Randall W. Engle

Individual differences in cognitive abilities may influence how people develop and maintain trustin automated systems, yet the mechanisms underlying this relationship remain poorlyunderstood. This study investigated how three foundational cognitive abilities — workingmemory (WM), attention control (AC), and fluid intelligence (Gf) — shape trust dynamics,particularly focusing on trust decline and recovery after automation failures. Participants (n=111)completed cognitive tests measuring WM, AC, and Gf before performing a dual-task paradigmwhere they monitored an automated memory aid exhibiting varying reliability levels (100%-65%-100%) while completing a mental arithmetic task. Results revealed that WM and Gfpositively predicted trust recovery after automation failures, while AC influenced long-term trustcalibration, with high-AC individuals showing more nuanced trust adjustments reflecting systemreliability history. Trust decline was uniform across ability levels, suggesting immediacy of trustloss. The relationship between abilities and trust may have been partially mediated byautomation schemas, with WM negatively correlating with all-or-none thinking aboutautomation. These findings suggest different cognitive abilities play distinct roles in trustdynamics: WM supports monitoring automation performance, Gf enables effective schemaupdating, and AC facilitates appropriate deployment of these resources for trust calibration. Thiswork advances our understanding of trust mechanisms and has implications for designingautomation interfaces that accommodate individual cognitive differences.

Efficient hypothesis sampling requires avoiding resampling: returning to hypotheses that have already been rejected. Doing so could depend on maintaining information about which hypotheses have been ruled out, or on disengaging from previously activated hypotheses once they are no longer relevant. Participants (N = 302) completed a hypothe- sis sampling task in which they tested hypotheses from six alternatives until identifying the correct one, along with batteries measuring working memory capacity (WMC), fluid intelligence (Gf), and attention control (AC). In latent- variable models, higher WMC, Gf, and AC were each associated with less resampling when considered separately, but only Gf uniquely predicted resampling once their overlapping contributions were partialled out. Mediation anal- yses showed that AC accounted for the WMC association but not the Gf association. Separating AC into selective and sustained components, selective but not sustained AC predicted resampling, although not uniquely after ac- counting for Gf and WMC, and neither component mediated the Gf effect. Overall, this pattern is consistent with the idea that avoiding rejected hypotheses depends more on disengaging from outdated representations than on maintaining information in working memory. The unique Gf association beyond AC suggests that this contributionis not fully reducible to domain-general attention control.

In hypothesis sampling, individuals progressively constrain a space of candidate explanations, eliminating al-ternatives as evidence accumulates. Central to this process is the capacity to avoid returning to hypotheses already rejected, yet the cognitive mechanisms supporting that avoidance remain unresolved. We tested whether efficient sampling depends on working memory capacity (WMC), associated with the active maintenance of tested and avoided hypotheses, or fluid intelligence (Gf), associated with disengagement from them. A large community sample completed a hypothesis sampling task alongside latent-variable assessments of WMC, Gf, and attention control (AC), with AC further decomposed into selective and sustained components. Structural equation models showed that Gf uniquely predicted resampling behavior, fully subsuming the effect of WMC. AC did not mediate the relationship between Gf and resampling; in a reverse mediation, AC's association with resampling was itself accounted for by Gf. Within AC, only selective attention predicted resampling, while sustained attention showed no association. These findings indicate that efficient hypothesis sampling depends on attentional disengagement, the capacity to suppress previously activated representations and reallocate attention toward new candidates, rather than on active maintenance. The residual con-tribution of Gf beyond attention control further suggests that additional processes, plausibly involving relational integration, support the structured reconfiguration of the hypothesis space.Keywords: hypothesis sampling; fluid intelligence; working memory capacity; attention control; disengagement; selective attention; relational integration

Attention control predicts academic achievement, professional success, and health outcomes. However, the neural basis of stable, individual differences in attention control remains unclear. Prior research has emphasized momentary fluctuations in attentional engagement, often overlooking enduring individual differences. Here, we applied the quasi-periodic pattern analysis of infraslow functional magnetic resonance imaging (fMRI) dynamics in a large sample ( N = 196) to test whether trait attention control is reflected in network-level brain activity as well as the locus coeruleus (LC). Using latent-variable measures of attention control, working memory capacity, and fluid intelligence, we isolated the unique contribution of attention control across rest, 1-back, and 3-back conditions. As cognitive demand increased, individuals with higher attention control exhibited more coordinated activity of the frontoparietal control network (FPCN): they showed enhanced coupling with the dorsal attention network (DAN), and greater engagement with the LC and stronger decoupling from the default mode network (DMN). Even at rest, high attention individuals demonstrated stronger FPCN–DAN coupling and little to no correlation between FPCN–DMN, indicating that attentional capacity is reflected in both task-evoked reconfiguration and baseline network architecture. These findings reveal how attention control, as an ability, is instantiated in the brain’s dynamic architecture.

Working memory capacity (WMC) has long served as a central indicator of individual differences in complex cognition. However, growing evidence suggests that a substantial portion of its predictive power may reflect attention control (AC)—including goal maintenance, interference management, and inhibition—rather than storage capacity alone. This review synthesizes findings across six domains: (1) perception and sensory discrimination, (2) learning and problem solving, (3) cognitive control and decision making, (4) retrieval and memory performance, (5) multitasking and real-world performance, and (6) clinical applications. Across these areas, WMC-related effects frequently align with demands on AC, though the strength and nature of this alignment vary by domain. We highlight the importance of incorporating reliable AC measures and recommend latent-variable approaches to more clearly separate storage, control, and representational processes underlying complex performance.

Most research on student AI use asks who adopts these tools, treating AI use as a stable personal characteristic. Yet each time students encounter a problem, they face a fresh decision about whether to seek AI help. We examined when students choose AI assistance by analyzing both stable individual differences and moment-to-moment decisions. Undergraduate students (N = 398) completed 30 quiz trials, freely choosing on each trial to answer independently, consult an AI chatbot, or rely entirely on AI. Before each decision, students rated their topic expertise. We also measured cognitive abilities, AI attitudes, and demographics. Multilevel modeling revealed that almost half the variance in AI use reflected stable individual differences; the other half varied within individuals across trials. Self-assessed expertise accounted for this within-person variation: students sought AI help primarily when they felt unknowledgeable about a topic. At the between-person level, fluid intelligence predicted more chatbot consultation but less full delegation, suggesting higher-ability students use AI strategically rather than as a substitute for thinking. Male students and those with favorable AI attitudes showed greater willingness to delegate entirely. These findings reframe AI use as adaptive help-seeking rather than a fixed trait. Educational interventions that target metacognitive calibration, helping students accurately judge when they need assistance, may prove more effective than blanket AI policies that assume uniform use patterns.

Attention control predicts academic achievement, professional success, and health outcomes. However, the neural basis of stable, trait-level differences in attention control remains unclear. Prior research has emphasized momentary fluctuations in attentional engagement, often overlooking enduring individual differences. Here, we applied the quasi-periodic pattern (QPP) analysis of infraslow fMRI dynamics in a large sample (N = 191) to test whether trait attention control is reflected in network-level brain activity as well as the locus coeruleus. Using latent-variable measures of attention control, working memory capacity, and fluid intelligence, we isolated the unique contribution of attention control across rest, 1-back, and 3-back conditions. Under heightened cognitive demand, individuals with higher attention control exhibited more flexible coordination of the frontoparietal control network (FPCN): they showed stronger decoupling from the default mode network (DMN), enhanced coupling with the dorsal attention network (DAN), and greater engagement with the locus coeruleus (LC). Even at rest, high attention individuals demonstrated stronger FPCN-DAN coupling, and little to no correlation between FPCN-DMN, indicating that attentional capacity is reflected in both task-evoked reconfiguration and baseline network architecture. These findings reveal how attention control, as a trait, is instantiated in the brain’s dynamic architecture.

Mental speed theories of intelligence suggest that people are smarter because they are faster. Weargue that attention control plays an important and fundamental role in mediating the relationshipbetween basic sensory processes and more complex cognitive processes such as fluidintelligence. One of the most successful paradigms for establishing a mental speed theory ofintelligence is the inspection time task. In this paper, we examine the mental speed and theattention control perspectives on the inspection time task and its relationship with fluidintelligence. Integrating experimental and correlational approaches, we find that attention controlexplains the inspection time task’s correlation with fluid intelligence and working memorycapacity. Attention control and inspection time are correlated beyond their relationship withother measures of processing speed. Further, while we find no evidence that selective attentionspecifically is related to inspection time performance, both attention control and inspection timepredicted declines in accuracy as participants sustained their attention over time; other measuresof processing speed did not predict sustained attention performance. Collectively, these resultsindicate that inspection time is related to the ability to control attention, especially the ability tosustain attention over time.

Military selection tests leave room for improvement when predicting work‐relevant outcomes. We tested whether measures of attention control, working memory capacity, and fluid intelligence improved the prediction of training success above and beyond composite scores used by the U.S. Military. For student air traffic controllers, commonality analyses revealed that attention control explained 9.1% ( R = .30) of the unique variance in academic performance, whereas the Armed Forces Qualification Test explained 5.2% ( r = .23) of the unique variance. For student naval aviators, incremental validity estimates were small and nonsignificant. For student naval flight officers, commonality analyses revealed that attention control measures explained 11.8% ( R = .34) of the unique variance in aviation preflight indoctrination training performance and 4.3% ( R = .21) of the unique variance in flight performance. Although these point estimates are based on relatively small samples, they provide preliminary evidence that attention control measures might improve training outcome classification accuracy in real‐world samples of military personnel.