arXiv subjects

Michael Rosenblum

Publications and source records attributed to Michael Rosenblum.

At least 19 recordsLinked to original sources

arXiv2026

Cluster dynamics in a two-group Stuart-Landau model analyzed by the second-order phase reduction

We analyze cluster states in an ensemble of Stuart-Landau oscillators with two subpopulations of different frequencies. Our main goal is to compare the descriptions of the system's dynamics obtained via the standard first-order phase approximation and the second-order phase reduction. We demonstrate that the second-order model not only provides quantitative improvements in the description but also reveals new dynamical states not present in the standard Kuramoto theory. In particular, it describes bistability of synchronous states in the minimal setup of two coupled oscillators and the existence of three-cluster states, forbidden in the first-order phase description by the Watanabe-Strogatz theory. The very good agreement between the second-order approximation results and the results of numerical simulations of the original Stuart-Landau network highlights the usefulness of high-order phase-reduction models.

nlin.AO↗

arXiv2026

Improving Variance Estimation for Covariate Adjustment with Binary Outcomes

Covariate adjustment is a general method for improving precision when estimating treatment effects in randomized trials and is recommended by the FDA in its 2023 guidance when baseline variables are prognostic for the primary outcome. We focus on a method highlighted in that guidance called ``standardization" (or ``g-computation") for estimating the marginal treatment effect. We address the question of how to reliably estimate variance for binary outcomes when marginal outcome probabilities are close to 0 or 1. We propose an influence function-based leave-one-out cross-validated (IF-LOO) variance estimator for the standardized difference-in-means average treatment effect. Through simulation studies, we show that this estimator provides appropriate type-I error control and performs reliably in challenging settings where existing methods can yield inflated type-I error or fail entirely, such as when outcome events are rare or sample sizes are small. In addition to having desirable statistical properties, we derive a closed-form expression for the proposed estimator, enabling straightforward and reliable implementation by study statisticians. The robust finite-sample performance and ease of implementation suggest the IF-LOO variance estimator is a prudent default choice for standardization in clinical trials.

stat.ME↗

arXiv2025

Inferring collective synchrony observing spiking of one or several neurons

We tackle a quantification of synchrony in a large ensemble of interacting neurons from the observation of spiking events. In a simulation study, we efficiently infer the synchrony level in a neuronal population from a point process reflecting spiking of a small number of units and even from a single neuron. We introduce a synchrony measure (order parameter) based on the Bartlett covariance density; this quantity can be easily computed from the recorded point process. This measure is robust concerning missed spikes and, if computed from observing several neurons, does not require spike sorting. We illustrate the approach by modeling populations of spiking or bursting neurons, including the case of sparse synchrony.

q-bio.NC↗

arXiv2024

Phase reduction explains chimera shape: when multi-body interaction matters

We present an extension of the Kuramoto-Sakaguchi model for networks, deriving the second-order phase approximation for a paradigmatic model of oscillatory networks - an ensemble of non-identical Stuart-Landau oscillators coupled pairwisely via an arbitrary coupling matrix. We explicitly demonstrate how this matrix translates into the coupling structure in the phase equations. To illustrate the power of our approach and the crucial importance of high-order phase reduction, we tackle a trendy setup of non-locally coupled oscillators exhibiting a chimera state. We reveal that our second-order phase model reproduces the dependence of the chimera shape on the coupling strength that is not captured by the typically used first-order Kuramoto-like model. Our derivation contributes to a better understanding of complex networks' dynamics, establishing a relation between the coupling matrix and multi-body interaction terms in the high-order phase model.

nlin.AO↗

arXiv2024

Methodological Problems in Every Black-Box Study of Forensic Firearm Comparisons

Reviews conducted by the National Academy of Sciences (2009) and the President's Council of Advisors on Science and Technology (2016) concluded that the field of forensic firearm comparisons has not been demonstrated to be scientifically valid. Scientific validity requires adequately designed studies of firearm examiner performance in terms of accuracy, repeatability, and reproducibility. Researchers have performed ``black-box'' studies with the goal of estimating these performance measures. As statisticians with expertise in experimental design, we conducted a literature search of such studies to date and then evaluated the design and statistical analysis methods used in each study. Our conclusion is that all studies in our literature search have methodological flaws that are so grave that they render the studies invalid, that is, incapable of establishing scientific validity of the field of firearms examination. Notably, error rates among firearms examiners, both collectively and individually, remain unknown. Therefore, statements about the common origin of bullets or cartridge cases that are based on examination of ``individual" characteristics do not have a scientific basis. We provide some recommendations for the design and analysis of future studies.

stat.AP↗

arXiv2024

A unified quantification of synchrony in globally coupled populations with the Wiener order parameter

We tackle the quantification of synchrony in globally coupled populations. Furthermore, we treat the problem of incomplete observations when the population mean field is unavailable, but only a small subset of units is observed. We introduce a new order parameter and demonstrate its efficiency for quantifying synchrony via monitoring general observables, regardless of whether the oscillations can be characterized in terms of the phases. Under condition of a significant irregularity in the dynamics of the coupled units, this order parameter provides a unified description of synchrony in populations of units of various complexity. The main examples include noise-induced oscillations, coupled strongly chaotic systems, and noisy periodic oscillations. Furthermore, we explore how this parameter works for the standard Kuramoto model of coupled regular phase oscillators. The most significant advantage of our approach is its ability to infer and quantify synchrony from the observation of a small percentage of the units and even from a single unit, provided the observations are sufficiently long.

nlin.CD↗

arXiv2023

High-order phase reduction for coupled 2D oscillators

Phase reduction is a general approach to describe coupled oscillatory units in terms of their phases, assuming that the amplitudes are enslaved. For such a reduction, the coupling should be small, but one also expects the reduction to be valid for finite coupling. This paper presents a general framework allowing us to obtain coupling terms in higher orders of the coupling parameter for generic two-dimensional oscillators and arbitrary coupling terms. The theory is illustrated with an accurate prediction of Arnold's tongue for the van der Pol oscillator exploiting higher-order phase reduction.

nlin.CD↗

arXiv2023

Commentary on Guyll et al. (2023): Misuse of Statistical Method Results in Highly Biased Interpretation of Forensic Evidence

Since the National Academy of Sciences released their report outlining paths for improving reliability, standards, and policies in the forensic sciences NAS (2009), there has been heightened interest in evaluating and improving the scientific validity within forensic science disciplines. Guyll et al. (2023) seek to evaluate the validity of forensic cartridge-case comparisons. However, they make a serious statistical error that leads to highly inflated claims about the probability that a cartridge case from a crime scene was fired from a reference gun, typically a gun found in the possession of a defendant. It is urgent to address this error since these claims, which are generally biased against defendants, are being presented by the prosecution in an ongoing homicide case where the defendant faces the possibility of a lengthy prison sentence (DC Superior Court, 2023).

stat.AP↗

arXiv2023

Desynchronizing two oscillators while stimulating and observing only one

Synchronization of two or more self-sustained oscillators is a well-known and studied phenomenon, appearing both in natural and designed systems. In some cases, the synchronized state is undesired, and the aim is to destroy synchrony by external intervention. In this paper, we focus on desynchronizing two self-sustained oscillators by short pulses delivered to the system in a phase-specific manner. We analyze a non-trivial case when we cannot access both oscillators but stimulate only one. The following restriction is that we can monitor only one unit, be it a stimulated or non-stimulated one. First, we use a system of two coupled Rayleigh oscillators to demonstrate how a loss of synchrony can be induced by stimulating a unit once per period at a specific phase and detected by observing consecutive inter-pulse durations. Next, we exploit the phase approximation to develop a rigorous theory formulating the problem in terms of a map. We derive exact expressions for the phase -- isostable coordinates of this coupled system and show a relation between the phase and isostable response curves to the phase response curve of the uncoupled oscillator. Finally, we demonstrate how to obtain phase response information from the system using time series and discuss the differences between observing the stimulated and unstimulated oscillator.

nlin.CD↗

arXiv2022

Inferring oscillator's phase and amplitude response from a scalar signal exploiting test stimulation

The phase sensitivity curve or phase response curve (PRC) quantifies the oscillator's reaction to stimulation at a specific phase and is a primary characteristic of a self-sustained oscillatory unit. Knowledge of this curve yields a phase dynamics description of the oscillator for arbitrary weak forcing. Similar, though much less studied characteristic, is the amplitude response that can be defined either using an ad hoc approach to amplitude estimation or via the isostable variables. Here, we discuss the problem of the phase and amplitude response inference from observations using test stimulation. Although PRC determination for noise-free neuronal-like oscillators perturbed by narrow pulses is a well-known task, the general case remains a challenging problem. Even more challenging is the inference of the amplitude response. This characteristic is crucial, e.g., for controlling the amplitude of the collective mode in a network of interacting units -- a task relevant to neuroscience. Here, we compare the performance of different techniques suitable for inferring the phase and amplitude response, particularly with application to macroscopic oscillators. We suggest improvements to these techniques, e.g., demonstrating how to obtain the PRC in case of stimuli of arbitrary shape. Our main result is a novel technique denoted by IPID-1, based on the direct reconstruction of the Winfree equation and the analogous first-order equation for isostable dynamics. The technique works for signals with or without well-pronounced marker events and pulses of arbitrary shape; in particular, we consider charge-balanced pulses typical in neuroscience applications. Moreover, this technique is superior for noisy and high-dimensional systems. Additionally, we describe an error measure that can be computed solely from data and complements any inference technique.

nlin.AO↗

arXiv2022

Using Targeted Maximum Likelihood Estimation to Estimate Treatment Effect with Longitudinal Continuous or Binary Data: A Systematic Evaluation of 28 Diabetes Clinical Trials

The primary analysis of clinical trials in diabetes therapeutic area often involves a mixed-model repeated measure (MMRM) approach to estimate the average treatment effect for longitudinal continuous outcome, and a generalized linear mixed model (GLMM) approach for longitudinal binary outcome. In this paper, we considered another estimator of the average treatment effect, called targeted maximum likelihood estimator (TMLE). This estimator can be a one-step alternative to model either continuous or binary outcome. We compared those estimators by simulation studies and by analyzing real data from 28 diabetes clinical trials. The simulations involved different missing data scenarios, and the real data sets covered a wide range of possible distributions of the outcome and covariates in real-life clinical trials for diabetes drugs with different mechanisms of action. For all the settings, adjusted estimators tended to be more efficient than the unadjusted one. In the setting of longitudinal continuous outcome, the MMRM approach with visits and baseline variables interaction appeared to dominate the performance of the MMRM considering the main effects only for the baseline variables while showing better or comparable efficiency to the TMLE estimator in both simulations and data applications. For modeling longitudinal binary outcome, TMLE generally outperformed GLMM in terms of relative efficiency, and its avoidance of the cumbersome covariance fitting procedure from GLMM makes TMLE a more advantageous estimator.

stat.AP↗

arXiv2022

Combining Covariate Adjustment with Group Sequential, Information Adaptive Designs to Improve Randomized Trial Efficiency

In clinical trials, there is potential to improve precision and reduce the required sample size by appropriately adjusting for baseline variables in the statistical analysis. This is called covariate adjustment. Despite recommendations by regulatory agencies in favor of covariate adjustment, it remains underutilized leading to inefficient trials. We address two obstacles that make it challenging to use covariate adjustment. A first obstacle is the incompatibility of many covariate adjusted estimators with commonly used boundaries in group sequential designs (GSDs). A second obstacle is the uncertainty at the design stage about how much precision gain will result from covariate adjustment. We propose a method that modifies the original estimator so that it becomes compatible with GSDs, while increasing or leaving unchanged the estimator's precision. Our approach allows the use of any asymptotically linear estimator, which covers many estimators used in randomized trials. Building on this, we propose using an information adaptive design, that is, continuing the trial until the required information level is achieved. Such a design adapts to the amount of precision gain and can lead to faster, more efficient trials, without sacrificing validity or power. We evaluate estimator performance in simulations that mimic features of a completed stroke trial.

stat.ME↗

arXiv2021

Optimizing Precision and Power by Machine Learning in Randomized Trials, with an Application to COVID-19

The rapid finding of effective therapeutics requires the efficient use of available resources in clinical trials. The use of covariate adjustment can yield statistical estimates with improved precision, resulting in a reduction in the number of participants required to draw futility or efficacy conclusions. We focus on time-to-event and ordinal outcomes. A key question for covariate adjustment in randomized studies is how to fit a model relating the outcome and the baseline covariates to maximize precision. We present a novel theoretical result establishing conditions for asymptotic normality of a variety of covariate-adjusted estimators that rely on machine learning (e.g., l1-regularization, Random Forests, XGBoost, and Multivariate Adaptive Regression Splines), under the assumption that outcome data is missing completely at random. We further present a consistent estimator of the asymptotic variance. Importantly, the conditions do not require the machine learning methods to converge to the true outcome distribution conditional on baseline variables, as long as they converge to some (possibly incorrect) limit. We conducted a simulation study to evaluate the performance of the aforementioned prediction methods in COVID-19 trials using longitudinal data from over 1,500 patients hospitalized with COVID-19 at Weill Cornell Medicine New York Presbyterian Hospital. We found that using l1-regularization led to estimators and corresponding hypothesis tests that control type 1 error and are more precise than an unadjusted estimator across all sample sizes tested. We also show that when covariates are not prognostic of the outcome, l1-regularization remains as precise as the unadjusted estimator, even at small sample sizes (n = 100). We give an R package adjrct that performs model-robust covariate adjustment for ordinal and time-to-event outcomes.

stat.ME↗

arXiv2021

Optimizing charge-balanced pulse stimulation for desynchronization

Collective synchronization in a large population of self-sustained units appears both in natural and engineered systems. Sometimes this effect is in demand, while in some cases, it is undesirable, which calls for control techniques. In this paper, we concentrate on pulsatile control, with the goal to either increase or decrease the level of synchrony. We quantify this level by the entropy of the phase distribution. Motivated by possible applications in neuroscience, we consider pulses of a realistic shape. Exploiting the noisy Kuramoto-Winfree model, we search for the optimal pulse profile and the optimal stimulation phase. For this purpose, we derive an expression for the change of the phase distribution entropy due to the stimulus. We relate this change to the properties of individual units characterized by generally different natural frequencies and phase response curves and the population's state. We verify the general result by analyzing a two-frequency population model and demonstrating a good agreement of the theory and numerical simulations.

nlin.CD↗

arXiv2021

Two Mechanisms of Remote Synchronization in a Chain of Stuart-Landau Oscillators

Remote synchronization implies that oscillators interacting not directly but via an additional unit (hub) adjust their frequencies and exhibit frequency locking while the hub remains asynchronous. In this paper, we analyze the mechanisms of remote synchrony in a small network of three coupled Stuart-Landau oscillators using recent results on high-order phase reduction. We analytically demonstrate the role of two factors promoting remote synchrony. These factors are the non-isochronicity of oscillators and the coupling terms appearing in the second-order phase approximation. We show a good correspondence between our theory and numerical results for small and moderate coupling strengths.

nlin.CD↗

arXiv2021

Real-time estimation of phase and amplitude with application to neural data

Computation of the instantaneous phase and amplitude via the Hilbert Transform is a powerful tool of data analysis. This approach finds many applications in various science and engineering branches but is not proper for causal estimation because it requires knowledge of the signal's past and future. However, several problems require real-time estimation of phase and amplitude; an illustrative example is phase-locked or amplitude-dependent stimulation in neuroscience. In this paper, we discuss and compare three causal algorithms that do not rely on the Hilbert Transform but exploit well-known physical phenomena, the synchronization and the resonance. After testing the algorithms on a synthetic data set, we illustrate their performance computing phase and amplitude for the accelerometer tremor measurements and a Parkinsonian patient's beta-band brain activity.

q-bio.NC↗

arXiv2020

The Impact of Time Series Length and Discretization on Longitudinal Causal Estimation Methods

The use of observational time series data to assess the impact of multi-time point interventions is becoming increasingly common as more health and activity data are collected and digitized via wearables, social media, and electronic health records. Such time series may involve hundreds or thousands of irregularly sampled observations. One common analysis approach is to simplify such time series by first discretizing them into sequences before applying a discrete-time estimation method that adjusts for time-dependent confounding. In certain settings, this discretization results in sequences with many time points; however, the empirical properties of longitudinal causal estimators have not been systematically compared on long sequences. We compare three representative longitudinal causal estimation methods on simulated and real clinical data. Our simulations and analyses assume a Markov structure and that longitudinal treatments/exposures are binary-valued and have at most a single jump point. We identify sources of bias that arise from temporally discretizing the data and provide practical guidance for discretizing data and choosing between methods when working with long sequences. Additionally, we compare these estimators on real electronic health record data, evaluating the impact of early treatment for patients with a life-threatening complication of infection called sepsis.

stat.ME↗

arXiv2020

Model-Robust Inference for Clinical Trials that Improve Precision by Stratified Randomization and Covariate Adjustment

Two commonly used methods for improving precision and power in clinical trials are stratified randomization and covariate adjustment. However, many trials do not fully capitalize on the combined precision gains from these two methods, which can lead to wasted resources in terms of sample size and trial duration. We derive consistency and asymptotic normality of model-robust estimators that combine these two methods, and show that these estimators can lead to substantial gains in precision and power. Our theorems cover a class of estimators that handle continuous, binary, and time-to-event outcomes; missing outcomes under the missing at random assumption are handled as well. For each estimator, we give a formula for a consistent variance estimator that is model-robust and that fully captures variance reductions from stratified randomization and covariate adjustment. Also, we give the first proof (to the best of our knowledge) of consistency and asymptotic normality of the Kaplan-Meier estimator under stratified randomization, and we derive its asymptotic variance. The above results also hold for the biased-coin covariate-adaptive design. We demonstrate our results using three completed, phase 3, randomized trial data sets of treatments for substance use disorder, where the variance reduction due to stratified randomization and covariate adjustment ranges from 1% to 36%.

stat.ME↗