveness of these regions mediates age-related decline. We replicate a recent demonstration of such mediation, showing specific association with brain regions most activated by cognitive demand, and robustness to moderate confounding by unmodeled variables. By showing that this mediation model is moderated by the variety of regular physical activities, more reliably than by their frequency or duration, we identify a potential modifiable lifestyle factor that may help promote successful aging.

Listening to spoken language engages domain-general multiple demand (MD; frontoparietal) regions of the human brain, in addition to domain-selective (frontotemporal) language regions, particularly when comprehension is challenging. However, there is limited evidence that the MD network makes a functional contribution to core aspects of understanding language. In a behavioural study of volunteers (n = 19) with chronic brain lesions, but without aphasia, we assessed the causal role of these networks in perceiving, comprehending, and adapting to spoken sentences made more challenging by acoustic-degradation or lexico-semantic ambiguity. We measured perception of and adaptation to acoustically degraded (noise-vocoded) sentences with a word report task before and after training. Participants with greater damage to MD but not language regions required more vocoder channels to achieve 50% word report, indicating impaired perception. Perception improved following training, reflecting adaptation to acoustic degradation, but adaptation was unrelated to lesion location or extent. Comprehension of spoken sentences with semantically ambiguous words was measured with a sentence coherence judgement task. Accuracy was high and unaffected by lesion location or extent. Adaptation to semantic ambiguity was measured in a subsequent word association task, which showed that availability of lower-frequency meanings of ambiguous words increased following their comprehension (word-meaning priming). Word-meaning priming was reduced for participants with greater damage to language but not MD regions. Language and MD networks make dissociable contributions to challenging speech comprehension: Using recent experience to update word meaning preferences depends on language-selective regions, whereas the domain-general MD network plays a causal role in reporting words from degraded speech.

Fluid intelligence encompasses a wide range of abilities such as working memory, problem-solving, and relational reasoning. In the human brain, these abilities are associated with the Multiple Demand Network, traditionally thought to involve combined activity of specific regions predominantly in the prefrontal and parietal cortices. However, the structural basis of the interactions between areas in the Multiple Demand Network, as well as their evolutionary basis among primates, remains largely unexplored. Here, we exploit diffusion MRI to elucidate the major white matter pathways connecting areas of the human core and extended Multiple Demand Network. We then investigate whether similar pathways can be identified in the putative homologous areas of the Multiple Demand Network in the macaque monkey. Finally, we contrast human and monkey networks using a recently proposed approach to compare different species' brains within a common organizational space. Our results indicate that the core Multiple Demand Network relies mostly on dorsal longitudinal connections and, although present in the macaque, these connections are more pronounced in the human brain. The extended Multiple Demand Network relies on distinct pathways and communicates with the core Multiple Demand Network through connections that also appear enhanced in the human compared with the macaque.

Ventrolateral prefrontal cortex (vlPFC), dorsolateral prefrontal cortex (dlPFC), and temporal cortex (TE) all contribute to visual decision-making. Accumulating evidence suggests that vlPFC may play a central role in multiple cognitive operations, perhaps resembling domain-general regions of the human frontal lobe. We trained monkeys in a task calling for learning, retrieval, and spatial selection of rewarded target objects. Recordings of neural activity covered large areas of vlPFC, dlPFC, and TE. Results suggested a central role for vlPFC in each cognitive operation with strong coding of each task feature, while only location was strongly coded in dlPFC and current object identity in TE. During target selection, target location was communicated first from vlPFC to dlPFC, followed by extensive mutual support. In vlPFC, stimulus identities were independently coded in different task operations. The results suggest a central role for the inferior frontal convexity in controlling successive operations of a complex, multi-step task.

While much variance in general intelligence or g is genetic, a substantial environmental component suggests a possible role for parent-child interaction. In particular, previous evidence suggests the importance of parental scaffolding, or provision of cognitive structure to shape child behaviour. A role for scaffolding is consistent with the proposal that, in adult cognition, a critical aspect of g is decomposition of complex problems into a structure of simpler parts. Building on previous work, we recruited 162 parents attending Children's Centres with a child aged 2–4 years, and examined parental scaffolding during a book-sharing activity. Scaffolding was measured as the first principal component of a variety of parental behaviours, including sensitivity, focusing attention, extending comprehension, and promoting child participation. Child g was measured as the first principal component of a broad cognitive battery, including language, attention, working memory, and executive function. Importantly, we assessed contributions of the parent's own intelligence, education, and family income. Though these variables were all associated with both child g and parental scaffolding, scaffolding remained predictive of child g even once the influence of these variables was removed. In contrast to the correlation with cognitive proficiency, scaffolding did not predict child pro-social behaviour. We suggest that parental scaffolding supports the child's development of a broad skill of attentional structuring, promoting the across-the-board cognitive proficiency that is reflected in g.

Chronic and recurrent forms of clinical depression can persist for a lifetime and often respond poorly to intervention. Psychological formulations implicate rigid, negative expectations of self and world which are resistant to updating with new information, a phenomenology consistent with a Bayesian account of brain function. Bayesian predictive processing models suggest that sensory data which is represented with low precision (high uncertainty) in the brain cannot exert much influence on existing beliefs, giving rise to the hypothesis that persistent forms of depression may be characterised by disturbances in sensory precision optimization. We optimized a computational model with data from a cross-modal (visual, auditory, somatic) covert attention task to estimate sensory precision in persistently depressed participants relative to healthy controls. Results suggested that both sensory precision and the salience of attentional targets were attenuated in depressed participants across sensory modalities, contributing to a suppression of contextual prediction error in this group. These outcomes provide support for a novel theoretical account of depression chronicity and suggest avenues for enhancing the effectiveness of psychological interventions for this population.

Anhedonia – a pervasive loss of interest or enjoyment in relation to previously valued stimuli - is a central feature of clinical depression. A series of studies investigated alterations in perceptual sensitivity and attention, and their relation to anhedonia, in participants with persistent depressive illness in comparison to healthy controls. Study 1 measured perceptual sensitivity in the somatic and auditory modalities while minimising attentional influences. Study 2 maximised attentional influences to assess their impact on perceptual sensitivity in the same modalities. Perceptual sensitivity did not differ between groups in either study. Both studies found evidence of lateralised attentional deficits suggestive of subtle left-sided neglect in the depressed samples. In both studies, lateralised attentional deficits were associated with anhedonia above and beyond depression severity. Study 3 demonstrated through simulation that a single lateralised attentional lapse model could predict the key outcomes in both empirical studies, and additionally demonstrated the plausibility of a resource conservation strategy as a data-generating process among depressed participants. In conclusion, the studies support the novel insight that anhedonia may be related to disruptions in brain networks subserving stimulus-driven attention, with a potential contribution of resource conservation strategies that reduce access to precise sensory information.

Disrupted bodily awareness may reflect mechanisms which drive both symptomatology and disorder maintenance in clinical depression. We investigated attentional capture by somatic signals in clinically depressed individuals and healthy controls as one aspect of bodily awareness. Attentional performance in a cross-modal covert orienting task was globally disrupted when depressed participants attempted to ignore uninformative somatic cues, in a largely modality-specific manner. In a subsequent study employing a similar paradigm, attention was disproportionately captured in depressed participants by informative somatic cues, in a similarly modality-specific way. Taken together, the results suggest that the salience of somatic signals is amplified in clinical depression, and that attempts to ignore them may have a disruptive effect on attentional processing more generally in this population.

Classical executive tasks, such as Wisconsin card-sorting and verbal fluency, are widely used as tests of frontal lobe control functions. Since the pioneering work of Shallice and Burgess (1991), it has been known that complex, naturalistic tasks can capture deficits that are missed in these classical tests. Matching this finding, deficits in several classical tasks are predicted by loss of fluid intelligence, linked to damage in a specific cortical “multiple-demand” (MD) network, while deficits in a more naturalistic task are not. To expand on these previous results, we examined the effect of focal brain lesions on three new tests – a modification of the previously-used Hotel task, a new test of task switching after extended delays, and a test of decision-making in imagined real-life scenarios. As potential predictors of impairment we measured volume of damage to a priori MD and default mode (DMN) networks, as well as cortical damage outside these networks. Deficits in the three new tasks were substantial, but were not explained by loss of fluid intelligence, or by volume of damage to either MD or DMN networks. Instead, deficits were associated with diverse lesions, and not strongly correlated with one another. The results confirm that naturalistic tasks capture cognitive deficits beyond those measured by fluid intelligence. We suggest, however, that these deficits may not arise from specific control operations required by complex behaviour. Instead, like everyday activities, complex tasks combine a rich variety of interacting cognitive components, bringing many opportunities for processing to be disturbed.

The default mode network (DMN) is often associated with representing semantic, social, and situational content of contexts and episodes. The DMN may therefore be important for contextual decision-making, through representing situational constraints and simulating common courses of events. Most decision-making paradigms, however, use symbolic stimuli and instead implicate cognitive control regions, such as the multiple demand (MD) system. This fMRI study aimed to contrast the brain mechanisms underlying decision-making based on rich naturalistic contexts or symbolic cues. While performing an ongoing task, 40 human participants (25 female) responded to different sounds. For one sound, the stimulus-response mapping was fixed; responses for the other sounds depended on the visual context: either lifelike scenes or letter symbols, varying across participants. Despite minimal behavioral differences between the groups, posterior DMN regions showed increased activity during context-dependent decision-making using the naturalistic scenes only, compared with symbolic cues. More anterior temporal and frontal DMN regions showed a different pattern, with sensitivity to the need for contextual control, but not to the type of context. Furthermore, in the scenes group, widespread DMN regions showed stronger representation of not just the context but also the sound whose significance it modulated. In comparison, the MD system showed strong univariate activity for every decision, but, intriguingly, somewhat reduced activity in the case of a scene-based but demanding context-dependent decision. Depending on context, we suggest, either DMN or MD regions may play a prominent role in selection and control of appropriate behavior.SIGNIFICANCE STATEMENT Contextual knowledge is widely believed to be important for guiding real-world goal-directed behavior. Much remains to be understood, however, regarding the underlying brain mechanisms. Using a novel paradigm to contrast decisions based on richly meaningful naturalistic scenes with decisions based on symbolic cues, we find that both multiple demand regions and default mode regions may contribute to the cognitive control of behavior. Rich semantic context enhances representation not just of the context itself, but also of the contents of the decision that it controls. Dependence of a decision on naturalistic context can also reverse the common pattern of multiple demand regions responding more, and default mode regions responding less, to more difficult decisions.