Lag Window Research Articles

Mean air temperature as a risk factor for stroke mortality in São Paulo, Brazil.

In Brazil, chronic diseases account for the largest percentage of all deaths among men and women. Among the cardiovascular diseases, stroke is the leading cause of death, accounting for 10% of all deaths. We evaluated associations between stroke and mean air temperature using recorded mortality data and meteorological station data from 2002 to 2011. A time series analysis was applied to 55,633 mortality cases. Ischemic and hemorrhagic strokes (IS and HS, respectively) were divided to test different impact on which subgroup. Poisson regression with distributed lag non-linear model was used and adjusted for seasonality, pollutants, humidity, and days of the week. HS mortality was associated with low mean temperatures for men relative risk (RR) = 2.43 (95% CI, 1.12-5.28) and women RR = 1.39 (95% CI, 1.03-1.86). RR of IS mortality was not significant using a 21-day lag window. Analyzing the lag response separately, we observed that the effect of temperature is acute in stroke mortality (higher risk among lags 0-5). However, for IS, higher mean temperatures were significant for this subtype with more than 15-day lag. Our findings showed that mean air temperature is associated with stroke mortality in the city of São Paulo for men and women and IS and HS may have different triggers. Further studies are needed to evaluate physiologic differences between these two subtypes of stroke.

Bayesian Inference and Prediction of Wave-induced Ship Motion based on Discrete-frequency Model Approximations

In this paper, we investigate the use of a discrete-frequency approximation for stochastic processes, modelling wave-induced ship motion and assess its prediction performance. The proposed estimator is obtained by adapting Bayesian spectral inference methods. We study the relationship between the lag window and the prediction performance and suggest a minimum requirement on window length for minimising the prediction error. We show an application to prediction of the pitch, heave, and roll motion from experiments of a scale-model container ship.

No Association Between Proton Pump Inhibitor Use and Risk of Alzheimer's Disease.

The objective of the study was to investigate whether proton pump inhibitor (PPI) use is associated with an increased risk of clinically verified Alzheimer's disease (AD). A Finnish nationwide nested case-control study MEDALZ includes all community-dwelling individuals with newly diagnosed AD during 2005-2011 (N=70,718), and up to four age-, sex-, and region of residence-matched comparison individuals for each case (N=282,858). Data were extracted from Finnish nationwide health-care registers. PPI use was derived from purchases recorded in the Prescription register data since 1995 and modeled to drug use periods with PRE2DUP method. AD was the outcome measure. PPI use was not associated with risk of AD with 3-year lag window applied between exposure and outcome (adjusted odds ratio (OR) 1.03, 95% confidence interval (CI) 1.00-1.05). Similarly, longer duration of use was not associated with risk of AD (1-3 years of use, adjusted OR 1.01 (95% CI 0.97-1.06); ≥3 years of use adjusted OR 0.99 (95% CI 0.94-1.04)). Higher dose use was not associated with an increased risk (≥1.5 defined daily doses per day, adjusted OR 1.03 (95% CI 0.92-1.14)). In conclusion, we found no clinically meaningful association between PPI use and risk of AD. The results for longer duration of cumulative use or use with higher doses did not indicate dose-response relationship.

Quantile Spectral Analysis for Locally Stationary Time Series

Summary Classical spectral methods are subject to two fundamental limitations: they can account only for covariance-related serial dependences, and they require second-order stationarity. Much attention has been devoted lately to quantile-based spectral methods that go beyond covariance-based serial dependence features. At the same time, covariance-based methods relaxing stationarity into much weaker local stationarity conditions have been developed for a variety of time series models. Here, we combine those two approaches by proposing quantile-based spectral methods for locally stationary processes. We therefore introduce a time varying version of the copula spectra that have been recently proposed in the literature, along with a suitable local lag window estimator. We propose a new definition of local strict stationarity that allows us to handle completely general non-linear processes without any moment assumptions, thus accommodating our quantile-based concepts and methods. We establish a central limit theorem for the new estimators and illustrate the power of the proposed methodology by means of a simulation study. Moreover, in two empirical studies (namely of the Standard & Poor's 500 series and a temperature data set recorded in Hohenpeissenberg), we demonstrate that the new approach detects important variations in serial dependence structures both across time and across quantiles. Such variations remain completely undetected and are actually undetectable, via classical covariance-based spectral methods.

Impact of energy efficiency interventions in public housing buildings on cold-related mortality: a case-crossover analysis.

Interventions to mitigate fuel poverty and particularly energy efficiency façade retrofitting (EEFR) have demonstrated positive impacts on health but the impacts of EEFR interventions on cold-related mortality have not been studied in depth. We evaluated the impact of EEFR interventions in Barcelona on the association between cold outdoor temperatures and mortality (from all natural causes and from neoplasms, circulatory system and respiratory system causes) from 1986 to 2012. A time-stratified case-crossover analysis was used. Relative risks (RR) for death related to extreme cold (lowest fifth percentile) in the no-intervention and intervention groups were obtained for temperature lag windows covering the day of the death and the previous 20 days (0-2, 3-5, 6-8, 9-11, 12-14, 15-17, 18-20). The statistical significance of the observed changes was evaluated using the RR for the cold temperature-intervention interaction. In men, interventions significantly increased the extreme cold-death association for the lag window 15-17 [interaction RR 2.23, 95% confidence interval (CI) 1.14-4.36]. The impacts were stronger for respiratory system causes and in men aged 75 or older. In women, on lag window 0-2, the extreme cold-death association was not significantly reduced when analysing all natural causes of death (interaction RR 0.46, 95% CI 0.21-1.01), but it was reduced significantly when analysing only deaths from neoplasms, circulatory system and respiratory system causes together. The impacts were stronger in women who died from circulatory system causes, in women with no education and in those aged 75 or older. EEFR interventions had differentiated effects on cold-related mortality in men and women. Differentiated effects were also observed by cause, educational level and age.

The relative accuracy and interpretability of five sequential analysis methods: A simulation study.

A variety of sequential analysis methods exist to quantify close temporal associations between events from direct observation data. In the present study, we compared the relative accuracy and interpretability of five sequential-analysis methods using simulated data. The methods included three existing approaches (event lag, concurrent interval, and time window) and two proposed modifications of the event lag approach (event lag with contiguous pauses and event lag with noncontiguous pauses) designed to address limitations of the existing approaches. We evaluated accuracy on the basis of the extent to which the mean contingency estimates produced by each method approximated a known mean (i.e., zero). We evaluated interpretability on the basis of the extent to which the contingency estimates produced by each method were independent from chance estimates of the two-event sequence. The results indicated that the event lag with contiguous pauses method produced the most accurate and interpretable estimates of contingency. This modified method prevents the total number of event types from influencing contingency estimates, thus solving a problem associated with the traditional event lag method.

Prediction of daily modes of South Asian monsoon variability and its association with Indian and Pacific Ocean SST in the NCEP CFS V2

The prediction capability of daily modes of variability for South Asian monsoon from climate forecast system version 2 of national centers for environmental prediction with respect to observed precipitation has been assessed. The space–time structure of the daily modes for summer monsoon rainfall has been identified by using multi-channel singular spectrum analysis (MSSA). The MSSA is applied on daily anomalies of rainfall data over the South Asian monsoon region (40°E–160°E, 30°S–35°N) for the period of 2001–2013 with a lag window of 61 days for June–July–August–September season. The broad spectrum around 45 and 50 days was obtained from the observed and model data during the time domain of our study. The space–time structure of the modes obtained from the model shows good resemblance with respect to the observation. The observed northeastward propagation of oscillatory mode is well simulated by the model. The significant improvement in the space–time structure, period of oscillation, and propagation of oscillatory modes was found in the model. The observed connectivity of oscillatory and persisting modes with the sea surface temperature of Indian and Pacific Ocean has also been investigated and it was found that the model is able to predict it reasonably well.

Aggregation of spectral density estimators

Given stationary time series data, we study the problem of finding the best linear combination of a set of lag window spectral density estimators with respect to the mean squared risk. We present an aggregation procedure and prove a sharp oracle inequality for its risk. We also provide simulations demonstrating the performance of our aggregation procedure, given Bartlett and other estimators of varying bandwidths as input. This extends work by P. Rigollet and A. Tsybakov on aggregation of density estimators.

OP0029 The Risk of Serious Infections in Patients Receiving Rituximab for Rheumatoid Arthritis

BackgroundIn the Europe, rituximab (RTX), an anti-CD20 monoclonal antibody, is reserved for patients with rheumatoid arthritis (RA) who have failed initial anti-TNF therapy. There is a concern over whether B-cell...

Statistical characterization of an ensemble of functional neural networks

This work uses a complex network approach to analyze temporal sequences of electrophysiological signals of brain activity from freely behaving rats. A network node represents a neuron and a network link is included between a pair of nodes whenever their firing rates are correlated. The framework of time varying graph (TVG) is used to deal with a very large number (>30 000) of time dependent networks, which are set up by taking into account correlations between neuron firing rates in a moving time lag window of suitable width. Statistical distributions for the following network measures are obtained: size of the largest connected cluster, number of edges, average node degree, and average minimal path. We find that the number of networks with highly correlated activity in distinct brain areas has a fat-tailed distribution, irrespective of the behavioral state of the animal. This contrasts with short-tailed distributions for surrogates obtained by shuffling the original data, and reflects the fact that neurons in the neocortex and hippocampus often act in precise temporal coordination. Our results also suggest that functional neuronal networks at the millimeter scale undergo statistically nontrivial rearrangements over time, thus delimitating an empirical constraint for models of brain activity.

Adaptive windowed cross Wigner–Ville distribution as an optimum phase estimator for PSK signals

Cross time-frequency distribution (XTFD), unlike the quadratic time-frequency distribution (QTFD), can represent phase information for phase shift keying (PSK) signals such as binary phase shift keying (BPSK) and quadrature phase shift keying (QPSK). However, it suffers from interference due to duplicated terms that cause inaccuracy in the phase estimate. Thus, a time-dependent lag window is proposed in this article where the window width is adaptively adjusted according to the localized lag autocorrelation function (LLAC). This adaptive windowed cross Wigner-Ville distribution (AWXWVD) is proven capable to remove most of the duplicated terms. The instantaneous information bearing phase (IIB-phase) is then estimated from the peak of the cross time-frequency representation (XTFR). Simulation results show that the AWXWVD is an optimum phase estimator as it meets the Cramer-Rao lower bound (CRLB) at SNR>=5 dB and SNR>=6 dB for BPSK and QPSK signals respectively. The AWXWVD also outperforms the conventional IQ demodulator as a phase estimator with difference in the IIB-phase variance of about 5 dB.

Revised eddy covariance flux calculation methodologies – effect on urban energy balance

ABSTRACTEddy covariance (EC) measurements of turbulent fluxes of momentum, sensible heat and latent heat – in addition to net radiation measurements – were conducted for three consecutive years in an urban environment: Helsinki, Finland. The aims were to: (1) quantify the detection limit and random uncertainty of turbulent fluxes, (2) assess the systematic error caused by EC calculation-procedure choices on the energy balance residual and (3) report the energy balance of the world's northernmost urban flux station. The mean detection limits were about 10% of the observed flux, and the random uncertainty was 9–16%. Of all fluxes, the latent heat flux – as measured with a closed-path gas analyser – was most prone to systematic calculation errors due to water vapour interactions with tube walls: using a lag window that is too small can cause a 15% lack of data (due to the dependency of lag time on relative humidity) and omitting spectral corrections can cause on average a 26% underestimation of the flux. The systematic errors in EC calculation propagate into the energy balance residual and can be larger than the residual itself: for example, omitting spectral corrections overestimates the residual by 13% or 18% on average, depending on the analyser.

Improved Windowed Cross Wigner-Ville Distribution for PSK Class Signals

Instantaneous information bearing phase is an important parameters in the analysis of a digitally phase modulated signal. Cross time-frequency distributions (XTFD) have been proposed recently to represent the class of Phase Shift Keying (PSK) signals. However, due to the presence of duplicated terms, the estimation of the instantaneous bearing phase is inaccurate. A fixed lag window width can be used as a kernel but it is insufficient to completely remove the duplicated terms. Here, a new improved windowed cross Wigner-Ville distribution (XWVD) is proposed as a method to obtain a better cross time-frequency representation (XTFR) for binary phase shift keying (BPSK) and quadrature phase shift keying (QPSK) signals. By adaptively adjusting the lag window width, the duplicated terms are removed and an accurate representation of the cross TFR are obtained at low SNR conditions, as low as 3dB.

Estimating surface CO<sub>2</sub> fluxes from space-borne CO<sub>2</sub> dry air mole fraction observations using an ensemble Kalman Filter

Abstract. We have developed an ensemble Kalman Filter (EnKF) to estimate 8-day regional surface fluxes of CO2 from space-borne CO2 dry-air mole fraction observations (XCO2) and evaluate the approach using a series of synthetic experiments, in preparation for data from the NASA Orbiting Carbon Observatory (OCO). The 32-day duty cycle of OCO alternates every 16 days between nadir and glint measurements of backscattered solar radiation at short-wave infrared wavelengths. The EnKF uses an ensemble of states to represent the error covariances to estimate 8-day CO2 surface fluxes over 144 geographical regions. We use a 12×8-day lag window, recognising that XCO2 measurements include surface flux information from prior time windows. The observation operator that relates surface CO2 fluxes to atmospheric distributions of XCO2 includes: a) the GEOS-Chem transport model that relates surface fluxes to global 3-D distributions of CO2 concentrations, which are sampled at the time and location of OCO measurements that are cloud-free and have aerosol optical depths <0.3; and b) scene-dependent averaging kernels that relate the CO2 profiles to XCO2, accounting for differences between nadir and glint measurements, and the associated scene-dependent observation errors. We show that OCO XCO2 measurements significantly reduce the uncertainties of surface CO2 flux estimates. Glint measurements are generally better at constraining ocean CO2 flux estimates. Nadir XCO2 measurements over the terrestrial tropics are sparse throughout the year because of either clouds or smoke. Glint measurements provide the most effective constraint for estimating tropical terrestrial CO2 fluxes by accurately sampling fresh continental outflow over neighbouring oceans. We also present results from sensitivity experiments that investigate how flux estimates change with 1) bias and unbiased errors, 2) alternative duty cycles, 3) measurement density and correlations, 4) the spatial resolution of estimated flux estimates, and 5) reducing the length of the lag window and the size of the ensemble. At the revision stage of this manuscript, the OCO instrument failed to reach its orbit after it was launched on 24 February 2009. The EnKF formulation presented here is also applicable to GOSAT measurements of CO2 and CH4.

Exact simulation of Gaussian Time Series from Nonparametric Spectral Estimates with Application to Bootstrapping

The circulant embedding method for generating statistically exact simulations of time series from certain Gaussian distributed stationary processes is attractive because of its advantage in computational speed over a competitive method based upon the modified Cholesky decomposition. We demonstrate that the circulant embedding method can be used to generate simulations from stationary processes whose spectral density functions are dictated by a number of popular nonparametric estimators, including all direct spectral estimators (a special case being the periodogram), certain lag window spectral estimators, all forms of Welch's overlapped segment averaging spectral estimator and all basic multitaper spectral estimators. One application for this technique is to generate time series for bootstrapping various statistics. When used with bootstrapping, our proposed technique avoids some --- but not all --- of the pitfalls of previously proposed frequency domain methods for simulating time series.

An autocorrelation model with place dependence to account for the effect of harmonic number on fundamental frequency discrimination

Fundamental frequency (f0) difference limens (DLs) were measured as a function of f0 for sine- and random-phase harmonic complexes, bandpass filtered with 3-dB cutoff frequencies of 2.5 and 3.5 kHz (low region) or 5 and 7 kHz (high region), and presented at an average 15 dB sensation level (approximately 48 dB SPL) per component in a wideband background noise. Fundamental frequencies ranged from 50 to 300 Hz and 100 to 600 Hz in the low and high spectral regions, respectively. In each spectral region, f0 DLs improved dramatically with increasing f0 as approximately the tenth harmonic appeared in the passband. Generally, f0 DLs for complexes with similar harmonic numbers were similar in the two spectral regions. The dependence of f0 discrimination on harmonic number presents a significant challenge to autocorrelation (AC) models of pitch, in which predictions generally depend more on spectral region than harmonic number. A modification involving a "lag window"is proposed and tested, restricting the AC representation to a limited range of lags relative to each channel's characteristic frequency. This modified unitary pitch model was able to account for the dependence of f0 DLs on harmonic number, although this correct behavior was not based on peripheral harmonic resolvability.

Effects of relative frequency, absolute frequency, and phase on fundamental frequency discrimination: Data and an autocorrelation model

Fundamental frequency (F0) difference limens (DLs) were measured versus F0 for sine- and random-phase harmonic complexes bandpass-filtered into low- or high-frequency regions, with 3-dB passbands of 2.5–3.5 and 5–7 kHz, respectively. In all cases, F0 DLs decreased dramatically with increasing F0 as approximately the tenth harmonic appeared in the passband. Generally, F0 DLs were similar in both frequency regions for complexes with similar harmonic numbers and phase relationships. However, F0 DLs were larger in the high-frequency than the low-frequency region for random-phase complexes containing only harmonics above the tenth, suggesting a possible role for additional fine-structure information in the low-frequency region. The dependence of F0 discrimination on relative frequency presents a significant challenge to autocorrelation (AC) models of pitch perception, in which predictions generally depend more on absolute frequency and phase locking. To represent this relative frequency effect, a ‘‘lag window’’ modification to the Meddis and O’Mard [J. Acoust. Soc. Am. 102, 1811–1820 (1997)] AC model was introduced, restricting each channel’s AC representation to a limited range of lags relative to the center frequency (CF). Thus in each channel, the AC model responds best to F0’s whose dominant harmonics fall near the CF. [Work supported by NIH Grant Nos. 5T32-DC-00038 and R01-DC-05216.]

Performance of quadratic time~frequency distributions as instantaneous frequency estimators

General performance analysis of the shift covariant class of quadratic time-frequency distributions (TFDs) as instantaneous frequency (IF) estimators, for an arbitrary frequency-modulated (FM) signal, is presented. Expressions for the estimation bias and variance are derived. This class of distributions behaves as an unbiased estimator in the case of monocomponent signals with a linear IF. However, when the IF is not a linear function of time, then the estimate is biased. Cases of white stationary and white nonstationary additive noises are considered. The well-known results for the Wigner distribution (WD) and linear FM signal, and the spectrogram of signals whose IF may be considered as a constant within the lag window, are presented as special cases. In addition, we have derived the variance expression for the spectrogram of a linear FM signal that is quite simple but highly signal dependent. This signal is considered in the cases of other commonly used distributions, such as the Born-Jordan and the Choi-Williams distributions. It has been shown that the reduced interference distributions outperform the WD but only in the case when the IF is constant or its variations are small. Analysis is extended to the IF estimation of signal components in the case of multicomponent signals. All theoretical results are statistically confirmed.

The bias of lag window estimators of the fractional difference parameter

An approximation for the bias in lag window estimators of the degree of differencing in fractionally integrated time series models is derived. The expression obtained is compared with the observed bias from simulations for various windows.

Adaptive instantaneous frequency estimation of multicomponent FM signals using quadratic time-frequency distributions

An adaptive approach to the estimation of the instantaneous frequency (IF) of nonstationary mono- and multicomponent FM signals with additive Gaussian noise is presented. The IF estimation is based on the fact that quadratic time-frequency distributions (TFDs) have maxima around the IF law of the signal. It is shown that the bias and variance of the IF estimate are functions of the lag window length. If there is a bias-variance tradeoff, then the optimal window length for this tradeoff depends on the unknown IF law. Hence, an adaptive algorithm with a time-varying and data-driven window length is needed. The adaptive algorithm can utilize any quadratic TFD that satisfies the following three conditions: First, the IF estimation variance given by the chosen distribution should be a continuously decreasing function of the window length, whereas the bias should be continuously increasing so that the algorithm will converge at the optimal window length for the bias-variance tradeoff, second, the time-lag kernel filter of the chosen distribution should not perform narrowband filtering in the lag direction in order to not interfere with the adaptive window in that direction; third, the distribution should perform effective cross-terms reduction while keeping high resolution in order to be efficient for multicomponent signals. A quadratic distribution with high resolution, effective cross-terms reduction and no lag filtering is proposed. The algorithm estimates multiple IF laws by using a tracking algorithm for the signal components and utilizing the property that the proposed distribution enables nonparametric component amplitude estimation. An extension of the proposed TFD consisting of the use of time-only kernels for adaptive IF estimation is also proposed.