Humans can quickly recognize objects in a dynamically changing world. This ability is showcased by the fact that observers succeed at recognizing objects in rapidly changing image sequences, at up to ...13 ms/image. To date, the mechanisms that govern dynamic object recognition remain poorly understood. Here, we developed deep learning models for dynamic recognition and compared different computational mechanisms, contrasting feedforward and recurrent, single-image and sequential processing as well as different forms of adaptation. We found that only models that integrate images sequentially via lateral recurrence mirrored human performance (N = 36) and were predictive of trial-by-trial responses across image durations (13-80 ms/image). Importantly, models with sequential lateral-recurrent integration also captured how human performance changes as a function of image presentation durations, with models processing images for a few time steps capturing human object recognition at shorter presentation durations and models processing images for more time steps capturing human object recognition at longer presentation durations. Furthermore, augmenting such a recurrent model with adaptation markedly improved dynamic recognition performance and accelerated its representational dynamics, thereby predicting human trial-by-trial responses using fewer processing resources. Together, these findings provide new insights into the mechanisms rendering object recognition so fast and effective in a dynamic visual world.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Increasing numbers of patients are presenting worldwide to emergency departments with suspected myocardial infarction. The use of point-of-care troponin assays might enable faster decision-making in ...this high-risk population and reduce the burden on emergency facilities. Here, we evaluate the diagnostic performance of a point-of-care high-sensitivity troponin I assay.
We conducted a prospective cohort study including patients presenting to the emergency department with suspected myocardial infarction from July 2013 to July 2016. A diagnostic algorithm for a high-sensitivity troponin I point-of-care assay was developed in a derivation data set with 669 patients and validated in an additional 610 patients.
The derived 0/1 h algorithm for the point-of-care assay consisted of an admission troponin I <4 ng/L and a δ from 0 h to 1 h <3 ng/L for rule out and an admission troponin I ≥90 ng/L or a δ from 0 h to 1 h ≥20 ng/L for rule in of non-ST-elevation myocardial infarction. Application to the validation cohort showed a negative predictive value of 99.7% (95% CI, 98.1%-100.0%) and 48.0% of patients ruled out, whereas 14.6% were ruled in with a positive predictive value of 86.5% (95% CI, 77.6%-92.8%). The diagnostic performance of the point-of-care high-sensitivity assay was highly comparable to guideline-recommended use of a laboratory-based high-sensitivity troponin assay.
The clinical application of a 0/1 h diagnostic algorithm based on a high-sensitivity troponin I point-of-care assay is safe, and diagnostic performance is comparable to a laboratory-based high-sensitivity troponin I assay.
Background
The Society of Cardiovascular Angiography and Interventions (SCAI) have recently proposed a new classification of cardiogenic shock (CS) dividing patients into five subgroups.
Objective
...Aim of this study was to apply the SCAI classification to a cohort of patients presenting with CS and to evaluate its ability to predict 30‐day survival.
Methods
SCAI CS subgroups were interpreted based on the recent consensus statement and then applied to N = 1,007 consecutive patients presenting with CS or large myocardial infarction (MI) between October 2009 and October 2017. The association between SCAI classification and 30‐day all‐cause mortality was assessed by logistic regression analysis.
Results
Mean age in the study cohort was 67 (±15) years, 72% were male. Mean lactate at baseline was 6.05 (±5.13) mmol/l and 51% of the patients had prior cardiac arrest. Overall survival probability was 50.6% (95% confidence interval CI 47.5–54.0%). In view of the SCAI classification, the survival probability was 96.4% (95% CI 93.7–99.0%) in class A, 66.1% (95% CI 50.2–87.1%) in class B, 46.1% (95% CI 40.6–52.4%) in class C, 33.1% (95% CI 26.6–41.1%) in class D, and 22.6% (95% CI 17.1–30.0%) in class E. Higher SCAI classification was significantly associated with lower 30‐day survival (p < .01).
Conclusion
In this large clinical cohort, the SCAI classification was significantly associated with 30‐day survival. This finding supports the rationale of the SCAI CS classification and calls for a validation in a prospective trial.
The human visual system is an intricate network of brain regions that enables us to recognize the world around us. Despite its abundant lateral and feedback connections, object processing is commonly ...viewed and studied as a feedforward process. Here, we measure and model the rapid representational dynamics across multiple stages of the human ventral stream using time-resolved brain imaging and deep learning. We observe substantial representational transformations during the first 300 ms of processing within and across ventral-stream regions. Categorical divisions emerge in sequence, cascading forward and in reverse across regions, and Granger causality analysis suggests bidirectional information flow between regions. Finally, recurrent deep neural network models clearly outperform parameter-matched feedforward models in terms of their ability to capture the multiregion cortical dynamics. Targeted virtual cooling experiments on the recurrent deep network models further substantiate the importance of their lateral and top-down connections. These results establish that recurrent models are required to understand information processing in the human ventral stream.
Abstract
Background
Emergency departments worldwide are increasingly adopting rapid diagnosis of patients with suspected myocardial infarction (MI) based on high-sensitivity troponin. We set out to ...assess the diagnostic accuracy of a high-sensitivity cardiac troponin I (hs-cTnI) assay in a prospective study.
Methods
In a cohort study including 1800 patients presenting with suspected acute MI, we developed and temporally validated a 0/1 h diagnostic algorithm using the Siemens Atellica IM hs-cTnI assay. The algorithm was established in the first 928 patients and validated in the following 872 patients.
Results
The derived algorithm consisted of a baseline rule-out of non–ST-segment elevation MI using a cutoff <3 ng/L in patients with symptom onset ≥3 h or an admission troponin I level <6 ng/L with a Δ change of <3 ng/L from 0 h to 1 h. For rule-in, an admission troponin I level ≥120 ng/L or an increase within the first hour ≥12 ng/L was required. Application of the algorithm to the validation cohort showed a negative predictive value of 99.8% (95% CI, 98.7%–100.0%), sensitivity of 99.1% (95% CI, 95.1%–100.0%), and 48.3% of patients ruled out, whereas 15.1% were ruled in with a positive predictive value of 68.0% (95% CI, 59.1%–75.9%) and specificity of 94.4% (95% CI, 92.5%–96.0%). The diagnostic performance was comparable to guideline-recommended application of an established hs-cTnI assay in a rapid 0/1 h strategy.
Conclusions
The Siemens hs-cTnI assay is well suited for application in rapid diagnostic stratification of patients with suspected MI.
Study Registration
www.clinicaltrials.gov (NCT02355457)
Discrimination among patients with type 1 myocardial infarction (T1MI), type 2 myocardial infarction (T2MI), and myocardial injury is difficult.
The aim of this study was to investigate the ...discriminative value of a 29-biomarker panel in an emergency department setting.
Patients presenting with suspected myocardial infarction (MI) were recruited. The final diagnosis in all patients was adjudicated on the basis of the fourth universal definition of MI. A panel of 29 biomarkers was measured, and multivariable logistic regression analysis was used to evaluate the associations of these biomarkers with the diagnosis of MI or myocardial injury. Biomarkers were chosen using backward selection. The model was internally validated using bootstrapping.
Overall, 748 patients were recruited (median age 64 years), of whom 138 had MI (107 T1MI and 31 T2MI) and 221 had myocardial injury. In the multivariable model, 4 biomarkers (apolipoprotein A-II, N-terminal prohormone of brain natriuretic peptide, copeptin, and high-sensitivity cardiac troponin I) remained significant discriminators between T1MI and T2MI. Internal validation of the model showed an area under the curve of 0.82. For discrimination between MI and myocardial injury, 6 biomarkers (adiponectin, N-terminal prohormone of brain natriuretic peptide, pulmonary and activation-regulated chemokine, transthyretin, copeptin, and high-sensitivity troponin I) were selected. Internal validation showed an area under the curve of 0.84.
Among 29 biomarkers, 7 were identified to be the most relevant discriminators between subtypes of MI or myocardial injury. Regression models based on these biomarkers allowed good discrimination. (Biomarkers in Acute Cardiac Care BACC; NCT02355457)
Display omitted
We present the first experimental realization of coherent Bragg scattering off a one-dimensional system-two strings of atoms strongly coupled to a single photonic mode-realized by trapping atoms in ...the evanescent field of a tapered optical fiber, which also guides the probe light. We report nearly 12% power reflection from strings containing only about 1000 cesium atoms, an enhancement of 2 orders of magnitude compared to reflection from randomly positioned atoms. This result paves the road towards collective strong coupling in 1D atom-photon systems. Our approach also allows for a straightforward fiber connection between several distant 1D atomic crystals.
Increased afterload results in ‘pathological’ cardiac hypertrophy, the most important risk factor for the development of heart failure. Current in vitro models fall short in deciphering the ...mechanisms of hypertrophy induced by afterload enhancement. The aim of this study was to develop an experimental model that allows investigating the impact of afterload enhancement (AE) on work-performing heart muscles in vitro. Fibrin-based engineered heart tissue (EHT) was cast between two hollow elastic silicone posts in a 24-well cell culture format. After 2 weeks, the posts were reinforced with metal braces, which markedly increased afterload of the spontaneously beating EHTs. Serum-free, triiodothyronine-, and hydrocortisone-supplemented medium conditions were established to prevent undefined serum effects. Control EHTs were handled identically without reinforcement. Endothelin-1 (ET-1)- or phenylephrine (PE)-stimulated EHTs served as positive control for hypertrophy. Cardiomyocytes in EHTs enlarged by 28.4 % under AE and to a similar extent by ET-1- or PE-stimulation (40.6 or 23.6 %), as determined by dystrophin staining. Cardiomyocyte hypertrophy was accompanied by activation of the fetal gene program, increased glucose consumption, and increased mRNA levels and extracellular deposition of collagen-1. Importantly, afterload-enhanced EHTs exhibited reduced contractile force and impaired diastolic relaxation directly after release of the metal braces. These deleterious effects of afterload enhancement were preventable by endothelin-A, but not endothelin-B receptor blockade. Sustained afterload enhancement of EHTs alone is sufficient to induce pathological cardiac remodeling with reduced contractile function and increased glucose consumption. The model will be useful to investigate novel therapeutic approaches in a simple and fast manner.
The effect of temperature changes on the light output of LAB based liquid scintillator is investigated in a range from
-
5
to
30
∘
C with
α
-particles and electrons in a small scale setup. Two PMTs ...observe the scintillator liquid inside a cylindrically shaped aluminum cuvette that is heated or cooled and the temperature dependent PMT sensitivity is monitored and corrected. The
α
-emitting isotopes in dissolved radon gas and in natural Samarium (bound to a LAB solution) excite the liquid scintillator mixtures and changes in light output with temperature variation are observed by fitting light output spectra. Furthermore, also changes in light output by compton electrons, which are generated from external calibration
γ
-ray sources, is analysed with varying temperature. Assuming a linear behaviour, a combined negative temperature coefficient of
(
-
0.29
±
0.01
)
%
/
∘
C
is found. Considering hints for a particle type dependency, electrons show
(
-
0.17
±
0.02
)
%
/
∘
C
, whereas the temperature dependency seems stronger for
α
-particles, with
(
-
0.35
±
0.03
)
%
/
∘
C
. Due to a high sampling rate, a pulse shape analysis can be performed and shows an enhanced slow decay component at lower temperatures, pointing to reduced non-radiative triplet state de-excitations.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Purpose
The aim of this study was to investigate whether textural features of tumour hypoxia, assessed with serial
18
Ffluoromisonidazole (FMISO)-PET, were able to predict clinical outcome in ...patients with head and neck squamous cell carcinoma (HNSCC, T1-4, N+, M0) during chemoradiotherapy (CRT).
Methods
In a preliminary evaluation of a prospective trial, tumour hypoxia was evaluated in 29 patients via serial FMISO-PET before and during CRT. All patients received an initial
18
Ffluorodeoxyglucose (FDG)-PET before CRT, and tumour regions were defined on this FDG-PET. The first-order metrics tumour-to-background ratio (TBR
mean
, TBR
max
, TBR
peak
), coefficient of variation, total lesion uptake and integral non-uniformity were calculated for all scans. Further, 3 second-order (textural) features from two grey-level matrices were calculated, as well as differential non-uniformity (
u
diff
). Prognostic value was examined by median split for group separation (GS) in Kaplan-Meier estimates and correlated with overall survival (OS), quantified via log-rank tests (
p
≤ 0.05) and group-relative hazard ratios (HR).
Results
Within a median follow-up of 29.6 months (95% CI: 16.8–48.0 months), no first-order metrics predicted OS with a significant GS (all
p
> 0.05) on any FMISO-PET scan. Only
u
diff
before and in week 2 during CRT (
p
= 0.03, HR = 10.8 and
p
= 0.05, HR = 5.2) and non-uniformity from grey-level run length matrix in week 2 separated prognostic groups (
p
= 0.05, HR = 5.3); lower values were correlated with better OS. Further, the decrease in
u
diff
from before CRT to week 2 was correlated with better OS (
p
= 0.04, HR = 9.4). FDG-PET before CRT did not predict outcome in any measure.
Conclusions
Textural features on FMISO-PET scans before CRT, in week 2 and, to a limited degree, the change of features during CRT, were able to identify head and neck squamous cell carcinoma patients with better OS, suggesting that a higher homogeneity of the degree of hypoxia in tumours could correlate with a better outcome after CRT.
PDF
