{
  "artiushin-2026-spider-atlas": {
    "subject": "A three-dimensional immunofluorescence atlas of the synganglion of the hackled-orb weaver spider Uloborus diversus, built from whole-mount synapsin staining and registered to a common reference volume. The atlas maps the distribution of classical neurotransmitters (GABA, acetylcholine, serotonin, octopamine/tyramine) and neuropeptides (allatostatin A, CCAP, FMRFamide, proctolin) across sub- and supraesophageal neuropils in a web-building species.",
    "claims": "The work resolves transmitter architecture across leg, opisthosomal, pedipalpal, and cheliceral neuropils, describes layered organization of the arcuate body into four sublayers with differential transmitter content, and documents two previously uncharacterized protocerebral structures: the tonsillar neuropil (serotonin core, TDC2 shell) and a candidate protocerebral bridge with stratified transmitter identity. Mushroom bodies are identified as AstA-exclusive, with cholinergic and GABAergic globuli cells.",
    "inferences": "Together the tonsillar neuropil and candidate protocerebral bridge are proposed as components of a spider equivalent of the insect central complex, reframing prior cursorial-spider neuroanatomy by situating orb-weaver brain organization within a broader arthropod central-complex lineage and nominating specific structures as substrates for web-building behavior."
  },
  "bouyeure-2026-fear-rsa": {
    "hypotheses": "The paper tests whether the fear network encodes threatening cues through a generalized between-item representation whenever threat is acquired, whether reversal recruits two simultaneous representational strategies — generalization for currently dangerous cues and item-specific stabilization for cues whose valence changes — in dissociable regions, and whether prefrontal context-specific coding is the substrate linking context distinctiveness to context-dependent fear renewal.",
    "claims": "RSA on fMRI data across acquisition, reversal, and test phases establishes CS+ > CS- cue generalization in dACC, SFG, caudate, and insula during acquisition; during reversal, newly dangerous CS-+ cues acquire a generalized fear-network pattern while changing-contingency cues exhibit elevated item-specific stability in precuneus and IFG. Context representations become more distinct in dmPFC and lateral PFC during reversal, and PFC context specificity predicts subsequent reinstatement of acquisition fear traces at test.",
    "inferences": "The brain adapts to shifting contingencies through a flexible composition of generalization and item-specific coding rather than a single representational scheme, and prefrontal context coding is implicated as a candidate neural mechanism of fear renewal — a framing with direct implications for why exposure therapy gains do not always transfer across contexts."
  },
  "ejdrup-2026-dopamine": {
    "hypotheses": "The paper tests whether regional differences in striatal dopamine dynamics arise from release-related phenomena or from differential dopamine-transporter activity, and whether the classical phasic/tonic dichotomy corresponds to distinct receptor-level readouts by D1 and D2 receptors. It further asks whether DAT nanoclustering is sufficient to generate the observed dorsoventral gradient.",
    "claims": "A large-scale three-dimensional simulation of extracellular dopamine reproduces isolated hotspots without tonic baseline in dorsal striatum and a pervasive tonic-like signal in ventral striatum at 4 Hz pacemaker activity; parameter sweeps identify Vmax as the sole parameter generating the regional difference, consistent with immunostaining showing elevated DAT in DS and dSTORM showing greater DAT nanoclustering in VS. D1 receptor occupancy tracks phasic dopamine with ~50 ms delay, while D2 receptor occupancy integrates over seconds and is insensitive to brief firing pauses.",
    "inferences": "The prevailing view that phasic and tonic dopamine represent distinct release modes is refuted: both emerge from a single release process filtered through region-specific DAT activity, with nanoclustering proposed as the upstream regulator. D2 receptors cannot resolve individual bursts or pauses on behavioral timescales, constraining theories that assign D2 signaling a role in rapid event-level computation."
  },
  "gadeke-2026-guilt-insula": {
    "hypotheses": "The paper bets that responsibility for a choice leaving a partner worse off produces interpersonal guilt — a sharper fall in the chooser's momentary happiness — carried by the anterior insula, that a separate substrate represents the partner's reward surprises more strongly when the participant's own choice caused them, and that prefrontal connectivity with both regions shifts with who decides. Agency aversion, disappointment at one's own outcome and a social shift in risk attitude have to go with it.",
    "claims": "Across a behavioural study and an fMRI replication, happiness fell further after a partner's low outcome when the participant rather than the partner had chosen, whatever the participant's own outcome — ruling out self-directed disappointment — while a decision-maker cost independent of outcome ruled out agency aversion and unchanged risk premiums ruled out a shift in risk attitude. A happiness model weighting the partner's reward surprises from the participant's own choices fitted best, though the participant's own surprises weighed more. Left anterior insula and left superior temporal sulcus clusters carried the responsibility-contingent signals, and insula–prefrontal connectivity varied with condition and choice.",
    "inferences": "The convergence of behavioural, computational and neural evidence is read as localising interpersonal guilt to the anterior insula, with the superior temporal sulcus tracking a partner's outcomes gated by the participant's agency — social choice under risk as a responsibility-weighted affective computation rather than a reward comparison. A previously published neural guilt signature was recovered across the group but did not track individual differences in behavioural guilt. All of it rests on two samples of healthy adults, every brain result from the scanned study, and a partner whose choices were simulated.",
    "model": "supplied:runs/gadeke-2026-guilt-insula/summaries.answer.v4.json"
  },
  "headley-2026-inhibitory-rhythms": {
    "hypotheses": "The paper tests whether rhythmic inhibition onto distinct compartments of a layer 5 pyramidal neuron regulates integration in a compartment-specific and frequency-specific manner — specifically, whether perisomatic inhibition is optimally tuned to gamma while distal dendritic inhibition is optimally tuned to beta.",
    "claims": "In a biophysically realistic compartmental model supporting Na+, NMDA, and Ca2+ dendritic spikes, doubling distal inhibition reduces somatic firing from 5.5 Hz to 0.2 Hz by suppressing dendritic spikes, while doubling perisomatic inhibition reduces firing to 0.7 Hz by raising the AP voltage threshold. Frequency sweeps establish that beta (~20 Hz) is optimal for entraining dendritic spike onsets at distal sites and gates clustered apical input, whereas gamma (40–80 Hz) is optimal for phase-modulating perisomatic AP threshold and gates proximal and basal input, with the two effects remaining functionally orthogonal.",
    "inferences": "The compartmental division of labor provides a mechanistic grounding for the empirical association of parvalbumin-positive interneurons with gamma and somatostatin-positive interneurons with beta: each interneuron class targets the compartment whose integrative role aligns with its preferred rhythm. Rhythm and spatial target are therefore not independent features of inhibition but jointly specify what inhibition does to the postsynaptic computation."
  },
  "kammer-2026-foveal-feedback": {
    "hypotheses": "The paper tests whether primary visual cortex receives image-specific feedback about peripheral saccade targets before those targets reach the fovea, whether that feedback carries shape rather than semantic category information, and whether it shares a representational format with direct foveal stimulation.",
    "claims": "In a gaze-contingent fMRI paradigm in which peripheral targets disappear before fixation, multivariate classifiers decode target identity from foveal retinotopic V1 despite the absence of foveal stimulation; decoding generalizes across experimental and control conditions, follows a U-shaped eccentricity profile inconsistent with spillover or large receptive fields, and preserves shape but not category information. V1 loses category decoding in the feedback condition while lateral occipital cortex shows the inverse pattern, and trial-wise foveal decodability covaries with activity in the intraparietal sulcus.",
    "inferences": "Foveal V1 predicts the features of incoming saccade targets via top-down feedback in a shared representational format with bottom-up stimulation, positioning IPS as a candidate driver and nominating foveal prediction as a mechanism underlying transsaccadic perceptual stability. The shape-not-category profile constrains the feedback to low-to-mid-level visual content rather than abstract object identity."
  },
  "kolb-2026-igabasnfr2": {
    "hypotheses": "The paper tests the engineering bet that near-saturation mutagenesis at the Pf622 binding pocket and cpGFP linkers of iGABASnFR can yield a successor sensor with improvements large enough to cross qualitative capability thresholds for single-bouton, single-trial, and in vivo GABA imaging.",
    "claims": "Screening 3,947 variants in primary cultured neurons identifies iGABASnFR2 with 4.1-fold higher ΔF/F, 30% faster rise time, and synapse-appropriate affinity, together with a negative-going variant iGABASnFR2n; a 2.60 Å crystal structure (PDB 9D57) shows the cpGFP remains rigid on GABA binding. Side-by-side comparisons with the v1 sensor demonstrate single-bouton hippocampal detection, single-trial direction-selective GABA release in retina, and volume-transmitted GABA release in somatosensory cortex following whisker stimulation.",
    "inferences": "Both engineering hypotheses are confirmed: the screen yields multiple improved variants and the improvements are sufficient to unlock measurements that v1 cannot support, with the first in vivo demonstration of direction-selective retinal GABA release establishing the capability threshold has been crossed. The negative-going sensor offers independent corroboration of positive-going signals, addressing a recurrent concern about intensiometric sensor artifacts."
  },
  "meijer-2025-serotonin-orthogonal": {
    "hypotheses": "The paper tests whether phasic dorsal-raphe serotonin release drives a rapid offline-to-online state switch during quiet wakefulness, whether 5-HT codes surprise and thereby alters perceptual decisions in a steering-wheel task, and whether a strong neuromodulatory signal can leave behavior intact by occupying a subspace orthogonal to the choice-coding dimensions of the population.",
    "claims": "Across 7,478 neurons recorded with Neuropixels in 13 regions of 17 mice, optogenetic 5-HT stimulation dilates the pupil, suppresses hippocampal sharp wave ripples, and increases whisking and sniffing during quiet wakefulness while modulating 10–60% of neurons per region bidirectionally. The same stimulation produces no detectable change in psychometric slope, bias, lapse, reaction time, prior updating, or probabilistic choice-model weights, and the 5-HT effect axis in PCA space is significantly more orthogonal to the choice axis than shuffle null distributions, with orthogonality growing toward the choice moment.",
    "inferences": "The surprise-coding hypothesis is refuted by the absence of behavioral effects despite robust neural modulation; the orthogonal-subspace hypothesis is advanced as the reconciliation, proposing that neuromodulation can reshape population activity without perturbing ongoing computation when it targets dimensions the downstream readout ignores. This offers a mechanism by which 5-HT could gate state-dependent stimulus representations without interfering with in-progress decisions."
  },
  "meijer-2025-serotonin-additive-r1": {
    "hypotheses": "The paper tests whether phasic dorsal-raphe serotonin modulates brain-wide spiking additively rather than through multiplicative gain, whether the resulting modulation occupies a subspace geometrically orthogonal to choice-coding dimensions, and whether phasic 5-HT release drives a rapid offline-to-online internal-state switch during quiet wakefulness. A carried-over hypothesis that 5-HT codes surprise is also tested behaviorally.",
    "claims": "Optogenetic activation of DRN serotonergic neurons modulates 10-60% of neurons across 13 regions in 7,478 cells yet leaves perceptual decision behavior intact, while inducing pupil dilation, hippocampal sharp-wave-ripple suppression, and exploratory whisking. A per-neuron GLM yields significant main effects of choice and stimulation but a near-zero choice-by-stimulation interaction (-0.05 +/- 0.06), and a receptor-expression GLM shows that 5-HT1a and 5-HT2 receptors contribute oppositely signed effects that together explain roughly half the variance in regional modulation.",
    "inferences": "The near-zero interaction term eliminates multiplicative gain control as the brain-wide mode of 5-HT action, overturning the default computational framework for neuromodulators since Servan-Schreiber (1990) and demoting orthogonality to a geometric consequence of additivity rather than an independent finding. The receptor-expression results support a reconciliation in which a 5-HT1a-mediated baseline suppression dominates over the 5-HT2A-mediated evoked-response gain reported in visual cortex, while the absence of behavioral effects refutes the surprise hypothesis."
  },
  "rozak-2026-neurovascular-dl": {
    "hypotheses": "The paper advances two coupled bets: an engineering hypothesis that a deep-learning pipeline combining segmentation, longitudinal registration, radius estimation, and graph-theoretic analysis can recover network-level neurovascular measurements unavailable to conventional single-vessel approaches, and a biological hypothesis that optogenetic neuronal activation drives coordinated, graph-level microvascular responses that point-caliber measurement cannot resolve.",
    "claims": "The NOVAS3D UNet/UNETR ensemble outperforms ilastik on Dice, precision, and HD95; registration nearly doubles the tracked vessel count (241 to 412 per field) and radius estimation recovers simulated ground truth at R-squared = 0.68. Applied to Thy1-ChR2 cortex, the pipeline reveals baseline within-vessel radius heterogeneity of 24 +/- 28%, dilations occurring closer to labeled neurons than constrictions (16.1 vs 21.9 micrometers), dilators biased toward the surface, a 152 +/- 65% rise in network assortativity, and a 4% median gain in capillary-network efficiency under 4.3 mW/mm-squared photostimulation, all absent in green-light and wild-type controls.",
    "inferences": "The within-vessel heterogeneity and network coordination findings jointly rule out the assumption that better single-vessel segmentation alone can substitute for graph-level analysis of functional hyperemia, reframing neurovascular coupling as an intrinsically network-level phenomenon. Scope is bounded by the single-preparation training set and the untested 2-SD responder threshold, which condition all quantitative biology claims."
  },
  "scheller-2026-self-prioritization": {
    "hypotheses": "The paper tests whether arbitrary self-association acts as an automatic attentional salience signal at early perceptual stages, whether self-prioritization mobilizes absolute TVA processing capacity rather than merely redistributing fixed attentional weights, and whether social and perceptual salience capture attention through largely independent mechanisms that would produce additive effects when both are applied to the same stimulus.",
    "claims": "In a temporal-order-judgment task with hierarchical Bayesian TVA modeling across two experiments (N = 69, N = 71), self-associated stimuli are processed about 1.5 Hz faster than other-associated under perceptual decisions but the advantage vanishes or reverses under explicit social decoding, and a condition-specific capacity model decisively outperforms a weights-only model (Delta-loo = 14.2, weight 0.86). For other-associated stimuli the social-by-perceptual interaction is near zero (-0.54 Hz) consistent with additivity, while for self-associated stimuli the perceptual-salience benefit is markedly reduced (2.5 Hz vs 5.2 Hz for other, 6 Hz neutral).",
    "inferences": "Self-prioritization operates automatically at early, pre-attentive stages and reduces rather than amplifies under active social decoding, and it does so by recruiting additional processing capacity rather than merely reweighting attention. The sub-additive interaction in the self condition partially refutes the strong form of the independence hypothesis: social and perceptual salience stream independently except when self-relevance obligatorily dominates, forcing a qualified rather than unrestricted independence account."
  },
  "wengert-2026-kcnc1": {
    "hypotheses": "The paper tests three linked hypotheses about KCNC1 developmental and epileptic encephalopathy: that the recurrent A421V missense variant produces Kv3.1 loss of function through impaired membrane trafficking and reduced potassium current, that this loss selectively compromises parvalbumin-positive fast-spiking interneurons while sparing excitatory neurons, and that the resulting PV-interneuron dysfunction is sufficient to drive network hyperexcitability, seizures, cognitive impairment, and premature lethality.",
    "claims": "In a conditional Kcnc1-A421V/+ knock-in mouse, cortical PV interneurons show reduced Kv3.1 surface expression, reduced voltage-gated potassium current density, impaired maximal firing, and altered action-potential downstroke and half-width, while excitatory neurons are unaffected at both juvenile and adult ages. Inhibitory synaptic transmission is intact in juveniles but altered in young adults, and mutants exhibit hypersynchronous cortical discharges correlated with myoclonic seizures on EEG, spatial-learning and working-memory deficits, spontaneous seizures with SUDEP-like events, and death before 122 days.",
    "inferences": "The findings establish A421V as a loss-of-function variant that produces cell-type-selective excitability failure in PV interneurons, with excitatory neurons serving as a dissociating control that localizes the primary lesion to Kv3.1-dependent fast-spiking circuits. The age-dependent emergence of synaptic failure against a backdrop of early intrinsic PV dysfunction supports a progressive circuit model in which PV-interneuron compromise is causally upstream of the encephalopathy phenotype, providing a mechanistic foundation for the broader KCNC1 disease spectrum."
  }
}