Łukasz Kidziński is a research associate in the Neuromuscular Biomechanics Lab at Stanford, applying state-of-the-art computer vision and reinforcement learning algorithms for broadening our understanding of human movement and performance. Previously he was a researcher in the CHILI group, Computer-Human Interaction in Learning and Instruction, at the EPFL in Switzerland, where he was developing methods for measuring and improving engagement of users in massive online open courses. He obtained a Ph.D. degree at Université Libre de Bruxelles in mathematical statistics, working on frequency domain methods for time series of functional data.

Academic Appointments

Honors & Awards

  • Data Science Fellow, Stanford Data Science Initiative (2018)

Professional Education

  • Master of Science, University of Warsaw, Computer science (2009)

All Publications

  • Automated Classification of Radiographic Knee Osteoarthritis Severity Using Deep Neural Networks. Radiology. Artificial intelligence Thomas, K. A., Kidzinski, L., Halilaj, E., Fleming, S. L., Venkataraman, G. R., Oei, E. H., Gold, G. E., Delp, S. L. 2020; 2 (2): e190065


    Purpose: To develop an automated model for staging knee osteoarthritis severity from radiographs and to compare its performance to that of musculoskeletal radiologists.Materials and Methods: Radiographs from the Osteoarthritis Initiative staged by a radiologist committee using the Kellgren-Lawrence (KL) system were used. Before using the images as input to a convolutional neural network model, they were standardized and augmented automatically. The model was trained with 32116 images, tuned with 4074 images, evaluated with a 4090-image test set, and compared to two individual radiologists using a 50-image test subset. Saliency maps were generated to reveal features used by the model to determine KL grades.Results: With committee scores used as ground truth, the model had an average F1 score of 0.70 and an accuracy of 0.71 for the full test set. For the 50-image subset, the best individual radiologist had an average F1 score of 0.60 and an accuracy of 0.60; the model had an average F1 score of 0.64 and an accuracy of 0.66. Cohen weighted kappa between the committee and model was 0.86, comparable to intraexpert repeatability. Saliency maps identified sites of osteophyte formation as influential to predictions.Conclusion: An end-to-end interpretable model that takes full radiographs as input and predicts KL scores with state-of-the-art accuracy, performs as well as musculoskeletal radiologists, and does not require manual image preprocessing was developed. Saliency maps suggest the model's predictions were based on clinically relevant information. Supplemental material is available for this article. © RSNA, 2020.

    View details for DOI 10.1148/ryai.2020190065

    View details for PubMedID 32280948

  • Deep neural networks enable quantitative movement analysis using single-camera videos. Nature communications Kidziński, Ł., Yang, B., Hicks, J. L., Rajagopal, A., Delp, S. L., Schwartz, M. H. 2020; 11 (1): 4054


    Many neurological and musculoskeletal diseases impair movement, which limits people's function and social participation. Quantitative assessment of motion is critical to medical decision-making but is currently possible only with expensive motion capture systems and highly trained personnel. Here, we present a method for predicting clinically relevant motion parameters from an ordinary video of a patient. Our machine learning models predict parameters include walking speed (r = 0.73), cadence (r = 0.79), knee flexion angle at maximum extension (r = 0.83), and Gait Deviation Index (GDI), a comprehensive metric of gait impairment (r = 0.75). These correlation values approach the theoretical limits for accuracy imposed by natural variability in these metrics within our patient population. Our methods for quantifying gait pathology with commodity cameras increase access to quantitative motion analysis in clinics and at home and enable researchers to conduct large-scale studies of neurological and musculoskeletal disorders.

    View details for DOI 10.1038/s41467-020-17807-z

    View details for PubMedID 32792511

  • The turning and barrier course reveals gait parameters for detecting freezing of gait and measuring the efficacy of deep brain stimulation. PloS one O'Day, J., Syrkin-Nikolau, J., Anidi, C., Kidzinski, L., Delp, S., Bronte-Stewart, H. 2020; 15 (4): e0231984


    Freezing of gait (FOG) is a devastating motor symptom of Parkinson's disease that leads to falls, reduced mobility, and decreased quality of life. Reliably eliciting FOG has been difficult in the clinical setting, which has limited discovery of pathophysiology and/or documentation of the efficacy of treatments, such as different frequencies of subthalamic deep brain stimulation (STN DBS). In this study we validated an instrumented gait task, the turning and barrier course (TBC), with the international standard FOG questionnaire question 3 (FOG-Q3, r = 0.74, p < 0.001). The TBC is easily assembled and mimics real-life environments that elicit FOG. People with Parkinson's disease who experience FOG (freezers) spent more time freezing during the TBC compared to during forward walking (p = 0.007). Freezers also exhibited greater arrhythmicity during non-freezing gait when performing the TBC compared to forward walking (p = 0.006); this difference in gait arrhythmicity between tasks was not detected in non-freezers or controls. Freezers' non-freezing gait was more arrhythmic than that of non-freezers or controls during all walking tasks (p < 0.05). A logistic regression model determined that a combination of gait arrhythmicity, stride time, shank angular range, and asymmetry had the greatest probability of classifying a step as FOG (area under receiver operating characteristic curve = 0.754). Freezers' percent time freezing and non-freezing gait arrhythmicity decreased, and their shank angular velocity increased in the TBC during both 60 Hz and 140 Hz STN DBS (p < 0.05) to non-freezer values. The TBC is a standardized tool for eliciting FOG and demonstrating the efficacy of 60 Hz and 140 Hz STN DBS for gait impairment and FOG. The TBC revealed gait parameters that differentiated freezers from non-freezers and best predicted FOG; these may serve as relevant control variables for closed loop neurostimulation for FOG in Parkinson's disease.

    View details for DOI 10.1371/journal.pone.0231984

    View details for PubMedID 32348346

  • Pre-operative gastrocnemius lengths in gait predict outcomes following gastrocnemius lengthening surgery in children with cerebral palsy. PloS one Rajagopal, A., Kidzinski, L., McGlaughlin, A. S., Hicks, J. L., Delp, S. L., Schwartz, M. H. 2020; 15 (6): e0233706


    Equinus deformity is one of the most common gait deformities in children with cerebral palsy. We examined whether estimates of gastrocnemius length in gait could identify limbs likely to have short-term and long-term improvements in ankle kinematics following gastrocnemius lengthening surgery to correct equinus. We retrospectively analyzed data of 891 limbs that underwent a single-event multi-level surgery (SEMLS), and categorized outcomes based on the normalcy of ankle kinematics. Limbs with short gastrocnemius lengths that received a gastrocnemius lengthening surgery as part of a SEMLS (case limbs) were 2.2 times more likely than overtreated limbs (i.e., limbs who did not have short lengths, but still received a lengthening surgery) to have a good surgical outcome at the follow-up gait visit (good outcome rate of 71% vs. 33%). Case limbs were 1.2 times more likely than control limbs (i.e., limbs that had short gastrocnemius lengths but no lengthening surgery) to have a good outcome (71% vs. 59%). Three-fourths of the case limbs with a good outcome at the follow-up gait visit maintained this outcome over time, compared to only one-half of the overtreated limbs. Our results caution against over-prescription of gastrocnemius lengthening surgery and suggest gastrocnemius lengths can be used to identify good surgical candidates.

    View details for DOI 10.1371/journal.pone.0233706

    View details for PubMedID 32502157

  • Reply: Limitations in the creation of an automatic diagnosis tool for dysgraphia NPJ DIGITAL MEDICINE Asselborn, T., Gargot, T., Kidzinski, L., Johal, W., Cohen, D., Jolly, C., Dillenbourg, P. 2019; 2
  • Reply: Limitations in the creation of an automatic diagnosis tool for dysgraphia. NPJ digital medicine Asselborn, T., Gargot, T., Kidziński, Ł., Johal, W., Cohen, D., Jolly, C., Dillenbourg, P. 2019; 2: 37

    View details for DOI 10.1038/s41746-019-0115-z

    View details for PubMedID 31304383

    View details for PubMedCentralID PMC6550145

  • Automatic real-time gait event detection in children using deep neural networks. PloS one Kidzinski, L., Delp, S., Schwartz, M. 2019; 14 (1): e0211466


    Annotation of foot-contact and foot-off events is the initial step in post-processing for most quantitative gait analysis workflows. If clean force plate strikes are present, the events can be automatically detected. Otherwise, annotation of gait events is performed manually, since reliable automatic tools are not available. Automatic annotation methods have been proposed for normal gait, but are usually based on heuristics of the coordinates and velocities of motion capture markers placed on the feet. These heuristics do not generalize to pathological gait due to greater variability in kinematics and anatomy of patients, as well as the presence of assistive devices. In this paper, we use a data-driven approach to predict foot-contact and foot-off events from kinematic and marker time series in children with normal and pathological gait. Through analysis of 9092 gait cycle measurements we build a predictive model using Long Short-Term Memory (LSTM) artificial neural networks. The best-performing model identifies foot-contact and foot-off events with an average error of 10 and 13 milliseconds respectively, outperforming popular heuristic-based approaches. We conclude that the accuracy of our approach is sufficient for most clinical and research applications in the pediatric population. Moreover, the LSTM architecture enables real-time predictions, enabling applications for real-time control of active assistive devices, orthoses, or prostheses. We provide the model, usage examples, and the training code in an open-source package.

    View details for PubMedID 30703141

  • Estimating the effect size of surgery to improve walking in children with cerebral palsy from retrospective observational clinical data. Scientific reports Rajagopal, A., Kidzinski, L., McGlaughlin, A. S., Hicks, J. L., Delp, S. L., Schwartz, M. H. 2018; 8 (1): 16344


    Single-event multilevel surgery (SEMLS) is a standard treatment approach aimed at improving gait for patients with cerebral palsy, but the effect of this approach compared to natural progression without surgical intervention is unclear. In this study, we used retrospective patient history, physical exam, and three-dimensional gait analysis data from 2,333 limbs to build regression models estimating the effect of SEMLS on gait, while controlling for expected natural progression. Post-hoc classifications using the regression model results identified which limbs would exhibit gait within two standard deviations of typical gait at the follow-up visit with or without a SEMLS with 73% and 77% accuracy, respectively. Using these models, we found that, while surgery was expected to have a positive effect on 93% of limbs compared to natural progression, in only 37% of limbs was this expected effect a clinically meaningful improvement. We identified 26% of the non-surgically treated limbs that may have shown a clinically meaningful improvement in gait had they received surgery. Our models suggest that pre-operative physical therapy focused on improving biomechanical characteristics, such as walking speed and strength, may improve likelihood of positive surgical outcomes. These models are shared with the community to use as an evaluation tool when considering whether or not a patient should undergo a SEMLS.

    View details for PubMedID 30397268

  • Automated human-level diagnosis of dysgraphia using a consumer tablet NPJ DIGITAL MEDICINE Asselborn, T., Gargot, T., Kidzinski, L., Johal, W., Cohen, D., Jolly, C., Dillenbourg, P. 2018; 1
  • Principal Components Analysis of Periodically Correlated Functional Time Series JOURNAL OF TIME SERIES ANALYSIS Kidzinski, L., Kokoszka, P., Jouzdani, N. 2018; 39 (4): 502–22

    View details for DOI 10.1111/jtsa.12283

    View details for Web of Science ID 000434361400003

  • Gene expression profiling of low-grade endometrial stromal sarcoma indicates fusion protein-mediated activation of the Wnt signaling pathway GYNECOLOGIC ONCOLOGY Przybyl, J., Kidzinski, L., Hastie, T., Debiec-Rychter, M., Nusse, R., van de Rijn, M. 2018; 149 (2): 388–93


    Low-grade endometrial stromal sarcomas (LGESS) harbor chromosomal translocations that affect proteins associated with chromatin remodeling Polycomb Repressive Complex 2 (PRC2), including SUZ12, PHF1 and EPC1. Roughly half of LGESS also demonstrate nuclear accumulation of β-catenin, which is a hallmark of Wnt signaling activation. However, the targets affected by the fusion proteins and the role of Wnt signaling in the pathogenesis of these tumors remain largely unknown.Here we report the results of a meta-analysis of three independent gene expression profiling studies on LGESS and immunohistochemical evaluation of nuclear expression of β-catenin and Lef1 in 112 uterine sarcoma specimens obtained from 20 LGESS and 89 LMS patients.Our results demonstrate that 143 out of 310 genes overexpressed in LGESS are known to be directly regulated by SUZ12. In addition, our gene expression meta-analysis shows activation of multiple genes implicated in Wnt signaling. We further emphasize the role of the Wnt signaling pathway by demonstrating concordant nuclear expression of β-catenin and Lef1 in 7/16 LGESS.Based on our findings, we suggest that LGESS-specific fusion proteins disrupt the repressive function of the PRC2 complex similar to the mechanism seen in synovial sarcoma, where the SS18-SSX fusion proteins disrupt the mSWI/SNF (BAF) chromatin remodeling complex. We propose that these fusion proteins in LGESS contribute to overexpression of Wnt ligands with subsequent activation of Wnt signaling pathway and formation of an active β-catenin/Lef1 transcriptional complex. These observations could lead to novel therapeutic approaches that focus on the Wnt pathway in LGESS.

    View details for PubMedID 29544705

  • Multimodal teaching analytics: Automated extraction of orchestration graphs from wearable sensor data JOURNAL OF COMPUTER ASSISTED LEARNING Prieto, L. P., Sharma, K., Kidzinski, L., Rodriguez-Triana, M. J., Dillenbourg, P. 2018; 34 (2): 193–203


    The pedagogical modelling of everyday classroom practice is an interesting kind of evidence, both for educational research and teachers' own professional development. This paper explores the usage of wearable sensors and machine learning techniques to automatically extract orchestration graphs (teaching activities and their social plane over time), on a dataset of 12 classroom sessions enacted by two different teachers in different classroom settings. The dataset included mobile eye-tracking as well as audiovisual and accelerometry data from sensors worn by the teacher. We evaluated both time-independent and time-aware models, achieving median F1 scores of about 0.7-0.8 on leave-one-session-out k-fold cross-validation. Although these results show the feasibility of this approach, they also highlight the need for larger datasets, recorded in a wider variety of classroom settings, to provide automated tagging of classroom practice that can be used in everyday practice across multiple teachers.

    View details for PubMedID 29686446

    View details for PubMedCentralID PMC5909982

  • Orchestration Load Indicators and Patterns: n-the-Wild Studies Using Mobile Eye-Tracking IEEE TRANSACTIONS ON LEARNING TECHNOLOGIES Prieto, L. P., Sharma, K., Kidzinski, L., Dillenbourg, P. 2018; 11 (2): 216–29
  • Automated human-level diagnosis of dysgraphia using a consumer tablet. NPJ digital medicine Asselborn, T., Gargot, T., Kidziński, Ł., Johal, W., Cohen, D., Jolly, C., Dillenbourg, P. 2018; 1: 42


    The academic and behavioral progress of children is associated with the timely development of reading and writing skills. Dysgraphia, characterized as a handwriting learning disability, is usually associated with dyslexia, developmental coordination disorder (dyspraxia), or attention deficit disorder, which are all neuro-developmental disorders. Dysgraphia can seriously impair children in their everyday life and require therapeutic care. Early detection of handwriting difficulties is, therefore, of great importance in pediatrics. Since the beginning of the 20th century, numerous handwriting scales have been developed to assess the quality of handwriting. However, these tests usually involve an expert investigating visually sentences written by a subject on paper, and, therefore, they are subjective, expensive, and scale poorly. Moreover, they ignore potentially important characteristics of motor control such as writing dynamics, pen pressure, or pen tilt. However, with the increasing availability of digital tablets, features to measure these ignored characteristics are now potentially available at scale and very low cost. In this work, we developed a diagnostic tool requiring only a commodity tablet. To this end, we modeled data of 298 children, including 56 with dysgraphia. Children performed the BHK test on a digital tablet covered with a sheet of paper. We extracted 53 handwriting features describing various aspects of handwriting, and used the Random Forest classifier to diagnose dysgraphia. Our method achieved 96.6% sensibility and 99.2% specificity. Given the intra-rater and inter-rater levels of agreement in the BHK test, our technique has comparable accuracy for experts and can be deployed directly as a diagnostics tool.

    View details for DOI 10.1038/s41746-018-0049-x

    View details for PubMedID 31304322

    View details for PubMedCentralID PMC6550155

  • Smart Environments and Analytics on Video-Based Learning Giannakos, M. N., Sampson, D. G., Kidzinski, L., Pardo, A., ACM ASSOC COMPUTING MACHINERY. 2016: 502–4
  • Semiautomatic Annotation of MOOC Forum Posts Liu, W., Kidzinski, L., Dillenbourg, P., Li, Y., Chang, M., Kravcik, M., Popescu, E., Huang, R., Kinshuk, Chen, N. S. SPRINGER-VERLAG SINGAPORE PTE LTD. 2016: 399–408
  • A Tutorial on Machine Learning in Educational Science Kidzinski, L., Giannakos, M., Sampson, D. G., Dillenbourg, P., Li, Y., Chang, M., Kravcik, M., Popescu, E., Huang, R., Kinshuk, Chen, N. S. SPRINGER-VERLAG SINGAPORE PTE LTD. 2016: 453–59
  • Estimation in Functional Lagged Regression JOURNAL OF TIME SERIES ANALYSIS Hormann, S., Kidzinski, L., Kokoszka, P. 2015; 36 (4): 541–61

    View details for DOI 10.1111/jtsa.12114

    View details for Web of Science ID 000355723900003

  • A Note on Estimation in Hilbertian Linear Models SCANDINAVIAN JOURNAL OF STATISTICS Hoermann, S., Kidzinski, L. 2015; 42 (1): 43–62

    View details for DOI 10.1111/sjos.12094

    View details for Web of Science ID 000349982500003

  • Dynamic functional principal components JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY Hormann, S., Kidzinski, L., Hallin, M. 2015; 77 (2): 319–48

    View details for DOI 10.1111/rssb.12076

    View details for Web of Science ID 000349206900001

  • Augmenting Collaborative MOOC Video Viewing with Synchronized Textbook Li, N., Kidzinski, L., Dillenbourg, P., Abascal, J., Barbosa, S., Fetter, M., Gross, T., Palanque, P., Winckler, M. SPRINGER-VERLAG BERLIN. 2015: 81–88