Dynamic Time Series Research Articles

Time-series dynamics in symbiont communities of acroporid corals undergoing cold stress in winter

Time-series dynamics in symbiont communities of acroporid corals undergoing cold stress in winter

INTELLIGENT MODEL FOR CLASSIFYING HEMODYNAMIC PATTERNS OF BRAIN ACTIVATION TO IDENTIFY NEUROCOGNITIVE MECHANISMS OF SPATIAL-NUMERICAL ASSOCIATIONS

The study presents the results of the development and testing of deep learning neural network architectures, which demonstrate high accuracy rates in classifying neurophysiological data, in particular hemodynamic brain activation patterns obtained by functional near-infrared spectroscopy, during solving mathematical problems on spatial-numerical associations. The analyzed signal represents a multidimensional time series of oxyhemoglobin and deoxyhemoglobin dynamics. Taking the specificity of the fNIRS signal into account, a comparative analysis of 2 types of neural network architectures was carried out: (1) architectures based on recurrent neural networks: recurrent neural network with long short-term memory, recurrent neural network with long short-term memory with fully connected layers, bidirectional recurrent neural network with long short-term memory, convolutional recurrent neural network with long short-term memory; (2) architectures based on convolutional neural networks with 1D convolutions: convolutional neural network, fully convolutional neural network, residual neural network. Trained long short-term memory recurrent neural network architectures showed worse results in accuracy in comparison with 1D convolutional neural network architectures. Residual neural network (model_Resnet) demonstrated the highest accuracy rates in three experimental conditions more than 88% in detecting age-related differences in brain activation during spatial-numerical association tasks considering the individual characteristics of the respondents’ signal.

Multiscale non-linear tale risk spillover effect from oil to stocks – The case of East European emerging markets

This paper investigates the multiscale non-linear risk transmission effect from Brent oil to eleven European emerging stock markets. Dynamic extreme risk time series are created using the FIAPARCH-CVaR approach. The MODWT transformation is applied to make three wavelet details that represent different time horizons. In the final step, the MODWT time series are fitted into the Markov switching model to examine the spillover phenomenon. The results indicate that the Czech and Hungarian stock markets endure the spillover effect in crisis regime in the short term, probably because these markets are among the most efficient emerging European markets. On the other hand, a relatively high spillover effect is found in a peaceful rather than a crisis regime in the case of Poland. This is probably because the Polish index lists almost 300 stocks, which means that oil shocks disperse to a large number of different industry sectors. In small and less developed markets, such as Estonia, Slovenia, Bulgaria, and Croatia, a high spillover effect exists in a tranquil regime because these countries have high oil consumption per capita. Lithuania and Latvia do not report the spillover effect in the short run, while this is true for all time horizons in the case of Slovakia.

Online Learning of Temporal Association Rule on Dynamic Multivariate Time Series Data

Online Learning of Temporal Association Rule on Dynamic Multivariate Time Series Data

Autoregressive With Slack Time Series Model for Forecasting a Partially-Observed Dynamical Time Series

Autoregressive With Slack Time Series Model for Forecasting a Partially-Observed Dynamical Time Series

Energy-efficient Online Continual Learning for Time Series Classification in Nanorobot-based Smart Health.

Nanorobots have been used in smart health to collect time series data such as electrocardiograms and electroencephalograms. Real-time classification of dynamic time series signals in nanorobots is a challenging task. Nanorobots in the nanoscale range require a classification algorithm with low computational complexity. First, the classification algorithm should be able to dynamically analyze time series signals and update itself to process the concept drifts (CD). Second, the classification algorithm should have the ability to handle catastrophic forgetting (CF) and classify historical data. Most importantly, the classification algorithm should be energy-efficient to use less computing power and memory to classify signals in real-time on a smart nanorobot. To solve these challenges, we design an algorithm that can Prevent Concept Drift in Online continual Learning for time series classification (PCDOL). The prototype suppression item in PCDOL can reduce the impact caused by CD. It also solves the CF problem through the replay feature. The computation per second and the memory consumed by PCDOL are only 3.572M and 1KB, respectively. The experimental results show that PCDOL is better than several state-of-the-art methods for dealing with CD and CF in energy-efficient nanorobots.

Fractal analysis of the medical insurance market: Theoretical and practical aspects

The latest experience obviously reflects that social-economic systems increasingly demonstrate an unforecastable behaviour. Financial markets are characterised by extremely guick change and crisis phenomena. Use of traditional modelling methods (in particular, the effective market hypothesis) provides no opportunity to obtain efficient forecasts about state and development of financial markets. One of promising methods in researching and modelling financial markets is the fractal analysis. Various natural and social phenomena posess fractality properties, namely there are similar structures on different scales. Generality of the fractal analysis methods allows their applying to study systems of any nature – from physical to economic and social ones. Fractality attributes of time series provide its dynamics pre-forecast and assess the time series predictability. The fractal analysis consists in establishing the extent of time series similarity to the fractal ones and defining a relation between the trend line and fractal dimension. The Hurst exponent assessment (persistence or anti-persistence of time series) predicts the further process development on the preliminary data basis. The fractal analysis efficiency during the unstable market periods is applicable for studying different social-economic systems (in particular, the Ukrainian medical insurance market). The research object is time series of gross insurance rewards on the Ukrainian medical insurance market. The research topic is the time series fractal analysis. The paper deals with peculiarities of the time series fractal analysis, the R/S analysis for time series of gross insurance rewards on the Ukrainian medical insurance market (via the MS Excel software). The Statistica 10.0.228.2 application generated the ARIMA model to predict dynamics of the gross insurance rewards on the Ukrainian medical insurance market. The obtained results may be used to conduct the R/S analysis of financial time series. In particular, that concerns those time series that describe the insurance market. Also, dynamics of the financial time series may be forecasted via these results as well.

Short-Term Load Forecasting Using Nonlinear Dynamic Time Series Autoregressive with Feedback for AZADI Control-Station Erbil-Iraq

The load forecasting is a human or computational technique for accurate preanticipation of electrical load to enhance reliable operation and optimal planning control of system plant for electrical energy flowing without facing any economical and technical limitations, therefore appropriate estimation for present and future consumption cost of electrical loads which are necessary to predict the load demand for generating near to accurate power. During advanced technology at the last few decades, artificial neural networks(ANNs) have been extensively employed in electrical system, they are trained using historical data obtained from plant station. This work is intended to be a study of short-term load forecasting (STLF) basis for a power predicted applied to the actual past load data displayed from Azadi station for Feb.2022 were used in training and validation system of neural grid. The result was evaluated by mean square percentage error of (32.7) for the forecasting dynamic time series method to solve the data over hours, days, and weeks in advance, using a kind of non-linear filtering. Short-term load forecasting tried out with main stages; predicted power load data sets, network training, and forecasting. Neural network used has 3-layers: an input, a hidden, and an output layer. The number of hidden layer neurons can be varied for the different network performance. The active power generation faces economical and technical challenges, therefore appropriate evaluation of loads are much needed. Index Terms— ANN, STLF, MW, Neurons.

Comparative evaluation of different spinal stability metrics

Direct in vivo measurements of spinal stability are not possible, leaving computational estimations (such as dynamic time series and structural analyses) as the feasible option. However, differences between different stability assessment approaches and metrics remain unclear. To explore this, we asked 32 participants to perform 35 cycles of repetitive lifts with and without load (4/2.6 kg for males/females). EMG signals and 3D kinematics were collected via 12 surface electrodes and 17 inertial sensors, and three dynamical stability measures were computed: short and long temporal and conventional maximum Lyapunov exponents (LyE) and maximum Floquet multipliers (FM). A dynamic subject-specific EMG-assisted musculoskeletal model computed four structural stability measures (critical muscle stiffness coefficient at which spine becomes unstable, average spine stiffness, minimum and geometric average of Hessian matrix eigenvalues). Across cycles, dynamical and structural stability outcomes varied noticeably. Temporal short-term LyE and all structural stability measures were more influenced by the cycle percentage (posture factor) than by phase (lifting, lowering) or load factor. The effect of all factors were non-significant for FM and long LyE, except for the posture on LyE-L with a small effect size. Pearson's correlations revealed a weak to moderate, or non-existent, correlation between structural and dynamical stability metrics, with small shared variances, underscoring their distinct and independent nature and theoretical foundations. Moreover, the low sensitivity of dynamic measures to posture and load factors, found in this study, calls for further examination. Considering the limitations and shortcomings of both dynamical and structural stability assessment approaches, there is a need for the development of improved musculoskeletal stability evaluation techniques.

Leveraging high resolution classifications and random forests for hindcasting decades of mesic ecosystem dynamics in the Landsat time series

Mesic ecosystems are fundamental to conservation efforts in semi-arid systems, but are threatened by climate change and development. Newer earth observation datasets, including Sentinel-1 and −2, provide opportunities to monitor mesic ecosystems at meaningful spatial scales, but are insufficient for measuring decadal-scale changes. Conversely, the Landsat time series has decades of data, but images are spatially coarse relative to many of the mesic ecosystem areas that sustain dryland systems, resulting in classifications with mixed pixels inadequate for effective monitoring. We developed a workflow that uses 10-m classifications produced from fusion of the Sentinel-1 and −2 time series (2017–2020) to estimate sub-pixel proportions of Landsat time series observations (2004–2020). Using random forest regression models, we quantified water resource proportions (WRP) of surface water, mesic vegetation, and upland land covers within each 30-m Landsat pixel. We incorporated ancillary covariates to account for varying topographic conditions, land cover, and climate. Results indicate that our approach consistently estimates sub-pixel proportions of Landsat pixels more accurately compared to spectral mixture analysis (SMA). The WRP product for surface water had up to 8% less error than SMA as measured by Mean Absolute Error (MAE) and up to 17% less error as measured by Root Mean Squared Error (RMSE). For mesic vegetation, the WRP product outperformed SMA by up to 4% (MAE) and 7% (RMSE). Finally, we demonstrated the ability of our time series to characterize historical water resource availability at a case study site with a well documented restoration history by qualitatively examining the mesic vegetation dynamics time series to identify system responses to restoration efforts. Our approach allows us to hindcast observations of Sentinel products and measure water resource dynamics with greater precision over larger temporal scales. We envision these WRP data to be useful for measuring the impacts of conservation interventions, disturbance recovery, or land use changes that pre-date the Sentinel time series.

Impacts of non-locality and memory kernel of fractional derivative, awareness and treatment strategies on HIV/AIDS prevalence

We propose a Liouville–Caputo non-linear fractional order α(0<α≤1) epidemic model to understand the transmission dynamics of infectious diseases, like HIV/AIDS, and explore the presence of memory kernel of fractional derivative for the precautionary measure against the transmission of disease to control HIV/AIDS prevalence. A proper sex education, which depends on number of HIV and AIDS, to unaware susceptible as well as fear of infection among these aware susceptible are employed here. Treatment compartment is incorporated to treat the symptomatic HIV and AIDS infected individuals.Using the Banach fixed point theorem, we first ensure the positivity, existence and uniqueness of solution, and biologically feasible region. Next after defining the basic reproduction number (R0), its uses on stability (local and global) and direction of bifurcation are manifested here. A quantitative assessment is conducted onR0and then we explore the sensitive parameter/s, which control the HIV/AIDS prevalence.Numerical experiment is conducted based on the Adams–Bashforth–Moulton predictor corrector method to obtain the approximate solution of the proposed model. Impacts of disease transmission rate, awareness, psychological fear of infection and treatment strategies on the HIV/AIDS prevalence are analysed based on non-linear dynamical tools, like, time series, phase diagram and spectra analysis. An effort is made to see how the order and kernel of the fractional derivative play a crucial role on the dynamics of the system and in the precautionary measure against the transmission of disease during transition as well as post-transition phases. We employ intervention strategies for controlling the spread as well as possibility of eradicating the HIV/AIDS prevalence. It is seen that the reduction ofαis comparable to awareness, psychological fear of infection and treatment strategies to control disease. Finally, it is shown that the use of fractional derivative in disease transmission dynamics is one of the most efficient applied mathematics content compatible with the natural characteristics of human memory.

A novel continuous delay hidden layer deep belief network and its application in life prediction of rolling bearings

The traditional rolling bearing trend prediction method is difficult to predict the bearing degradation process comprehensively and scientifically, and the feature extraction is highly dependent on the manual experience method, and the prediction ability of the bearing measurement data with strong dynamic characteristics and nonlinear is very limited. Therefore, how to build a high-precision life prediction model has become an urgent problem and challenge. In this paper, a continuous delay hidden layer deep belief network (CDHLDBN), kernel principal component analysis (KPCA)-CDHLDBN, is proposed based on the KPCA preprocessing method. Firstly, a continuous deep belief network (CDHLDBN) with a delay layer is proposed, which makes the output of the hidden layer more fully consider the influence of the previous moment on the output of this moment, so as to avoid the defect of DBN model which is difficult to process the data with strong dynamic time series. Then, all visible layers of a continuous DBN with a delay layer are de-averaged, and then the simulated bearing sequence data is trained and predicted using the KPCA-CDHLDBN life prediction model. The results show that KPCA-CDHLDBN model can predict nonlinear dynamic time series data better than other deep learning models. Finally, it is applied to the performance degradation life prediction of rolling bearings. The performance of the proposed method was verified by the bearing accelerated life vibration data provided by PRONOSTIA experimental platform, and the performance of the proposed method was compared with that of the traditional DBN and its improved prediction model. The results show that KPCA-CDHLDBN has better prediction accuracy and faster prediction speed.

Estimation of grain filling rate of winter wheat using leaf chlorophyll and LAI extracted from UAV images

The grain filling rate (GFR) plays a pivotal role in assessing winter wheat grain development and guiding crop variety selection and cultivation strategies. Presently, GFR monitoring primarily relies on destructive sampling methods, which are impractical for swift assessments in extensive agricultural fields. Our study introduces a novel approach for estimating winter wheat grain filling rate (GFR) using leaf chlorophyll content (LCC) and Leaf Area Index (LAI) extracted from UAV imagery. Initially, we established a quantitative relationship between GFR and LAI as well as LCC in winter wheat. Subsequently, UAV multispectral imagery was employed to track the time-series dynamics of LAI and LCC during the grain-filling stage. Ultimately, we constructed an estimation model for winter wheat GFR and achieved plot-scale GFR mapping through UAV-based imaging technology. To identify sensitive vegetation index combinations for LAI and chlorophyll content (SPAD), we employed adaptive re-weighted sampling and continuous projection algorithms, respectively. Following this, we employed partial least squares and random forest algorithms to develop LAI and SPAD estimation models, which were then compared to full-band regression. Finally, GFR estimation was based on the conversion models of LAI and SPAD to GFR. Our findings revealed that, despite variations in water supply, there were no significant differences in GFR trends among the irrigated treatments. However, in the nitrogenous fertilizer application treatments, GFR initially declined with increased nitrogenous fertilizer supply, reaching its peak between 21 and 25 days post-flowering, during which period the differences in GFR among the various nitrogenous fertilizer application treatments were not significant. As the reproductive period advanced, higher nitrogenous fertilizer treatments maintained a higher GFR for a more extended period and experienced a slower decline compared to the lower nitrogenous fertilizer treatments. LAI validation yielded an R2 of 0.84, a root mean square error (RMSE) of 0.39, and a normalized root mean square error (NRMSE) of 10.9%. SPAD validation resulted in R2, RMSE, and NRMSE values of 0.79%, 1.30%, and 12.5%, respectively. The GFR estimation achieved an R2 of 0.77, an RMSE of 0.32 mg/d, and an NRMSE of 11.3%. Mapping the GFR distribution in the field, coupled with water and fertilizer management, supports UAVs in monitoring crop growth and refining field-level water and fertilizer decisions.

A novel and effective method for quantifying complexity of nonlinear time series

In this work, we propose the discrete general cumulative residual Kullback-Leibler information(DGCKL)to measure the complexity of nonlinear time series. Theoretically, the novel method benefits from the core concept of distribution entropy (DistEn)and possesses the advantages of the general cumulative Kullback-Leibler information(GCKL), presenting significant superiorities in quantifying the complexity of time series. Furthermore, comparative experiments conducted on simulation data confirm thatDGCKLis robust to noise and inherits the excellent properties ofDistEn to extract the features of time series. In practical application, DGCKLis provided with the ability to detect the abnormal behaviors of the rail dynamic time series and identify the significant events in specific periods of stock indices. Supported by theoretical analysis, simulated research and experimental verifications, this work establishes DGCKLas a meaningful and practical model for quantifying the complexity of time series.

Effect of SUVR reference region selection on sampled iterative local approximation estimates of amyloid onset age

AbstractBackgroundAmyloid PET imaging quantifies Aß plaques informing longitudinal modeling methods such as sampled iterative local approximation (SILA) that provide individualized estimated amyloid onset age (EAOA). EAOA is defined as the transition to abnormal Aß accumulation that occurs in Alzheimer’s disease. This work investigates the effect of SUVR reference region selection on EAOA by comparing DVR‐trained versus SUVR‐trained SILA models with three different SUVR reference regions.MethodN = 585 participants (mean (SD) baseline age = 66.40 (7.89) years, 512 unimpaired, 44 MCI, 28 dementia, 1 unknown) from the Wisconsin Registry for Alzheimer’s Prevention and Wisconsin Alzheimer’s Disease Research Center underwent T1‐weighted MRI and dynamic 11C‐Pittsburgh Compound B (PiB) amyloid PET imaging (0‐70 minutes after bolus injection). Reconstructed dynamic PET time series were processed and quantified using MR‐guided imaging pipelines optimized for DVR (Logan, k2’ = 0.149 min‐1, cerebellum gray matter (CBLM‐GM) reference region) and SUVR (50‐70 min frames) quantification with CBLM‐GM, whole cerebellum (WHOLE‐CBLM), or subcortical white matter (WM) as a reference region. Mean cortical PiB DVR and SUVR were each calculated by averaging across eight bilateral ROIs. Group‐based trajectory modeling was applied to longitudinal DVR data to establish an amyloid positivity (A+) threshold (DVR&gt;1.14; equivalent to 14 centiloids). LME was used to define the relationship between DVR and SUVR and translate the DVR threshold to SUVR (SUVRCBLM GM&gt;1.28; SUVRWHOLE CBLM&gt;1.07; SUVRWM&gt;0.68). Within‐person paired differences between EAOADVR and EAOASUVR_REFERENCE_REGION are reported as (mean; [95% CI]).ResultEAOA derived from DVR and SUVR were compared for n = 198 (34%) DVR A+ participants. Amyloid vs. age data are shown for each quantification approach in Figure 1. EAOASUVR‐WHOLE‐CBLM (‐1.29 years; [‐1.51, ‐1.08]) had the highest agreement with EAOADVR compared to EAOASUVR‐CBLM‐GM (‐1.44 years; [‐1.75, ‐1.12]) and EAOASUVR‐WM (‐0.67 years; [‐2.13, 0.79]; Figure 2). EAOA variability was highest for DVR&gt;1.8 where fewer longitudinal observations were available for model training.ConclusionSUVR with whole cerebellum as a reference region resulted in the smallest difference in EAOA compared to DVR from the SILA model. Future work is ongoing to define SUVR thresholds that will optimize agreement with DVR‐trained EAOA results for harmonization of temporal estimates.

Effect of SUVR reference region selection on sampled iterative local approximation estimates of amyloid onset age

AbstractBackgroundAmyloid PET imaging quantifies Aβ plaques informing longitudinal modeling methods such as sampled iterative local approximation (SILA) that provide individualized estimated amyloid onset age (EAOA). EAOA is defined as the transition to abnormal Aβ accumulation that occurs in Alzheimer’s disease. This work investigates the effect of SUVR reference region selection on EAOA by comparing DVR‐trained versus SUVR‐trained SILA models with three different SUVR reference regions.MethodN = 585 participants (mean (SD) baseline age = 66.40 (7.89) years, 512 unimpaired, 44 MCI, 28 dementia, 1 unknown) from the Wisconsin Registry for Alzheimer’s Prevention and Wisconsin Alzheimer’s Disease Research Center underwent T1‐weighted MRI and dynamic 11C‐Pittsburgh Compound B (PiB) amyloid PET imaging (0‐70 minutes after bolus injection). Reconstructed dynamic PET time series were processed and quantified using MR‐guided imaging pipelines optimized for DVR (Logan, k2’ = 0.149 min−1, cerebellum gray matter (CBLM‐GM) reference region) and SUVR (50‐70 min frames) quantification with CBLM‐GM, whole cerebellum (WHOLE‐CBLM), or subcortical white matter (WM) as a reference region. Mean cortical PiB DVR and SUVR were each calculated by averaging across eight bilateral ROIs. Group‐based trajectory modeling was applied to longitudinal DVR data to establish an amyloid positivity (A+) threshold (DVR&gt;1.14; equivalent to 14 centiloids). LME was used to define the relationship between DVR and SUVR and translate the DVR threshold to SUVR (SUVRCBLM GM&gt;1.28; SUVRWHOLE CBLM&gt;1.07; SUVRWM&gt;0.68). Within‐person paired differences between EAOADVR and EAOASUVR_REFERENCE_REGION are reported as (mean; [95% CI]).ResultEAOA derived from DVR and SUVR were compared for n = 198 (34%) DVR A+ participants. Amyloid vs. age data are shown for each quantification approach in Figure 1. EAOASUVR‐WHOLE‐CBLM (‐1.29 years; [‐1.51, ‐1.08]) had the highest agreement with EAOADVR compared to EAOASUVR‐CBLM‐GM (‐1.44 years; [‐1.75, ‐1.12]) and EAOASUVR‐WM (‐0.67 years; [‐2.13, 0.79]; Figure 2). EAOA variability was highest for DVR&gt;1.8 where fewer longitudinal observations were available for model training.ConclusionSUVR with whole cerebellum as a reference region resulted in the smallest difference in EAOA compared to DVR from the SILA model. Future work is ongoing to define SUVR thresholds that will optimize agreement with DVR‐trained EAOA results for harmonization of temporal estimates.

IoT Botnet Detection Based on Anomalies of Multiscale Time Series Dynamics

In this work, we propose a solution for detecting botnet attacks in the Internet of Things (IoT) by identifying anomalies in the temporal dynamics of their devices. Given their limited computing capabilities, IoT devices are more vulnerable to attacks than conventional computers. In this scenario, botnets have a high degree of severity since they are used to trigger distributed denial-of-service attacks, which are amplified by a large number of IoT devices. Thus, solutions aiming to identify and mitigate the damage caused by botnets in IoT are urgent and essential. We evaluate the number of packets a device transmits, following a multiscale ordinal patterns transformation, and uses Isolation Forest for anomaly detection. By investigating how devices evolve, we can distinguish between normal and anomalous behaviors. We apply the proposed solution to detect two major botnets for IoT: Mirai and Bashlite. We evaluated our model throughout two experimental setups. The first, using a single model for all devices, reaching 99.5% of accuracy and 99.6% of specificity, and the second, by tuning a model per device, reaching 100% of accuracy. These results show that, with the proper transformation, it is possible to use simple methods for detecting anomalies in IoT devices' behaviors.

ARIMA approach for temperature and rainfall time series prediction in Punjab

The present study aims to explore the effectiveness of Seasonal Autoregressive Integrated Moving Average (SARIMA) models in forecasting meteorological time series data exhibiting seasonal patterns. We compared the performance of SARIMA models with different configurations and evaluate their forecasting accuracy using real-world meteorological datasetsfor three different agroclimatic zones of Punjab (sub mountainous region, central region and south west region) was analyzed to forecast mean monthly maximum air temperature, minimum air temperature and rainfall. The weather data was used from 1984-2022 for sub-mountainous zone (Ballowal Saunkhri), 1970-2022 for Central zone (Ludhiana) and 1977-2022 for south west zone (Bathinda). The results provide insights into the suitability and limitations of SARIMA models for meteorological forecasting and offer practical recommendations for practitioners and researchers in the field. The goodness of fit was tested against residuals using Ljung-Box test. The accuracy of the model was tested using Mean Absolute Error (MAE) and root square mean error (RMSE). The model achieved Mean Absolute Errors (MAE) ranging from 0.61 to 0.78 for maximum temperature, 0.74 to 0.49 for minimum temperature, and 32.12 to 45.44 for rainfall, with lower MAE values indicating higher predictive accuracy. The fitted model was able to capture dynamics of the temperature time series and produce a sensible forecast. However, the model was unable to forecast rainfall series efficiently.

A closer look at residential water demand elasticities in the short and long run

Hotter, drier summers and population growth are causing water management challenges for water authorities in New Zealand. Effective long-term planning requires understanding the dynamics of household responses to demand management and climate variables. This paper analyses a large dataset of household-level demand data over nine years. Dynamic panel time series models are used to estimate short and long-term elasticities, interactions between demand management and climate variables, and associations with property characteristics. The general belief is that water demand becomes more sensitive to price changes over the long term compared to the short term. Yet, previous studies that directly compared long and short-run elasticities had limitations, as they imposed a constraint on the long-run estimate, assuming it to be larger. The models presented in this paper are more flexible and can identify the opposite scenario should it be present. We find long-run price elasticity may not be larger for new high-value properties or high-value properties with small gardens. Low-value properties have smaller short-run and larger long-run price elasticity. High-value properties respond more to outdoor restrictions and associated conservation messages than low-value properties. We discuss the policy implications of these findings.

Enhanced Dynamic Laterality Based on Functional Subnetworks in Patients with Bipolar Disorder.

An ocean of studies have pointed to abnormal brain laterality changes in patients with bipolar disorder (BD). Determining the altered brain lateralization will help us to explore the pathogenesis of BD. Our study will fill the gap in the study of the dynamic changes of brain laterality in BD patients and thus provide new insights into BD research. In this work, we used fMRI data from 48 BD patients and 48 normal controls (NC). We constructed the dynamic laterality time series by extracting the dynamic laterality index (DLI) at each sliding window. We then used k-means clustering to partition the laterality states and the Arenas-Fernandez-Gomez (AFG) community detection algorithm to determine the number of states. We characterized subjects' laterality characteristics using the mean laterality index (MLI) and laterality fluctuation (LF). Compared with NC, in all windows and state 1, BD patients showed higher MLI in the attention network (AN) of the right hemisphere, and AN in the left hemisphere showed more frequent laterality fluctuations. AN in the left hemisphere of BD patients showed higher MLI in all windows and state 3 compared to NC. In addition, in the AN of the right hemisphere in state 1, higher MLI in BD patients was significantly associated with patient symptoms. Our study provides new insights into the understanding of BD neuropathology in terms of brain dynamic laterality.