Neural Efficiency as a Brain-Wide Signature of Pediatric Anxiety and a Biomarker for Treatment Response

Pediatric anxiety disorders are the most prevalent psychiatric conditions in youth, affecting nearly 11% of children and adolescents ages 3–17 years in the U.S (1). Further, these rates have been increasing over the past decade, with a national survey reporting an alarming 60% increase in the prevalence of anxiety in adolescents ages 12–17, from 10% in 2016 to 16% in 2023 (2). The median age of onset of pediatric anxiety is 6 years; if untreated, pediatric anxiety increases children’s risk for developing secondary mood disorders, with accompanying difficulties in social, academic, and emotional functioning (3,4). Cognitive behavioral therapy (CBT) is currently the gold-standard treatment for pediatric anxiety; nevertheless, as Linke et al. (2025) note, nearly half of youth with pediatric anxiety do not experience clinically meaningful improvement with CBT. Given that a longer duration of anxiety is associated with higher symptom severity and poorer post-treatment functioning (5), identifying biomarkers for treatment response is critical for the development of more effective approaches to early intervention in this disorder.

Over the past several years, investigators have attempted to identify neural metrics both that differentiate youth with anxiety disorders from healthy controls and that predict response to treatment for pediatric anxiety. With respect to the former goal, researchers have documented consistently that pediatric anxiety is associated with alterations in cortical and subcortical brain regions that are involved in emotion regulation, threat processing, reward processing, and attentional control, including the amygdala, insula, prefrontal cortex, and anterior cingulate cortex (4). In contrast, findings concerning neural markers of response to CBT have been more equivocal; while some studies have identified specific patterns of task- or rest-based neural activation or functional connectivity that significantly predict response to CBT (6), results of investigations in this area are often inconsistent and lack sufficient sensitivity to predict individuals’ treatment response (7,8). A major challenge to identifying reliable neural predictors of response to treatment for pediatric anxiety is the fact that the brain regions that are affected by this disorder and that are altered by CBT are distributed across brain-wide networks that interact to support complex cognitive and affective functions. Importantly, in a noteworthy advance in our efforts to identify biomarkers of treatment response in pediatric anxiety, Linke et al. (9) leveraged signals from the entire brain during both task and rest conditions to yield a promising metric of neural efficiency.

Neural efficiency is a key concept in cognitive neuroscience, originally proposed in a positron emission tomography study in which Haier and colleagues (10) found that higher intelligence was associated with reduced brain glucose metabolism during problem solving, which they posited indicated a more economical use of the brain’s energy. This neural efficiency hypothesis has been supported by studies using functional magnetic resonance imaging (fMRI), in which individuals with higher intelligence have been found to have lower levels of brain activation during completion of novel cognitive tasks. In subsequent studies, researchers documented decreases in brain activation with extensive practice or long-term expertise, collectively suggesting that the brain optimizes efficiency by minimizing energy consumption (11,12). More recently, researchers have operationalized neural efficiency as the degree to which individuals reconfigure intrinsic functional brain organization, typically assessed during rest, to meet specific task demands (13). Using this conceptualization of neural efficiency, investigators are finding that greater similarity in patterns of functional network connectivity between rest and task conditions (i.e., less brain reconfiguration) is associated not only with higher intelligence (13), but also with better task-switching (14) and faster response to threat stimuli (15). Given these data linking neural efficiency to cognitive and affective processes, it is reasonable to hypothesize that neural efficiency influences individuals’ response to treatment of forms of psychopathology that are characterized by difficulties in cognitive and affective functioning.

In this context, Linke et al. (9) examined the utility of neural efficiency during threat processing as a biomarker of both a diagnosis of pediatric anxiety and response to CBT in 103 adolescents with a diagnosis of anxiety and matched healthy controls. Administering an emotional dot-probe task in the scanner, Linke et al. found that anxious youth had significantly lower neural efficiency than did healthy controls at baseline, indicating that they required a greater degree of reconfiguration between threat-processing and resting states. Notably, this attenuated neural efficiency was associated with higher clinician- and parent-rated, but not self-reported, symptoms of anxiety. Further, in 80 anxious youth who completed CBT, Linke et al. found that adolescents with higher neural efficiency at baseline showed a steeper decrease in clinician-rated symptoms of anxiety. The association between neural efficiency at baseline and percent change in symptom reduction was moderate (r=−0.29), which corresponds to a clinically meaningful effect size (Cohen’s d=−0.61). Finally, in a smaller subsample of anxious youth (n=49) and healthy controls (n=40) with usable second fMRI scans, Linke et al. found that neural efficiency in anxious youth did not change with CBT, remaining significantly lower than that of healthy controls throughout the study.

Linke et al.’s study represents a major advance in our understanding of the brain basis of pediatric anxiety and response to CBT. We think that three findings in particular warrant consideration. First, Linke et al.’s finding that neural efficiency did not improve with CBT, apparently even with symptomatic improvement, combined with its moderate test-retest reliability when computed using partial correlations, suggests that neural efficiency is a trait-like construct. Therefore, neural efficiency may be a reliable biomarker for patient stratification, helping clinicians personalize interventions for pediatric anxiety. Given the medium-to-large effect size of using neural efficiency to predict treatment outcome, assessing this construct prior to treatment could differentiate anxious youth who are most likely to benefit from CBT from adolescents who may respond better to alternative interventions, thereby contributing to the development of a precision medicine approach to the treatment of pediatric anxiety. Thus, as Linke et al. noted, a critical direction for future research is to explore whether neuromodulation techniques like transcranial magnetic stimulation can increase neural efficiency, which may improve outcomes in treatment-resistant individuals.

Second, to determine whether neural efficiency was driven by specific functional connections between brain regions during rest or task, Linke et al. quantified the contribution of each pairwise connection to their calculation of neural efficiency. Interestingly, they found that only 5% of all functional connections contributed meaningfully to neural efficiency; further, none of these individual connections survived a stringent statistical threshold. This is an important and somewhat counterintuitive finding, suggesting that neural efficiency in threat processing is not attributable to any single critical pathway, but rather, is a distributed, network-wide phenomenon. This finding challenges the traditional region- or pathway-centric view of pediatric anxiety (4), highlighting instead the intricate and interconnected nature of brain organization during threat processing. Future studies of neural markers of pediatric anxiety and its treatment might extend Linke et al.’s finding by explicitly and systematically assessing global patterns of connectivity that underpin complex cognitive and affective functioning.

Finally, Linke et al. found that different aspects of neural efficiency were related to clinician- and parent-rated levels of adolescent anxiety, but not to the participants’ ratings of their own anxiety symptoms. The authors note that these discrepancies may be due to lower reliability of adolescents’ than of parents’ and clinicians’ ratings, or to adolescents’ potentially biased sensitivity when evaluating their own behavior. Attempting to understand informant discrepancies in symptom reporting has a long history in clinical science. Whereas some investigators have found that children with a clinically diagnosed anxiety disorder under-report symptoms relative to their parents, other researchers have reported the opposite pattern of results (16). It is intriguing that Linke et al. did not find neural efficiency to be related to the adolescents’ own ratings of their anxiety, which not only suggests that adolescents have limited insight about their symptoms but also documents a disconnect between individuals’ brain function and their perceptions of their behavior. Thus, neural efficiency appears to be a more reliable and sensitive indicator of impaired threat processing than it is of adolescents’ self-reported symptoms of anxiety. Adopting objective and mechanistically driven neural metrics should increase the precision of clinical assessment and facilitate the development of more effective intervention strategies for pediatric anxiety.

In conclusion, Linke et al.’s study is an important advance in identifying a reliable biomarker of the diagnosis of pediatric anxiety and a clinically significant predictor of response to CBT for this disorder. Given that researchers have found functional brain alterations to precede the onset of pediatric anxiety (17,18), it is important to extend Linke et al.’s study to investigate whether neural efficiency is associated with adolescents’ risk for developing an anxiety disorder. Investigators should also conduct longitudinal studies to investigate whether neural efficiency can be improved through psychological, pharmacological, or neuromodulatory interventions that, in turn, might reduce adolescents’ risk for the onset of anxiety, and whether the clinical utility of neural efficiency generalizes to other developmental periods and other disorders.