About

Benjamin Smarr is an assistant professor in the department of Bioengineering at UCSD, holding a 50% appointment at the Halicioglu Data Science Institute. His research focuses on accelerating positive social change through the use of abundant human time series data. Dr. Smarr mentors a diverse group of students and researchers working on projects related to health, data science, and biomedical research. His lab emphasizes the application of data science to real-world health issues, including women's health and mental illness detection. Through collaborations and leadership, he supports efforts that integrate wearable technology, clinical records, and new trials to advance biomedical research and healthspan.

Research topics

• Computer Science
• Medicine
• Machine Learning
• Artificial Intelligence
• Pathology
• Biology
• Psychiatry
• Internal medicine
• Psychology
• Developmental psychology
• Data science
• Engineering
• Embedded system

Selected publications

• Power Consumption Patterns Using Telemetry Data

  Open MIND · 2026-02-25

  preprint

  This paper examines the analysis of package power consumption using Intel's telemetry data. It challenges the prevailing belief that hardware choice is the primary determinant of a device's power consumption and instead emphasizes the significant role of user behavior. The paper includes two sections: Exploratory Data Analysis (EDA) and a linear model for power consumption. The EDA section provides valuable insights from Intel's telemetry data, comparing power consumption across countries, with a specific focus on power consumption patterns in the US and China. Our simple linear model affirms those patterns and highlight the possible importance of user behavior and its influence on power consumption. Ultimately, the paper underscores the need to understand power consumption patterns and identifies areas where stakeholders like Intel can make improvements to reduce environmental impact effectively and efficiently.

  DOI

• Testing the effectiveness of an intervention that aligns circadian rhythm with daily activities on student flourishing

  Open MIND · 2026-01-01

  otherOpen access

  Previous research has examined whether tracking circadian rhythm improves well-being, with some evidence suggesting that aligning daily activities with one’s natural chronotype (an individual’s natural preference for being alert and asleep), enhances well-being and productivity. However, there remains a need for more randomized, intervention-based studies that test behavioral alignment tools, and evaluate their effects on various aspects of well-being. Here, we aim to investigate whether an active intervention, involving the use of a circadian rhythm-tracking app that provides personalized recommendations over 5-6 weeks, improves students’ flourishing. We hypothesize that students randomly assigned to the intervention condition (using the app) will show greater improvements in well-being compared to those in the control condition.

  PublisherDOI

• Age-related patterns in distal skin temperature during naps

  SLEEP · 2026-03-17

  articleOpen access

  STUDY OBJECTIVES: Naps are common worldwide; despite their prevalence, physiological changes during naps are less well-characterized than during nighttime sleep. We aimed to characterize napping patterns and their associated physiological changes across age groups. METHODS: We used longitudinal wearable device data from 20,027 individuals from the TemPredict Study to assess (1) napping frequency and timing and (2) distal skin temperature changes around naps. RESULTS: Older age groups napped more frequently and consistently throughout the week. Individuals aged 30-49 years showed the greatest increase in nap frequency from weekdays to weekends, whereas younger (18-19 years) and older (>60 years) age groups showed the smallest increases. Sleep timing and distal skin temperature rhythms appeared tightly coupled, even during the daytime. Both sleep timing and distal skin temperature rhythms appeared phase-delayed among younger age groups relative to older age groups. Distal skin temperature was higher during naps on weekends relative to weekdays. Older individuals (>65 years) had lower distal skin temperatures during naps relative to younger individuals (≤30 years) (weekends: p<.0001; δ=-0.20; weekdays: p<.0001; δ=-0.19). Longer naps were generally associated with greater distal skin temperature before napping, especially among younger individuals (≤30 years). CONCLUSION: We observed age-related differences in napping patterns and distal skin temperature around and during naps. Future research should examine whether such distal skin temperature changes around naps relate to sleep pressure and age-related disease risk.

  PublisherDOI

• Dynamical model of aperiodic locomotor activity effects on mouse core body temperature removes transient perturbations from longitudinal temperature signals

  Scientific Reports · 2026-01-06 · 1 citations

  articleOpen accessSenior author

  Mammalian temperature changes across time due to multiple endogenous and exogenous factors including circadian rhythms, hormonal changes, and locomotor activity. These multiple factors make it difficult to disentangle each of their effects to understand their independent contributions. This is especially problematic due to the relatively high-amplitude, aperiodic heating effects of locomotor activity on core body temperature. These heating effects, combined with innate cooling effects back to core body temperature steady state, mean that locomotor activity can contribute apparent power to both circadian and ultradian rhythms in observed temperature data. We propose that the effect from locomotor activity to core body temperature is not simply the linear addition of circadian and ultradian oscillations, but rather a heating effect that can be offset by a cooling effect dependent on core temperature displacement from resting temperature. Since these effects appear to contribute power to independent rhythms in spectral analysis, in this work we develop an interpretable, parsimonious mathematical model of murine core body temperature that removes them in the time-domain. The model only depends on the initial observed core body temperature as well as minute-level locomotor activity data, making it robust to aperiodic mouse activity. We show that coefficients obtained after fitting the model to each mouse return physiologically relevant differences between sexes, as well as reflect directional changes within female mice between their non-estrous and estrous temperature data. We believe this work should be of use to researchers interested in how core body temperature dynamics change in response to experimental interventions, especially if locomotor activity may be affected as well.

  PublisherOA PDFDOI

• Sleep and temperature data from wearable devices support noninvasive detection of diabetes mellitus in a large-scale, retrospective analysis

  Communications Medicine · 2026-03-16

  articleOpen accessSenior author

  BACKGROUND: Diabetes Mellitus is a common, chronic metabolic disorder affecting the cardiovascular system, autonomic nervous system, and sleep quality. Diabetes affects diverse physiological data including heart rate variability, distal body temperature, and sleep duration. We hypothesized that biologically informed features from wearable device data, combined with appropriate application of longitudinal data, can capture physiological covariates of diabetes and support the noninvasive detection of diabetes. METHODS: We obtained 4 months and 7 days of wearables data (Oura Ring) from 389 individuals self-reporting diabetes and 10,820 people self-reporting no diabetes diagnosis from the TemPredict database. We selected 36 features of sleep, circadian disruption, and distal body temperature from literature and evaluated whether time windows of these features could be classified to be from individuals self-reporting diabetes (N = 236) or self-reporting no diabetes diagnosis (N = 282). RESULTS: Here we show longer time windows of input perform better, with the best algorithm (21-nights) achieving 0.88 Area under ROC (AUROC) and 0.80 Area under Precision Recall (AUPRC) (0.30 improvement over random). Feature analyses reveal the importance of further derived distal body temperature features (increase AUROC by 0.0724), especially to differentiate other chronic conditions from diabetes. The model achieves 0.80 AUROC and 0.28 improvement over random in AUPRC in an imbalanced cohort drawn from 6,658 individuals, emulating a general population. CONCLUSIONS: These results indicate the value of biologically informed features and longitudinal data for identifying people with diabetes and further, suggest that these methods could make such separations possible for other chronic conditions that affect sleep and inflammation.

  PublisherDOI

• Power Consumption Patterns Using Telemetry Data

  arXiv (Cornell University) · 2026-02-25

  articleOpen access

  This paper examines the analysis of package power consumption using Intel's telemetry data. It challenges the prevailing belief that hardware choice is the primary determinant of a device's power consumption and instead emphasizes the significant role of user behavior. The paper includes two sections: Exploratory Data Analysis (EDA) and a linear model for power consumption. The EDA section provides valuable insights from Intel's telemetry data, comparing power consumption across countries, with a specific focus on power consumption patterns in the US and China. Our simple linear model affirms those patterns and highlight the possible importance of user behavior and its influence on power consumption. Ultimately, the paper underscores the need to understand power consumption patterns and identifies areas where stakeholders like Intel can make improvements to reduce environmental impact effectively and efficiently.

  PublisherOA PDF

• Wearable-derived skin temperature dynamics during sleep reveal cardiovascular perfusion deficits through mechanistic modeling

  npj Digital Medicine · 2026-04-22

  articleOpen accessSenior author

  Classical statistics are commonly used to find differences between distributions of average skin temperature across populations. However, skin temperature is affected by many endogenous (within body) and exogenous (outside body) factors, and these factors induce causal changes in longitudinal skin temperature that can obfuscate the interpretation of average population differences. Moreover, interpretations are increasingly difficult to make when using temperature signals sampled longitudinally in uncontrolled settings. A potential way to better handle the inherent complexity of skin temperature dynamics in uncontrolled settings is to explicitly account for the effects of causal factors on the short- and long-term trajectories of temperature. In this work, we find that a physics-informed model of skin temperature and activity during sleep accounts for significantly more variance than an equally parsimonious linear model. Furthermore, this model enables separation of cohorts with cardiovascular conditions that are known to affect skin thermoregulation, an important improvement over classic statistical modeling.

  PublisherOA PDFDOI

• Augmenting Circadian Biology Research With Data Science

  Journal of Biological Rhythms · 2025-01-29 · 2 citations

  reviewOpen accessSenior authorCorresponding

  The nature of biological research is changing, driven by the emergence of big data, and new computational models to parse out the information therein. Traditional methods remain the core of biological research but are increasingly either augmented or sometimes replaced by emerging data science tools. This presents a profound opportunity for those circadian researchers interested in incorporating big data and related analyses into their plans. Here, we discuss the emergence of novel sources of big data that could be used to gain real-world insights into circadian biology. We further discuss technical considerations for the biologist interested in including data science approaches in their research. We conversely discuss the biological considerations for data scientists so that they can more easily identify the nuggets of biological rhythms insight that might too easily be lost through application of standard data science approaches done without an appreciation of the way biological rhythms shape the variance of complex data objects. Our hope is that this review will make bridging disciplines in both directions (biology to computational and vice versa) easier. There has never been such rapid growth of cheap, accessible, real-world research opportunities in biology as now; collaborations between biological experts and skilled data scientists have the potential to mine out new insights with transformative impact.

  PublisherDOI

• Chronobiologically-informed features from CGM data provide unique information for XGBoost prediction of longer-term glycemic dysregulation in 8,000 individuals with type-2 diabetes

  PLOS Digital Health · 2025-04-09 · 4 citations

  articleOpen accessSenior author

  Type 2 Diabetes causes dysregulation of blood glucose, which leads to long-term, multi-tissue damage. Continuous glucose monitoring devices are commercially available and used to track glucose at high temporal resolution so that individuals can make informed decisions about their metabolic health. Algorithms processing these continuous data have also been developed that can predict glycemic excursion in the near future. These data might also support prediction of glycemic stability over longer time horizons. In this work, we leverage longitudinal Dexcom continuous glucose monitoring data to test the hypothesis that additional information about glycemic stability comes from chronobiologically-informed features. We develop a computationally efficient multi-timescale complexity index, and find that inclusion of time-of-day complexity features increases the performance of an out-of-the-box XGBoost model in predicting the change in glucose across days. These findings support the use of chronobiologically-inspired and explainable features to improve glucose prediction algorithms with relatively long time-horizons.

  PublisherDOI

• Multiscale Average Absolute Difference (MSAAD): A Computationally Efficient and Nonparametric Adaptation of Line Length for Noisy, Uncontrolled Wearables Time Series

  Algorithms · 2025-09-12

  articleOpen accessSenior author

  With the rise in physiological data sampled from wearable devices, efficient methods must be developed to encode temporal information for the comparison of time series arising from uncontrolled monitoring. We present a fast, nonparametric method called Multiscale Average Absolute Difference (MSAAD) to extract multiscale temporal features from wearable device data for purposes ranging from statistical analysis to machine learning inference. MSAAD outperforms comparable algorithms like multiscale sample entropy (MSSE) and multiscale Katz Fractal Dimension (MS-KFD) in terms of calculation stability on short realizations and faster runtime. MSAAD outperforms MSSE and MS-KFD by being able to separate diabetic and non-diabetic cohorts with moderate and large effect sizes in both sexes. Furthermore, it is capable of capturing “critical slowing down” in the temperature dynamics of aging populations, a phenomenon that has been previously observed in controlled settings. We propose that MSAAD is a scalable, interpretable time series feature that is capable of identifying meaningful differences in physiological time series data without making assumptions regarding underlying process models. MSAAD could improve the ability to derive insight from time series data mining for health applications.

  PublisherOA PDFDOI

Recent grants

• NIH Grant F32HD081957

  NIH · $172k · 2017

• Understanding the impact of environmental disruption in biological timing systems through signal processing.

  NIH · $203k · 2017–2019

Frequent coauthors

• Ashley E. Mason

  University of California, San Francisco

  32 shared

• Stephan Dilchert

  Baruch College

  31 shared

• Janet D. Robishaw

  Florida Atlantic University

  17 shared

• Joanne B. Krasnoff

  University of California, San Francisco

  17 shared

• Steven K Shiba

  16 shared

• C. Abigail Temple

  Baruch College

  16 shared

• Subhasis Dasgupta

  15 shared

• Frederick Hecht

  University of California, San Francisco

  15 shared

Labs

• UCSD Smarr labPI