About

Dr. Michael C.K. Khoo is the Dean’s Professor of Biomedical Engineering with a courtesy joint appointment in Pediatrics at the University of Southern California. He received his undergraduate degree in Mechanical Engineering from Imperial College (University of London) in 1976 and his doctoral degree in Biomedical Engineering from Harvard University. Following postdoctoral work at the Brigham and Women’s Hospital in Boston, he joined USC faculty in 1983. Dr. Khoo has served as Department Chair of Biomedical Engineering and Co-Director of Education and Outreach in the NSF Engineering Research Center on Biomimetic Microelectronic Systems. His research interests lie at the intersection of bioengineering, cardiorespiratory physiology, and sleep medicine. He and his team develop computational models and employ novel experimental techniques to better understand and quantitatively characterize mechanisms responsible for abnormal regulation of autonomic and metabolic control, as well as sleep-wake states in chronic diseases and clinical syndromes such as sleep-related breathing disorders, hypertension, metabolic syndrome, and sickle cell disease. His work focuses on analyzing the complex dynamics of interactions among cardiovascular, respiratory, and sleep regulatory systems, utilizing physiology-based modeling and estimation techniques to interpret physiological data from noninvasive studies. Dr. Khoo’s Cardio-Respiratory Sleep Lab collaborates with physician-scientists and physiologists from institutions including Children’s Hospital of Los Angeles, USC Keck School of Medicine, Cincinnati Children’s Hospital Medical Center, Harvard Medical School, and UCLA. Recent research efforts include phenotyping the underlying mechanisms of sleep apnea in children with Down Syndrome through computational modeling and sleep measurements, as well as developing noninvasive methodologies for predicting vaso-occlusive crises in patients with sickle-cell disease. Dr. Khoo has been recognized as a Fellow of the IEEE, Biomedical Engineering Society, and the American Institute of Medical and Biological Engineering, and is a Fellow of the International Academy of Medical and Biological Engineering. He has served in leadership roles within the IEEE Engineering in Medicine and Biology Society, including Vice-President for Conferences and Chair of the 2012 EMBS International Conference. Currently, he is the Director of “SleepHuB,” a USC collaboratory addressing medical and societal issues in sleep health.

Research topics

• Internal medicine
• Medicine
• Cardiology
• Anesthesia
• Genetics
• Endocrinology
• Biology
• Surgery

Selected publications

• Wearable Device Photoplethysmography As a Viable Tool to Longitudinally Monitor Vasoconstriction Biomarkers for Predicting Vaso-Occlusive Crisis in Sickle Cell Disease: Feasibility and Validation Study

  JMIR Human Factors · 2025-12-01

  articleOpen access

  Background: Entrapment of sickled red blood cells in the microvasculature leads to sudden painful vaso-occlusive crises (VOCs) in sickle cell disease (SCD). This is potentially triggered by autonomic nervous system-mediated vasoconstriction in the microvasculature. Indeed, vasoconstriction biomarkers derived from a single night of laboratory-based fingertip photoplethysmography (PPG) recording were predictive of a higher frequency of future VOC in SCD. Noninvasive, remote, and longitudinal monitoring of autonomic vasoreactivity will facilitate the development of predictive biomarkers of imminent VOC. Objective: This study aimed to assess the feasibility and performance of a wearable wristband device to longitudinally monitor nocturnal peripheral autonomic vasoreactivity and to cross-validate the vasoconstriction parameters across the "gold-standard" finger sensor. Methods: A total of 12 patients with SCD and 6 healthy controls were recruited to wear a wristband device (Biostrap) with a PPG sensor on a nightly basis. For cross-validation studies, 50% (3/6) controls wore both the wristband and a sleep monitoring device (AliceNightOne) with a finger PPG sensor. We quantified autonomic vasoreactivity by processing PPG signals and deriving vasoconstriction parameters-magnitude of vasoconstriction (Mvasoc) and photoplethysmography amplitude coefficient of variation (PPGampCV). We performed a correlation analysis of the vasoconstriction parameters within each device to investigate whether Mvasoc and PPGampCV can be used as surrogate markers of vasoconstriction, and then cross-validated the PPGampCV across the wristband and finger PPG devices. Results: A total of 131 nocturnal PPG recordings were made with a wristband device (1-19 nights per participant; patients with SCD: n=79, 60%; controls: n=52, 40%). A total of 9 nocturnal recordings (3 nights per participant) were made with both wristband and finger sensor devices. Longitudinal continuous PPG recordings were feasible with the wearable device, with significant within-night and night-to-night variability in vasoconstriction parameters, suggesting dynamic changes in autonomic vasoreactivity. Mvasoc and PPGampCV significantly correlated within devices-the maximum overnight correlation was 0.82 (P<.001) for the finger sensor and 0.69 (P<.001) for the wristband sensor, suggesting that PPGampCV can serve as a surrogate for Mvasoc. Cross-validation analysis of PPGampCV across wristband and fingertip sensors showed statistically significant correlations on all 9 nights (overnight correlation coefficient ranging from 0.24-0.7), with some nightly segments of PPGampCV showing very strong correlation across devices. Conclusions: Wearable wristband devices are feasible tools for the collection of continuous PPG measurements and vasoconstriction parameters, which serve as objective markers of autonomic vasoreactivity in users with and without SCD. We have optimized the methods of quantifying vasoconstriction from wearable device PPG signals, and cross-validated them with standardized sensors. These findings enable large-scale, real-time monitoring of autonomic vasoreactivity along with pain outcomes for the development of vasoconstriction parameters as biomarkers imminent VOC in patients with SCD. This biomarker also has the potential to impact other diseases involving autonomic vascular dysregulation.

  PublisherDOI

• Cardiorespiratory Markers for Early Detection of Type 2 Diabetes: Machine Learning Models (Preprint)

  2025-08-08

  articleOpen accessSenior author

  <sec> <title>BACKGROUND</title> The global prevalence of type 2 diabetes mellitus (T2DM) poses significant challenges due to its association with increased cardiovascular risk and complications like cardiovascular autonomic neuropathy (CAN). Early identification of autonomic dysfunction in T2DM is important for timely interventions and improved clinical outcomes. </sec> <sec> <title>OBJECTIVE</title> This study investigates heart rate variability (HRV), frequency response function (FRF), and impulse response (IR) metrics as physiological markers for machine learning-based early prediction of T2DM-associated autonomic dysfunction. </sec> <sec> <title>METHODS</title> Using ECG and respiratory signals from two Physionet datasets, we derived complementary indices of cardiac autonomic nervous system function—HRV, FRF, and IR— for machine learning (ML) analysis. ML classifiers—logistic regression, linear SVM, and SVM with RBF kernel—assessed the predictive value of individual and combined feature sets under NearMiss-1 undersampling and SMOTE oversampling. While HRV derives autonomic metrics from ECG alone, FRF and IR utilize paired cardiorespiratory signals (ECG and respiratory signals), enabling modeling of frequency-domain (FRF) and causal time-domain (IR) interactions between cardiac and respiratory systems. This systems-based approach may capture subtle autonomic dysfunction in T2DM more effectively than HRV alone by reflecting integrated cardiorespiratory coupling. </sec> <sec> <title>RESULTS</title> IR metrics were the most informative standalone feature set, capturing causal cardiorespiratory interactions, achieving accuracy of 0.770 ± 0.179 (mean ± SD), precision of 0.783 ± 0.217, recall of 0.900 ± 0.224, and F1 score of 0.798 ± 0.140 with logistic regression and NearMiss-1. While HRV metrics were the least informative standalone feature set, the combined HRV + FRF feature set with NearMiss-1 achieved the highest performance, with accuracy of 0.830 ± 0.172, precision of 0.800 ± 0.183, recall of 0.933 ± 0.149, and F1 score of 0.853 ± 0.145 (SVM RBF). In SMOTE, the HRV + IR feature set performed best, yielding accuracy of 0.700 ± 0.128, precision of 0.783 ± 0.217, recall of 0.683 ± 0.207, and F1 score of 0.691 ± 0.097 with SVM RBF, surpassing standalone IR in most metrics, though IR alone retained superior recall (0.950 ± 0.112) and F1 score (0.708 ± 0.038). </sec> <sec> <title>CONCLUSIONS</title> As the strongest individual feature set, IR offered robust, interpretable prediction with lower feature complexity. By modeling causal cardiorespiratory interactions, IR outperformed HRV in detecting early autonomic dysfunction in T2DM. Combining HRV with IR or FRF, which captures frequency-specific cardiorespiratory coupling, further enhanced predictive performance, integrating complementary autonomic insights. These results emphasize systems-based IR metrics as objective markers for early T2DM-associated cardiovascular autonomic dysfunction detection and risk assessment, supporting their use in proactive diabetes management. By integrating physiologically relevant features into ML classification, this study advances noninvasive tools for early disease detection, personalized risk stratification, and targeted interventions. </sec>

  PublisherDOI

• Wearable Device Photoplethysmography is a Viable Tool to Longitudinally Monitor Vasoconstriction Biomarkers for Predicting Vaso-occlusive Crisis in Sickle Cell Disease (Preprint)

  2025-04-03

  preprintOpen access

  <sec> <title>BACKGROUND</title> Entrapment of sickled red blood cells in the microvasculature leads to sudden painful vaso-occlusive crises (VOC) in sickle cell disease (SCD). This is potentially triggered by autonomic nervous system mediated vasoconstriction in peripheral vasculature and resultant decrease in microvascular blood flow. Indeed, vasoconstriction biomarkers derived from a single night of laboratory-based fingertip photoplethysmography (PPG) recording were predictive of higher frequency of future VOC in SCD. Non-invasive, remote and longitudinal monitoring of autonomic vasoreactivity will facilitate development of predictive biomarkers of imminent VOC. </sec> <sec> <title>OBJECTIVE</title> To assess the feasibility and performance of wearable wristband device to longitudinally monitor nocturnal peripheral autonomic vasoreactivity and to cross-validate the vasoconstriction parameters across the “gold-standard” finger sensor. </sec> <sec> <title>METHODS</title> 11 subjects with SCD and 5 healthy controls were recruited to wear a wristband device (Biostrap) with PPG sensor on a nightly basis. For cross-validation studies, 3 subjects wore both the wristband and a sleep monitoring device (AliceNightOne) with finger PPG sensor. We quantified autonomic vasoreactivity by processing PPG signals and deriving vasoconstriction parameters - magnitude of vasoconstriction (Mvasoc) and Coefficient of Variation (PPGampCV). We performed correlation analysis of the vasoconstriction parameters within each device to investigate if Mvasoc and PPGamp CV can be surrogate markers of vasoconstriction and then cross-validated the PPGampCV across the wristband and finger PPG devices. </sec> <sec> <title>RESULTS</title> A total of 115 nocturnal PPG recordings were made with wristband device (1-19 nights/subject; SCD=70, controls=45). A total of 9 nocturnal recordings (3 nights/subject) were made with both wristband and finger sensor devices. Longitudinal continuous PPG recordings were feasible with wearable devices with subjects reporting ease of use. There was significant within night and night-to-night variability in vasoconstriction parameters suggesting dynamic changes in autonomic vasoreactivity. Mvasoc and PPGampCV significantly correlated within devices - the maximum overnight correlation was 0.82 (P&lt;.001) for finger sensor and 0.69 (P &lt;.001) for wristband sensor, suggesting that PPGampCV can serve as a surrogate for Mvasoc. Cross-validation analysis of PPGampCV across wristband and fingertip sensors showed statistically significant correlation across all nights (overnight correlation coefficient ranging from 0.24-0.7) with some nightly segments of PPGampCV showing near perfect correlation between devices. </sec> <sec> <title>CONCLUSIONS</title> Wearable wristband devices are feasible tools for collection of continuous PPG measurements and vasoconstriction parameters, which serve as objective markers of autonomic vasoreactivity in subjects with and without SCD. We have optimized the methods of quantifying vasoconstriction from wearable device PPG signals, while minimizing effects of artefacts and cross-validated them with standardized sensors. These findings enable large scale, real-time monitoring of autonomic vasoreactivity along with pain outcomes for the development of vasoconstriction parameters as a biomarker of imminent VOC in SCD. This biomarker has potential to impact not only SCD but also other diseases involving autonomic vascular dysregulation. </sec>

  PublisherDOI

• Beat-to-beat analysis of hemodynamic response to mental and psychological stress in sickle cell anemia

  Journal of sickle cell disease. · 2024-01-01 · 1 citations

  articleOpen access

  Objectives: Vaso-occlusive crises (VOCs) are a hallmark symptom of sickle cell disease (SCD). Physical stressors can trigger decreased microvascular blood flow and increase the risk for VOCs. However, the effect of mental and psychological stressors on vascular physiology in SCD is not well-established. We hereby examined fluctuations in continuous blood pressure (BP) to evaluate hemodynamic changes in SCD patients during mental and psychological stress. Methods: Thirteen SCD (HbSS) subjects from the Children's Hospital Los Angeles and 11 healthy (HbAA) volunteers were recruited. Continuous BP was recorded as subjects participated in two mental and one psychological stress tasks. Systolic beat-to-beat BP variability (BPV) measurements were calculated for each subject. Three very short-term BPV metrics served as outcome measures: standard deviation, coefficient of variation, and average real variability. Linear mixed effects models evaluated associations between patient factors and outcome measures. Results: SCD patients were associated with increased systolic BPV and exhibited a distinct increase in BPV in response to psychological stress. All subjects exhibited a decrease in systolic BPV in response to mental stress tasks. During mental stress, both groups displayed increased augmentation index, reflective of stress-induced vasoconstriction, while psychological stress in SCD patients led to both decreased mean arterial pressure and increased AI, suggestive of uncompensated vasoconstriction. Conclusion: These findings emphasize the impact of mental and psychological stressors on vascular function in SCD, the potential for monitoring physiological signals to predict VOC events, and the importance of counseling SCD patients on lifestyle practices to reduce their stress to prevent pain.

  PublisherOA PDFDOI

• Impact of opioid-free analgesia on pain severity and patient satisfaction after discharge from surgery: multispecialty, prospective cohort study in 25 countries

  British journal of surgery · 2024 · 12 citations

  • Medicine
  • Anesthesia
  • Surgery

  BACKGROUND: Balancing opioid stewardship and the need for adequate analgesia following discharge after surgery is challenging. This study aimed to compare the outcomes for patients discharged with opioid versus opioid-free analgesia after common surgical procedures. METHODS: This international, multicentre, prospective cohort study collected data from patients undergoing common acute and elective general surgical, urological, gynaecological, and orthopaedic procedures. The primary outcomes were patient-reported time in severe pain measured on a numerical analogue scale from 0 to 100% and patient-reported satisfaction with pain relief during the first week following discharge. Data were collected by in-hospital chart review and patient telephone interview 1 week after discharge. RESULTS: The study recruited 4273 patients from 144 centres in 25 countries; 1311 patients (30.7%) were prescribed opioid analgesia at discharge. Patients reported being in severe pain for 10 (i.q.r. 1-30)% of the first week after discharge and rated satisfaction with analgesia as 90 (i.q.r. 80-100) of 100. After adjustment for confounders, opioid analgesia on discharge was independently associated with increased pain severity (risk ratio 1.52, 95% c.i. 1.31 to 1.76; P < 0.001) and re-presentation to healthcare providers owing to side-effects of medication (OR 2.38, 95% c.i. 1.36 to 4.17; P = 0.004), but not with satisfaction with analgesia (β coefficient 0.92, 95% c.i. -1.52 to 3.36; P = 0.468) compared with opioid-free analgesia. Although opioid prescribing varied greatly between high-income and low- and middle-income countries, patient-reported outcomes did not. CONCLUSION: Opioid analgesia prescription on surgical discharge is associated with a higher risk of re-presentation owing to side-effects of medication and increased patient-reported pain, but not with changes in patient-reported satisfaction. Opioid-free discharge analgesia should be adopted routinely.

  DOI

• Nocturnal Autonomic and Vasoconstriction Parameters By Wearable Wrist Photoplethysmography Predict Intensity of Imminent Pain in Subjects with Sickle Cell Disease

  Blood · 2023-11-02 · 2 citations

  article

  Introduction: Sudden and unpredictable onset of pain is one of the biggest contributors to the morbidity of sickle cell disease (SCD), thought to occur due to obstruction of microvascular blood flow by rigid sickle-shaped red blood cells. Autonomic mediated vasoconstriction in the peripheral vasculature, with resulting decrease in tissue perfusion, is likely a triggering mechanism for the onset of vaso-occlusive pain, while the autonomic nervous system is also implicated in the pathogenesis of chronic neuropathic pain. Our analysis of 212 polysomnograms in SCD subjects showed that greater magnitude of nocturnal vasoconstriction (Mvasoc) is predictive of increased frequency of hospitalizations for vaso-occlusive pain crises (VOC) in the subsequent years (Chalacheva et al., Am J Hematol, 2020). In order to further develop this finding into a biomarker of more imminent pain, we monitored nocturnal photoplethysmography (PPG) at home with a wearable wristband device to measure vasoconstriction activity and heart rate variability, along with a daily electronic pain survey. Methods: 10 subjects with SCD were consented and underwent autonomic monitoring with a wristband device (Biostrap Inc) that uses an optical sensor to measure continuous PPG during sleep, with the raw PPG waveforms available from a cloud-based server. A unique biomarker (Mvasoc) was used to quantify the magnitude of vasoconstriction from the raw PPG signal, considering the frequency, magnitude, and duration of autonomic mediated spontaneous vasoconstrictions each night. Heart rate variability was derived from the pulse-to-pulse interval (PPI) of the PPG signal. Spectral indices of the PPI represent parasympathetic activity (high frequency power; HFP) and sympatho-vagal balance (Low to high Ratio; LHR = low frequency power(LFP)/HFP). Subjects concurrently documented daily pain incidence using an electronic text-based REDCap survey, with pain intensity measured on a scale of 1-10 if experiencing pain. Daily stress levels were also recorded on a scale of 1-10. Multivariate regression analysis was performed. Results: The subjects had a mean age of 25 years (range 14- 40 years) with an average hemoglobin of 8.8 g/dl . A total of 68 usable nocturnal PPG recordings were made (3-16 nights per subject). Therewere 25 incidents of pain on the day following autonomic monitoring with Biostrap, with the pain intensity ranging from 3-8. Multivariate analysis of nocturnal autonomic and vasoconstriction parameters revealed that a lower HFP ( p =0.0002) and a higher Mvasoc ( p=0.02) both preceded a higher intensity of pain the following day (Table 1). These results are consistent with a greater parasympathetic withdrawal and a higher magnitude of peripheral vasoconstriction predicting higher intensity of imminent pain. Age, gender, hemoglobin, stress levels and LHR were not predictive of pain the following day. Conclusion: Augmented nocturnal vasoconstriction and parasympathetic withdrawal noted prior to high intensity of pain imply that dysautonomia, with attendant changes in peripheral perfusion, has a significant role in pathogenesis of pain in SCD. Furthermore, autonomic vasoreactivity parameters were not significantly altered on the night after pain onset suggesting that the dysautonomia precedes pain. These results support the possibility of employing indices of nocturnal autonomic vasoreactivity as a predictive biomarker for imminent pain in SCD and provides a potential window for therapeutic intervention. This pilot study demonstrates the successful use of a wearable device PPG to remotely monitor nocturnal autonomic vasoreactivity and the ability to objectively quantify peripheral vasoconstriction responses. These findings will need to be applied in a larger cohort of SCD subjects with longer term monitoring for further validation and development of autonomic vasoreactivity as a predictive tool for pain in SCD.

  PublisherDOI

• Mcdaps: A Multi-Channel Physiological Signals Display and Analysis System for Clinical Researchers

  SSRN Electronic Journal · 2023-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Integrating Machine Learning with Biomedical Signal Processing and Systems Analysis: An Applications-based Course

  2023-07-24 · 2 citations

  articleSenior author

  The growing importance of data analytics in biomedicine is increasingly becoming recognized in biomedical engineering curricula through the introduction of machine learning classes that generally run in parallel to, but separately from, more traditional engineering courses, such as signal and systems analysis. We propose a new approach that systematically integrates signal processing and systems analysis with key techniques in machine learning. In the proposed course, the student obtains hands-on experience in applying algorithms that can be applied to practical problems of physiological signal conditioning, analysis and interpretation. This is achieved by exposing the student to a sequence of 4 applications-based modules that represent different biomedical engineering problems: human activity recognition from wearable devices, epileptic seizure detection, quantification of dynamic respiratory-cardiac coupling in humans under different conditions, and detection of sleep apnea episodes from heart rate variability data. Within each module, the student gains the experience of working with the data in question "from the ground up". We also introduce a general plan for assessment of student learning, and discuss the expected outcomes and limitations of this integrative approach to teaching.Clinical Relevance- The proposed course is targeted at biomedical engineering students at the senior undergraduate or first-year graduate level who are interested in learning how to analyze physiological signals. The course would also be suitable for clinician-scientists who have prior training in statistics with some exposure to engineering mathematics.

  PublisherDOI

• McDAPS: A multi-channel physiological signals display and analysis system for clinical researchers

  SoftwareX · 2023-07-01

  articleOpen accessSenior author

  We introduce McDAPS, an interactive software for assessing autonomic imbalance from non-invasive multi-channel physiological recordings. McDAPS provides a graphical user interface for data visualization, beat-to-beat processing and interactive analyses. The software extracts beat-to-beat RR interval systolic blood pressure, diastolic blood pressure, the pulse amplitude of photoplethysmogram and the pulse-to-pulse interval. The analysis modules include stationary and time-varying power spectral analyses, moving-correlation analysis and univariate analyses. Analyses can also be performed in batch mode if multiple datasets have to be processed in the same way. The program exports results in standard CSV format. McDAPS runs in MATLAB, and is supported on MS Windows and MAC OS systems. The MATLAB source code is available at https://github.com/thuptimd/McDAPS.git.

  PublisherDOI

• Functional near-infrared spectroscopy-based prefrontal cortex oxygenation during working memory tasks in sickle cell disease

  Neurophotonics · 2023-10-17 · 3 citations

  articleOpen accessSenior author

  SignificanceSickle cell disease (SCD), characterized by painful vaso-occlusive crises, is associated with cognitive decline. However, objective quantification of cognitive decline in SCD remains a challenge, and the associated hemodynamics are unknown.AimTo address this, we utilized functional near-infrared spectroscopy (fNIRS) to measure prefrontal cortex (PFC) oxygenation responses to N-back working memory tasks in SCD patients and compared them with healthy controls.ApproachWe quantified the PFC oxygenation rate as an index of cognitive activity in each group and compared them. In half of the participants, a Stroop test was administered before they started N-back to elevate their baseline stress level.ResultsIn SCD compared to healthy controls, we found that (1) under a high baseline stress level, there were significantly greater oxygenation responses during the 2-back task, further elevated with histories of stroke; (2) there was a marginally slower N-back response time, and it was even slower with a history of stroke; and (3) the task accuracy was not different.ConclusionsAdditional requirements for processing time, PFC resources, and PFC oxygenation in SCD patients offer an important basis for understanding their cognitive decline and highlight the potential of fNIRS for evaluating cognitive functions.

  PublisherOA PDFDOI

Recent grants

• NIH Grant R21HL090451

  NIH · $1.1M · 2011

• NIH Grant P41EB001978

  NIH · $15.3M · 2020

• Multimodal biophysical markers of vascular disease in hemoglobinopathies

  NIH · $9.6M · 2013–2019

• NIH Grant R01HL058725

  NIH · $558k · 2001

• NIH Grant R21HL076375

  NIH · $309k · 2007

Frequent coauthors

• Thomas D. Coates

  University of Southern California

  82 shared

• Richard B. Berry

  University of Florida

  47 shared

• Patjanaporn Chalacheva

  Carnegie Mellon University

  46 shared

• R. Kato

  Children's Hospital of Los Angeles

  45 shared

• Flavia Oliveira

  Universidade de Brasília

  42 shared

• John C. Wood

  42 shared

• Payal Shah

  42 shared

• Jon Detterich

  University of Southern California

  38 shared

Education

• Ph.D., Biomedical Engineering

  University of Southern California

  2000

• M.S., Biomedical Engineering

  University of Southern California

  1996

• B.S., Biomedical Engineering

  National University of Singapore

  1994

Awards & honors

• 2016 International Academy of Medical and Biological Enginee…
• 2012 IEEE Fellow, Engineering in Medicine & Biology Society
• 2010 IEEE Engineering in Medicine and Biology Society AdCom…
• 2005 Fellow, Biomedical Engineering Society