Large herbivores are among the most ecologically influential and extinction-prone animals. Megaherbivores, in particular, radically alter vegetation. Yet, few studies have tried to predict the impacts of herbivores-or their extinction-on plant species composition and community structure. First principles suggest that preferred food plants should be most strongly suppressed by herbivores and released by herbivore loss, but this intuition may be misleading if plant responses are strongly contingent on functional traits and/or plant-plant interactions. We sought to predict the responses of plant species to size-selective herbivore-exclusion treatments in an African savanna, using data on herbivore diets and plant functional traits. Our analysis had three stages. First, we identified plant traits that predicted selectivity (use relative to availability) by the dominant herbivore species excluded by each experimental treatment: megaherbivores (elephant, giraffe; ≥1000 kg), mesoherbivores (buffalo, zebra, impala; 40-600 kg), and dik-dik (5 kg). Several plant traits predicted selectivity across multiple herbivore species, but each species' diet was best predicted by a distinctive suite of traits. Second, we tested whether herbivore selectivity alone predicted plant responses. Elephant selectivity uniquely predicted plant responses in exclosures relative to unfenced control plots (R2 = 0.24-0.30); plant taxa strongly favored by elephants were ninefold more abundant inside exclosures. However, herbivore selectivity failed to predict differences in community structure between the different fenced exclusion treatments, suggesting that bottom-up effects of competition may intensify relative to consumptive effects as smaller herbivores are removed. Third, we show that including plant traits as covariates alongside elephant selectivity modestly improved predictability (R2 = 0.27-0.50). Despite various sources of uncertainty and imprecision inherent in our approach, we show that elephant foraging decisions are a major determinant of plant community dynamics. Our findings indicate that models based on readily attainable data can substantially predict plant community responses to the loss or reintroduction of megafauna. Future work can refine our approach by incorporating additional plant traits associated with tolerance and competition, along with mechanistic measurements of herbivore preferences and biomass consumption, to predict even more accurately how large-herbivore population declines and extinctions will impact plant communities.
Flavonoids, a ubiquitous class of plant polyphenolic compounds, are known for their wide spectrum of biological functions, exhibiting diverse physiological functions and possessing significant application value in pharmaceuticals, foods, and nutraceuticals. Thus, it is of great significance to conduct the toxicity assessment. However, it is impossible to perform the experimental testing for a vast number of flavonoid chemcials. In this case, in silico methods are promising to address this problem. In strict accordance with OECD principles, this study established quantitative structure-toxicity relationship (QSTR) models for predicting flavonoid acute intraperitoneal toxicity in mice by employing GA-MLR methodology. Read-Across (RA) methodology was employed to estimate the toxicity based on structural similarity. RASTR descriptors were then calculated and pooled together with QSTR descriptors to establish a q-RASTR model. Importantly, intelligent consensus modelling was implemented as another method to enhance model's stability and predictive performance. Finally, the optimal QSTR model satisfied rigorous internal and external validation benchmarks, with R2 = 0.7887, [Formula: see text] = 0.7327, [Formula: see text] = 0.8521-0.8772, [Formula: see text] = 0.9299. Based on three computational toxicology methods (QSTR, RA, and consensus modeling), the optimal model was consensus model 0 (average predictions). This model was then applied to predict the toxicity of a real external dataset lacking toxicity values. A comparative analysis with the predictions from an open-source VEGA tool was conducted to verify the applicability and predictive reliability of our model. This work offers mechanistic insights into the toxicological behavior of flavonoids and provides a rapid toxicity prediction tool for evaluating the safety of flavonoid-based chemicals.
Reliable prediction of hydrogen adsorption free energy (ΔGH) is essential for accelerating electrocatalyst discovery for the alkaline hydrogen evolution reaction (HER), yet practical machine-learning workflows remain limited by inconsistent energetic definitions, heterogeneous density functional theory (DFT) protocols, and poor out-of-distribution (OOD) generalization. Here, we present an energetically anchored machine-learning framework that integrates machine-learning interatomic potential (MLIP)-derived energetic descriptors with pretrained Crystal Hamiltonian Graph Neural Network (CHGNet) latent embeddings to predict DFT-defined hydrogen adsorption energetics across chemically diverse catalyst surfaces. The framework employs protocol-consistent single-point MLIP energy evaluation on reference geometries to construct thermodynamically aligned energetic descriptors while combining them with structural representations through gradient-boosting regression. Using curated data sets from Catalysis Hub and AQCat25 containing single-, bi-, and trimetallic adsorption systems, the embedding-augmented energetic model achieved high predictive accuracy (R2 = 0.976, MAE = 0.054 eV, RMSE = 0.105 eV) under site-level data splitting, substantially outperforming embedding-only and physicochemical-descriptor-only models. Shapley additive explanations analysis revealed a hierarchical learning mechanism in which the protocol-consistent energetic descriptor serves as the dominant thermodynamic anchor, whereas structural descriptors provide secondary refinements that improve adsorption-energy discrimination, particularly near the thermoneutral regime relevant to catalyst screening. Explicit evaluation of MLIP-only geometry workflows further demonstrated that the framework retains meaningful adsorption-energy ranking capability despite degradation introduced by MLIP-relaxed geometries. Additional OOD and transfer-learning analyses showed that predictive robustness depends strongly on the balance between compositional diversity and reference-protocol consistency. These results establish energetically anchored MLIP embeddings as an effective strategy for scalable post-DFT adsorption-energy refinement and data-efficient electrocatalyst screening while clarifying the practical limitations of MLIP-driven workflows for heterogeneous catalysis.
Stroke is a major complication of atrial fibrillation (AF), and risk prediction using the congestive heart failure, hypertension, age, diabetes, stroke, vascular disease, and sex category score (CHA₂DS₂-VASc) remains limited by residual heterogeneity. We aimed to identify plasma proteins associated with post-AF stroke and evaluate whether a protein score provides incremental predictive information beyond CHA₂DS₂-VASc. We analyzed 709 AF participants from the UK Biobank Pharma Proteomics Project, with 76 incident strokes. Stroke-related proteins were identified using multivariable Cox regression, least absolute shrinkage and selection operator (LASSO) Cox regression, and machine-learning approaches. A five-protein score was constructed, and its incremental value beyond CHA₂DS₂-VASc was assessed by discrimination, calibration, reclassification, clinical net benefit, and 1000-bootstrap internal validation. Mendelian randomization served as supportive genetic evidence. Five core proteins were identified: epidermal growth factor receptor (EGFR), V-type proton ATPase subunit D (ATP6V1D), neurotrophin 4 (NTF4), amnionless (AMN), and discoidin, CUB and LCCL domain-containing protein 2 (DCBLD2). Adding the five-protein score to CHA₂DS₂-VASc improved discrimination, increasing the concordance index from 0.681 to 0.768. The combined model had 3-, 5-, and 8-year receiver operating characteristic areas of 0.799, 0.775, and 0.806, respectively, and showed favorable five-year prediction error and calibration, with a Brier score of 0.0347, calibration intercept of -0.068, and calibration slope of 0.968. The score improved continuous net reclassification improvement (0.459; P = 0.028). Mendelian randomization provided supportive genetic evidence for AMN, EGFR, and DCBLD2. The five-protein score provided incremental predictive information beyond CHA₂DS₂-VASc for post-AF stroke risk assessment.
Prefrontal regions are implicated in explore-exploit decision-making during foraging. Older adults often show an exploitation bias, and this age period is also marked by deteriorating prefrontal myelination. To investigate whether these phenomena are linked, we examined whether lower magnetization transfer saturation (MTsat), a myelin-sensitive quantitative MRI (qMRI) measure, in these regions predicts greater exploitation bias during foraging, and whether cortical microstructure is a better predictor of bias than macrostructure (i.e., cortical thickness). Cognitively healthy older adults with familial risk of Alzheimer's disease (AD) (N=118, 60-88 years) completed a foraging task indexing explore-exploit decision-making. qMRI was used to derive MTsat values for the frontopolar cortex (FPC), medial orbitofrontal cortex (OFC), rostral middle frontal gyrus (rMFG), dorsal anterior cingulate cortex (dACC), as well as the locus coeruleus (LC), a core subcortical region strongly implicated in explore-exploit decision-making. Secondary analyses examined associations between available AD risk markers and foraging. Lower MTsat in the FPC, OFC, rMFG, and LC was associated with an exploitation bias, with LC and FPC emerging as the strongest predictors. No relationship was observed for the dACC. MTsat remained a significant predictor of foraging after controlling for cortical thickness. Observed associations were largely unrelated to AD risk markers. Individual differences in cortical microstructural integrity within a well-defined explore-exploit circuit are associated with an exploitative decision-making bias in older adults. These findings highlight the value of qMRI microstructural integrity markers, beyond standard macrostructural assays, in characterizing the neural correlates of exploitation biases in later life.
To test whether pre-operative primary CT-based 3D tumor volume (TV) improves prediction of cervical spread in oral squamous cell carcinoma (OSCC) beyond conventional size metrics. We retrospectively analyzed 132 consecutive primary OSCC patients (2014-2019). Contrast-enhanced CT datasets were semi-automatically segmented to obtain 3D TV (cm3), verified by a second observer. Multivariable logistic regression related log-TV to pathologic lymph-node metastasis (LNM), extranodal extension (ENE), and skip metastasis. Performance of radiologic TV, depth of invasion (DOI), and ellipsoid pathologic volume was compared by ROC/AUC and decision-curve analyses. Median CT-derived TV was 12.6 cm3. Log-TV independently predicted LNM (odds ratio 3.00, 95% CI 1.21-7.68; AUC 0.815), outperforming DOI (AUC 0.541) and pathologic volume (AUC 0.498). A non-linear pattern linked very small (< 1.8 cm3) and large (> 16 cm3) tumors to increased skip-metastasis probability (AUC 0.733). Radiologic and pathologic volumes showed minimal concordance (R2 = 0.0002). Decision-curve analysis demonstrated consistent net benefit of the volume model across clinically relevant threshold probabilities for elective neck dissection. CT-based 3D volumetry is an independent pre-operative predictor of cervical LNM, flags tumors prone to skip spread, and offers greater clinical utility than linear measures. Integrating radiologic volumetry into staging could refine nodal management-particularly in clinically node-negative patients-by supporting risk-adapted selection between sentinel lymph-node biopsy and elective neck dissection; prospective validation is warranted.
Permanent supportive housing (PSH) is a community-based housing model for people experiencing homelessness that provides stable housing with a range of voluntary supportive services, such as case management, healthcare, mental health, and substance use treatment services. Understanding (1) the impact of PSH on health outcomes and (2) factors that predict retention in PSH are essential for implementing support systems to address the needs of PSH residents. To address this need, 259 residents living in Santa Clara County, California were followed for two years to evaluate: (1) predictors of retention in PSH compared to transitioning to other stable non-PSH housing or returning to homelessness; (2) the longitudinal impact of PSH on mental, physical, and social health outcomes among residents remaining in PSH; and (3) moderators of these effects. Results indicate, first, that older age, being male, and lower mental health and PTSD symptoms increased the likelihood of remaining in PSH compared to moving to non-PSH stable housing. Lower mental health symptoms also predicted greater likelihood of PSH retention compared to returning to homelessness. Second, among residents who remained in PSH for 2 years, results indicated significant improvements in quality of life, health satisfaction, mental health symptoms, PTSD symptoms, substance use, and social isolation. Third, the effects of PSH on health outcomes were not significantly moderated by demographics. Study findings demonstrate the promise of PSH, in keeping people housed and significantly improving residents' health outcomes, but strategies are needed to improve mental health and substance use service delivery in PSH.
Circulating tumor DNA (ctDNA) analyses are informative as an early indicator of immunotherapy response in advanced non-small cell lung cancer (NSCLC); however, the clinical value of ctDNA molecular response requires further validation. As part of a prospective clinical protocol (NCT05995821), we conducted targeted error-correction sequencing of ctDNA (n=328) and matched WBC DNA (n=109) from 109 patients with metastatic NSCLC who received anti-PD-(L)1 either as monotherapy or in combination. Following cellular origin resolution of 2,818 variants, landmark molecular response (mR) was defined as undetectable ctDNA within 3-9 weeks of treatment initiation. Pre-treatment ctDNA burden, but not blood tumor mutation burden, predicted survival. Implementing a tumor-naïve WBC DNA-informed approach increased the number of evaluable cases without compromising the overall accuracy of landmark ctDNA molecular responses. A direct comparison of single-timepoint on-therapy ctDNA assessment with ctDNA dynamics from baseline to the 3-9-week interval, along with an analysis of heterogeneity in molecular response within the 3-9-week window, showed that undetectable ctDNA at the landmark timepoint can effectively predict survival outcomes. A significant enrichment in landmark ctDNA mR was noted among patients with progression-free survival (PFS) ≥6 months on immunotherapy (p=2.5e-05) or chemo-immunotherapy (p=0.02). Patients in the landmark mR group had longer progression-free (p=1.6e-06) and overall survival (p=2.5e-05) than those with molecular progression. Landmark ctDNA molecular response provides a real-time, accurate approach for monitoring immunotherapy clinical outcomes. Although not currently validated for regulatory use, these findings demonstrate the potential validity of ctDNA as an early endpoint of immunotherapy response.
Predictive optical coherence tomography (OCT) biomarkers for functional outcome after epiretinal membrane (ERM) surgery could aid clinicians in treatment decisions. Available studies on biomarkers for epiretinal membrane surgery remain inconclusive on their predictive value. The aim was to evaluate the association of ERM stage and disorganization of retinal inner layers (DRIL) with functional outcomes after ERM surgery. The study was conducted at the Department of Ophthalmology at the TUM University Hospital. For this retrospective cohort study, medical records were analyzed for a consecutive sample of patients diagnosed with idiopathic ERM between 01.01.2018 and 01.01.2023 and treated with ERM peeling. Clinical examination and OCT scans were evaluated prior to surgery, and at 1, 3, 6, and 12 months postoperatively. Patients with idiopathic epiretinal membrane were included if they had at least one follow-up examination between 3 and 12 months after surgery. Patients with secondary ERMs, other conditions affecting best-corrected visual acuity (BCVA), or a history of intraocular surgery (except for uncomplicated cataract surgery) were excluded. The criteria were met in 131 eyes of 126 patients, and the 12-month follow-up examination was completed in 64 eyes. Exploratory associations between functional outcomes (BCVA and metamorphopsia) 12 months after surgery and potential OCT biomarkers (ERM stage, DRIL) using regression analysis. The inclusion criteria were met for 131 eyes of 126 patients with a mean (SD) age of 70.3 (7.4) years. 62 (49.2%) patients were female, and 64 (50.8%) were male. Mean (SD) BCVA improved from 0.37 (0.21) logMAR to 0.15 (0.16) logMAR 12 months after surgery. Neither preoperative ERM stage (adjusted β=-0.003, 95% CI: -0.057-0.052, p = 0.927) nor DRIL category (adjusted β = 0.014, 95% CI: -0.056-0.084, p = 0.693) was associated with BCVA 12 months after surgery. Neither biomarker was associated with BCVA 12 months after surgery. BCVA at baseline remained the only significant parameter. Reported effect sizes can be used in future confirmatory studies.
Immune effector cell-associated neurotoxicity syndrome (ICANS) is a serious early adverse event of chimeric antigen receptor T (CAR-T) cell therapy. ICANS typically occurs concurrently with or follows cytokine release syndrome (CRS), suggesting CRS can be considered the primary risk factor for ICANS onset. Therefore, CRS characteristics in an individual patient may predict their risk for subsequently developing ICANS. We analyzed 154 patients with B cell lymphoma treated with commercial CAR-T products between 2020 and 2024; 38 patients (24.7%) developed ICANS after CRS. The cohort was split into a derivation set and a validation set. In the derivation cohort, univariate analysis identified two CRS-related factors strongly associated with ICANS: CRS onset within 24h and grade 2-4 CRS by day 3. Using these two variables, we created a simple predictive model that stratified patients into high-, intermediate-, and low-risk groups, with ICANS incidences of 47.4%, 31.0%, and 8.2%, respectively. The validation cohort confirmed this trend. These findings suggest that early CRS characteristics provide a practical, clinically applicable method for estimating ICANS risk and may support timely management decisions in CAR-T therapy.
Maeda et al. recently proposed a model to predict long-term survival after isolated surgical aortic valve replacement (SAVR) in the transcatheter aortic valve replacement (TAVR) era. While the model shows encouraging discrimination and calibration, several methodological and clinical limitations may restrict its broader applicability. The authors selected the final six-variable model primarily on the basis of maximal five-year C-statistic, without formal sample size justification or contemporary shrinkage-based criteria. Validation was restricted to internal resampling within the same registry, limiting evidence for transportability. Important prognostic domains, notably frailty and key anatomical and comorbidity variables, were not incorporated, and performance was not directly compared with established risk scores. Reporting only partially aligns with modern prediction model guidelines and omits decision curve analysis, leaving clinical utility uncertain. Overall, the model represents a valuable step but requires methodological refinement and external validation before guiding lifetime management between SAVR and TAVR.EBM Rating: Level V evidence. The article represents expert opinion derived from the author's clinical experience and interpretation of existing literature, without original experimental, randomized, controlled, cohort, or comparative analytic data.
Curved intertrochanteric varus osteotomy (CVO) is a joint-preserving procedure for osteonecrosis of the femoral head (ONFH); however, postoperative femoral head collapse remains a concern. Although the intact ratio has traditionally been assessed using plain radiography, the clinical relevance of computed tomography (CT)-based intact ratios remains unclear. This study aimed to evaluate the usefulness of the postoperative CT-based intact ratio in predicting the progression of collapse after CVO. We retrospectively reviewed data from 54 patients (58 hips) with non-traumatic ONFH who underwent CVO. Collapse progression was defined as an increase in collapse depth ≥ 3 mm compared with the immediate postoperative state. The patients were categorized into collapse and non-collapse groups. The intact ratio was measured on central coronal CT images of the femoral head. Receiver operating characteristic (ROC) curve analysis and Kaplan-Meier survival analysis were performed. With a mean follow-up of 47.4 months, collapse progression occurred in 12 hips (21.1%). The postoperative CT-based intact ratio was significantly higher in the non-collapse than in the collapse group (48.4 ± 14.3% vs. 31.8 ± 13.4%, P = 0.001). ROC analysis showed an optimal cutoff value of 39.2% (area under the curve = 0.809). Kaplan-Meier analysis revealed significantly lower three year survivorship with postoperative CT-based intact ratios < 39.2% versus ≥ 39.2% (50.6% vs. 92.1%, P < 0.001). The postoperative CT-based intact ratio was a strong predictor of femoral head collapse progression, with an optimal cutoff value of 39.2%. Applying conventional radiograph-based thresholds directly to CT evaluations may lead to an underestimation of the risk of collapse.
Limited range of motion (ROM) and arthrofibrosis are complications that affect approximately 1 to 13% of patients after primary total knee arthroplasty (TKA). Manipulation under anesthesia (MUA) is the preferred treatment when failure to achieve adequate ROM in the early postoperative period occurs. This study aims to identify predictors for MUA that may guide surgeons in preoperative risk stratification. The Premier Healthcare Database was queried to identify patients aged 18 years or older who underwent elective total knee arthroplasty. Patients who underwent manipulation under anesthesia within 90 days of index surgery were compared with patients who did not. Demographics, comorbidities, and medication usage were compared between cohorts using chi-squared and t-tests. Akaike information criterion and Bayesian information criterion minimization was done to create an optimal mixed-effects model to identify risk factors. In total, 975,235 TKAs performed between 2015 and 2021 were identified. Of these, 1.55% (15,139) of patients required MUA. Patients in the MUA group were younger (62.47 ± 9.19 vs. 67.03 ± 9.24, P < 0.001) and more likely to be Black (15.91% vs. 8.29%, P < 0.001). Mixed-effects analysis revealed a decreased risk of MUA associated with perioperative dexamethasone (adjusted odds ratio [aOR] 0.925, 95% CI, 0.889-0.962, P < 0.001), daily prednisone usage (aOR 0.433, 95% CI, 0.347-0.542, P < 0.001), and angiotensin II receptor blockers (ARBs) (aOR 0.882, 95% CI, 0.840-0.927, P < 0.001). Younger age, female sex, and Black race were associated with an increased risk of MUA after TKA, while perioperative dexamethasone, daily steroids, and ARB use were protective. These findings may serve to aid surgeons in preoperative risk stratification.
Accurate assessment of pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) remains challenging in breast cancer. This prospective single-center study compared [68Ga]Ga-FAPI-04 PET/MRI, [18F]FDG PET/CT, and contrast-enhanced MRI for predicting pathologic response after NAC. Women with biopsy-confirmed stage II-III breast cancer underwent baseline and post-therapy [68Ga]Ga-FAPI-04 PET/MRI, [18F]FDG PET/CT, and contrast-enhanced MRI before surgery. PET parameters were analyzed for primary tumors and axillary nodes. Selected variables were evaluated using LASSO and receiver-operating-characteristic analyses. Stromal FAP expression was assessed immunohistochemically in paired specimens. Twenty-four patients completed the protocol, yielding 25 primary lesions and 44 metastatic lymph nodes across 27 axillary compartments. The pCR rate was 60.00% for primary lesions and 72.73% for axillary nodes. For primary-lesion response, post-therapy [68Ga]Ga-FAPI-04 SUVmax showed the highest performance (AUC, 0.84; sensitivity, specificity, and accuracy, 80.00% each); the best [18F]FDG parameter was ΔTBR% (AUC, 0.747). For nodal response, post-therapy [68Ga]Ga-FAPI-04 SULmean showed the highest performance (AUC, 0.89; sensitivity, 91.67%; specificity, 81.25%; accuracy, 84.09%) and exceeded the best [18F]FDG parameter on DeLong testing. MRI AUCs were 0.733 and 0.770 for primary and nodal disease, respectively. Stromal FAP expression correlated positively with [68Ga]Ga-FAPI-04 SUVmax. Post-therapy [68Ga]Ga-FAPI-04 PET may provide useful adjunctive information for response assessment after NAC in breast cancer, particularly for axillary nodal evaluation, potentially supporting more individualized preoperative planning. ClinicalTrials.gov, NCT07553741, retrospectively registered.
Cultural repertoires can influence access to resources and fitness across species, impacting individuals' environmental niches and social networks. Such behaviours may induce potentially reversible changes in gene expression via epigenetic mechanisms like DNA methylation, thus capturing cultural behaviour on a molecular level. Some Indo-Pacific bottlenose dolphin (Tursiops aduncus) in Shark Bay, Western Australia, use marine sponges as foraging tools and exhibit differences in diet, sociality and survival compared to dolphins that use the same habitat but lack the tool-using know-how ('non-spongers'). We investigated the relationship between this culturally transmitted behaviour and skin DNA methylation patterns at 29 812 cytosine-phosphate-guanine (CpG) sites of the HorvathMammalMethylChip40. Machine learning models, based on differences in DNA methylation at specific CpG sites, showed moderate ability to discriminate between spongers (n = 23) and non-spongers (n = 73; area under the curve = 0.68). Permutation tests indicated that our main machine learning model performed significantly better than random at differentiating spongers from non-spongers, suggesting that DNA methylation patterns contain information associated with sponging behaviour. Our study on the nexus between epigenetics and cultural evolution opens a promising avenue of research on the molecular underpinnings of cultural practises, important to the biology of humans and other animals. This article is part of the theme issue 'Ecological epigenetics at the intersection of behaviour and life history variation in non-model animals'.
暂无摘要（点击查看详情）
暂无摘要（点击查看详情）
Developing tools for estimating heterogeneous treatment effects (HTEs) and individualized treatment effects has been an area of active research in recent years. While these tools have proven to be useful in many contexts, a concern when deploying such methods is the degree to which incorporating HTE into a prediction model provides an advantage over predictive methods that do not allow for heterogeneity in treatment effect across individuals. To address this concern, we propose a procedure that evaluates the extent to which an HTE model provides a predictive advantage. Specifically, our procedure targets the gain in predictive performance from using a flexible predictive model incorporating HTE versus an alternative partial linear model which is similar to the HTE-utilizing model except that it is constrained to not allow heterogeneity in treatment effect. By drawing upon recent work in using nested cross-validation techniques for prediction error inference, we generate confidence intervals for this measure of gain in predictive performance, which allows one to directly calculate the level at which one is confident of a substantial HTE-modeling gain in prediction-a quantity that we refer to as the predictive HTE p-value. Our procedure is generic and can be directly used to assess the benefit of modeling HTE for any method that incorporates treatment effect heterogeneity.
During the oxidation of ammonia (NH3), pyrolysis reactions strongly influence ignition delay times, intermediate radical populations, and production of nitrogen oxides (NOx), none of which are adequately predicted by NH3-containing chemical kinetic models. Moreover, simultaneous, multispecies data sets relevant to NH3 pyrolysis, useful for elucidating areas for model refinement, are generally lacking. Thus, the pyrolysis of 0.1-1.0% NH3 dilute in argon was studied behind reflected shock waves (2100-3500 K, pressures near 1 atm) using a combination of laser absorption diagnostics to measure postshock imidogen radical (NH, 336.1 nm), amino radical (NH2, 597.4 nm), and NH3 (225.3 nm) concentrations, as well as preshock NH3 concentrations (10.4 μm). In this way, the first simultaneous measurements of NH, NH2, and NH3 during NH3 pyrolysis were obtained, and refined measurements of the oscillator strength of the targeted NH2 absorption feature were enabled. Key discrepancies between these multispecies data and existing model predictions motivated the development of a refined NH3 pyrolysis model, resulting in improved predictions of measured time-histories. Predictions of late-time NH2 consumption and early-time NH formation were significantly improved. Remaining model discrepancies motivate the generation of similar multispecies data sets using hydrazines and hydrazoic acid as precursors for important NH3-relevant species.