Biology subjects

Itzkovitch, A.

Publications and source records attributed to Itzkovitch, A..

5 recordsLinked to original sources

bioRxiv2025

From Big Data to Small Scales: Machine Learning Enhances Microclimate Model Predictions

1. Microclimates are critical for understanding how organisms interact with their environments, influencing behaviour, physiology, and species distributions. However, traditional physical heat-balance models for predicting ground temperatures in microhabitats often exhibit biases due to unaccounted environmental complexities and poorly constrained parameters. These limitations can hinder ecological research and conservation planning, particularly in the context of climate change. 2. In this study, we demonstrate how high-resolution drone-based mapping and machine learning can improve the accuracy of microclimate models. Using drone imagery, we generated detailed environmental maps, including solar radiation, vegetation indices, and skyview factors, to parameterize a physical heat-balance model. Validation with thermal maps derived from drone-mounted infrared cameras revealed systematic errors in the physical models predictions, including over- and underestimations under specific environmental conditions. To address these errors, we applied a random forest machine learning model to predict and correct biases in new prediction maps. 3. Our results show that machine learning reduced mean absolute errors by over 30% and mean square errors by 50%, while consistently narrowing the range of prediction inaccuracies. Key factors driving biases, such as vegetation cover, solar radiation, and height above ground, were identified, offering valuable insights for improving physical models. The machine learning corrections not only improved accuracy but also highlighted parameters and processes that were previously underrepresented or oversimplified in traditional models. 4. These findings illustrate the potential of combining machine learning with physical modelling to enhance microclimate predictions. This approach provides ecologists and conservation practitioners with a powerful tool to generate accurate, fine-scale microclimate maps, enabling better understanding of species responses to climate change and informing climate-resilient habitat management and conservation strategies.

ecology↗

bioRxiv2025

Mechanisms of long term non-reinforced preference change: functional connectivity changes in a longitudinal functional MRI study

Behavioral change studies mostly focus on external reinforcements to modify preferences. Cue-approach training (CAT) is a paradigm that influences preferences by the mere association of stimuli, sensory cues, and a rapid motor response, without external reinforcements. The behavioral effect has been shown to last for months after less than one hour of training. Here, we used a modified version of CAT by changing the neutral-cue to a number that represented a monetary amount of reward that the participants accumulate (i.e. incentive-cue). After a single training session, we compared behavioral performance and functional connectivity (FC), as measured using functional magnetic resonance imaging, between two groups, one receiving a neutral-cue and the other receiving an incentive-cue, at 5 time points across one year. We replicated the maintenance of behavioral changes after 6-months for the non-reinforced neutral-cue participants, but not for the reinforced group. The reinforced training group showed higher FC within the limbic system, whereas the non-externally reinforced group showed higher functional connectivity within and between default-mode and dorsal-attention networks. Our findings offer putative neural correlates for both reinforced and non-reinforced preference changes that are maintained over time and which could be implemented in future behavioral change interventions. Significance StatementThe current work examines the neural mechanisms of non-externally reinforced preference change and its maintenance over time, using both a neutral cue and a modified version of the cue-approach training paradigm. While both groups initially exhibited preference shifts, only the non-reinforced group maintained preference changes over one year, suggesting an enduring internal reinforcement mechanism. We identified distinct patterns in functional connectivity related to behavioral maintenance, find that non-external reinforcement increases the connectivity of the default mode network while external reinforcement elevates connectivity between limbic areas. These findings enhance our understanding of sustainable behavior change and advocate for non-external reinforcement in behavioral interventions.

neuroscience↗

bioRxiv2025

Assessing Garmin Stress Level Score Against Heart Rate Variability Measurements

Daily stressors elicit physiological and mental responses impacting health, cognition, and behavior. Accurately assessing stress responses in natural settings remains challenging despite extensive research, though wrist-worn devices have the potential to address this gap through remote data collection. The Garmin fitness tracker provides a stress score largely based on HRV which must be validated prior to use in research. This study aimed to assess the stress score given by the Garmin Vivosmart 4 against HR and HRV from ECG recordings derived by the Polar H10 chest strap. A pilot study of 29 participants was conducted, followed by power calculations and preregistration of the main study which included 60 participants. Data were collected simultaneously from both devices during a laboratory session of restful and mental-stress-inducing tasks. Garmins stress score, mean HR, SD2/SD1, and HF power exhibited significant differences between stress and rest conditions. Moreover, Garmins stress score correlated significantly with HR, RMSSD, and SD2/SD1. Our findings suggest that physiological responses to mental stress were influenced by sex and tonic HRV. The study suggests that the GSS is indicative of mental stress, with its accessibility and noninvasive nature promising widespread utilization in various research domains.

physiology↗

bioRxiv2022

A passive computational marker for individual differences in non-reinforced learning

Cue-Approach Training (CAT) is a paradigm that enhances preferences without external reinforcmeents, suggesting a potential role for internal learning processes. Here, we developed a novel Bayesian computational model to quantify anticipatory response patterns during the training phase of CAT. This phase includes individual items and thus this marker is potentially of internal learning signals at the item level. Our model, fitted to meta-analysis data from 29 prior CAT experiments, was able to predict individual differences in non-reinforced preference changes using a key computational marker. Crucially, two new experiments manipulated the training procedure to influence the models predicted learning marker. As predicted and preregistered, the manipulation successfully induced differential preference changes, supporting a causal role of our model. These findings demonstrate powerful potential of our computational framework for investigating intrinsic learning processes. This framework could be used to predict preference changes and opens new avenues for understanding intrinsic motivation and decision-making. TeaserBayesian modeling of response time predicts individual differences in non reinforced preference change.

neuroscience↗

bioRxiv2020

A probabilistic model for indel evolution: differentiating insertions from deletions

Insertions and deletions (indels) are common molecular evolutionary events. However, probabilistic models for indel evolution are under-developed due to their computational complexity. Here we introduce several improvements to indel modeling: (1) while previous models for indel evolution assumed that the rates and length distributions of insertions and deletions are equal, here, we propose a richer model that explicitly distinguishes between the two; (2) We introduce numerous summary statistics that allow Approximate Bayesian Computation (ABC) based parameter estimation; (3) We develop a neural-network model-selection scheme to test whether the richer model better fits biological data compared to the simpler model. Our analyses suggest that both our inference scheme and the model-selection procedure achieve high accuracy on simulated data. We further demonstrate that our proposed indel model better fits a large number of empirical datasets and that, for the majority of these datasets, the deletion rate is higher than the insertion rate. Finally, we demonstrate that indel rates are negatively correlated to the effective population size across various phylogenomic clades.

bioinformatics↗