Uncertainty is the enemy of selection — and 'brains in sync' does not mean 'hearts in sync'
The mechanistic centrepiece this week is a Nature Neuroscience paper showing that when you are unsure which task you are doing, the brain fails to suppress irrelevant features, neural representations of different features become entangled, and errors rise — a neuronal account of why divided or uncertain attention is costly. Alongside it, a quasi-experimental Nature Human Behaviour cohort of 5,227 Italian students ties early social-media adoption to standardized-test losses of up to a quarter of a standard deviation, mediated by smartphone intrusion into key moments of the day. Two findings push back on comfortable assumptions: longitudinal hyperscanning shows brain-to-brain synchrony can fall as people grow closer, and a validity-controlled clinical sample finds no additive cognitive deficit when adult ADHD co-occurs with anxiety or depression. Every item is peer-reviewed this week; three are preclinical or mechanistic, and several are single-site, so sizing is noted item by item.
grade-10 mathematics score for students who opened a social-media account in grade 6 versus matched late adopters (95% CI −0.39 to −0.15); n=5,227, difference-in-differences (Nature Human Behaviour, 2026)
Where the human evidence is
Count by statusEvidence-stage map
Stage × promise (1–10)Attention domain impact
Strength 0–10Why not knowing which task you are doing makes you worse at both
MechanismXue et al. (Nature Neuroscience, Sep 2026). When the relevant feature is uncertain, the brain stops suppressing the irrelevant one; the two representations blur together and decisions degrade. The practical corollary is to resolve the task before attending to the stimulus.
Switching between the outer and inner world — a framework, and two circuit dissociations
The Attentional Switching Framework: one shared system decides when to switch between external and internal attention, separate systems hold each state
Kam (mind-wandering), Spreng (default network), Mills and Cleary propose a dual-process account of how attention moves between the external world and internal thought. A 'resource-dependent' regulatory mechanism, drawing on a common set of frontoparietal regions and pathways, determines when and how the switch occurs regardless of direction; 'resource-independent' domain-specific representations, in largely dissociable networks, maintain each state once entered. The framework is intended to reconcile conflicting findings on whether external-to-internal and internal-to-external switches share machinery. It is a synthesis, not new data, and its predictions await preregistered tests; it is included because it gives the mind-wandering literature a testable structure that maps onto the default-mode versus dorsal-attention dichotomy in the book.
Read the paperToo little and too much medial prefrontal activity impair different stages of behavioural flexibility in rats
Renström and colleagues (Nottingham, with Boehringer Ingelheim) tested two schizophrenia-linked prefrontal abnormalities in rats performing operant reversal learning. Functional inhibition of medial prefrontal cortex ('hypofrontality') impaired early reversals, when adapting to a rule change was still new; GABAergic disinhibition (excess activity) impaired later, well-practised serial reversals. The stage dependence matters for the attention literature because 'prefrontal engagement' is often treated as monotonic — here both directions of dysregulation hurt, at different points in learning. Rodent work with pharmacological manipulations; translation to human cognitive flexibility is inferential.
Read the paperAnterior cingulate excitatory neurons handle post-error adjustment while inhibitory neurons restrain impulsivity — a cell-type dissociation within 'cognitive control'
Tan and colleagues combined the five-choice serial reaction time task with c-Fos mapping, fibre photometry and bidirectional chemogenetics in anterior cingulate cortex. Under high attentional demand the ACC shifted from broad to selective recruitment. CaMKIIα excitatory neurons were more engaged during decision-making and mediated post-error slowing; GAD1 inhibitory neurons surged after premature responses and, when manipulated, altered impulsivity. The result splits 'cognitive control' into an error-monitoring channel and an impulse-restraint channel with different cell types, relevant to ADHD models where both are implicated. Mouse study, single lab; the journal is mid-tier.
Read the paperClinical and lifespan — what co-occurs with ADHD, and what ageing does not take
When symptom and performance validity are controlled, adults with ADHD plus anxiety or depression show no additive neurocognitive deficit (n=752)
Harrison, Davin and Kwan analysed 752 postsecondary students referred for psychoeducational assessment, restricted to those passing established validity criteria, and compared ADHD-only, ADHD with elevated anxiety/depression, anxiety-only, depression-only and assessment-seeking controls across intellectual, executive, memory and achievement measures. Group means fell in the average range with heavy overlap; adults with ADHD and internalizing symptoms did not perform worse than ADHD alone, significant differences were small and inconsistent, and rates of scores at or below the 16th percentile were comparable. The message is that diagnostic comorbidity should not be taken as evidence of objective cognitive impairment. Referred, help-seeking sample; cross-sectional; single centre.
Read the paperADHD is associated with ~48% higher odds of gastrointestinal symptoms across 23 studies and 1.9 million people
Blake, Cannon and Shantikumar (Warwick) pooled studies with DSM/ICD-defined ADHD and reported functional GI symptoms (organic GI disease excluded). ADHD was associated with increased odds of every symptom examined: any GI symptom OR 1.48 (95% CI 1.35–1.62), IBS 1.58 (1.39–1.78), constipation 1.97 (1.49–2.61) and encopresis 4.32 (2.50–7.47). Candidate mechanisms — microbiome differences, vagal dysregulation, irregular eating, stimulant side effects — are not adjudicated, and the authors explicitly cannot separate intrinsic biology from medication effects. Observational and heterogeneous, but large; clinically, it argues for asking about gut symptoms in ADHD care.
Read the paperPreregistered null: mild-to-moderate age-related hearing loss does not degrade memory-guided auditory attention in adults aged 40–80
Pec, Alain and Paul tested the preregistered hypothesis that greater hearing loss would weaken the ability to use long-term memory to guide auditory attention — a hippocampus-dependent function thought vulnerable in ageing. Sixty-two adults learned associations between background scenes and the location of a probe-tone target. Contrary to prediction, memory-guided attention was preserved and uncorrelated with hearing thresholds; only verbal identification of the background sounds was worse in those with more hearing loss. A clean disconfirmation of the 'cognitive load from hearing loss erodes attention' account, at least for this task and severity range. Modest n and a single task, so it constrains rather than refutes the sensory-deprivation hypothesis.
Read the paperDecide what the task is before you start attending to it: ambiguity about which feature or goal matters produces measurable interference from the irrelevant one (Xue et al., Nature Neuroscience). Set a single explicit objective per work block rather than 'seeing what comes up'.
For children, delaying a first social-media account matters more than policing daily minutes: the Italian cohort shows the earliest adopters lost the most, and the mediating pathway is phone use before sleep, on waking and during homework — the specific windows to protect (Gui et al., Nature Human Behaviour).
Do not use brain-to-brain synchrony, or its consumer proxies, as a scorecard for connection; in the PLOS Biology cohort it fell as intergenerational pairs became closer and more predictable to each other.
A pre-task walk is reasonable but should not be sold as an attention enhancer: in the crossover fMRI study it altered neural recruitment without improving switch performance, while practice on the task itself produced the real gains (Asahara et al., JMRI).
In adults with ADHD, co-occurring anxiety or depression should not be read as extra cognitive impairment without validity-controlled testing; conversely, functional gut symptoms are common enough (OR ~1.5) to ask about routinely (Harrison et al., Arch Clin Neuropsychol; Blake et al., J Atten Disord).