Nadia Beyzaei1, Seraph Bao1, Yanyun Bu2, Linus Hung3, Hebah Hussaina1, Khaola Safia Maher1, Melvin Chan1, Heinrich Garn4, Gerhard Kloesch5, Bernhard Kohn6, Boris Kuzeljevic7, Scout McWilliams1, Karen Spruyt8, Emmanuel Tse1, Hendrik F Machiel Van der Loos9, Calvin Kuo10, Osman S Ipsiroglu11. 1. H-Behaviours Research Lab, BC Children's Hospital Research Institute, Vancouver, BC, Canada. 2. H-Behaviours Research Lab, BC Children's Hospital Research Institute, Vancouver, BC, Canada; Department of Mechanical Engineering, Faculty of Applied Science, University of British Columbia, Vancouver, BC, Canada. 3. H-Behaviours Research Lab, BC Children's Hospital Research Institute, Vancouver, BC, Canada; School of Kinesiology, Faculty of Education and Department of Computer Science, Faculty of Science, University of British Columbia, Vancouver, BC, Canada. 4. Austrian Institute of Technology, Austria. 5. Department of Neurology, Medical University of Vienna, Vienna, Austria. 6. School of Kinesiology, Faculty of Education and Department of Computer Science, Faculty of Science, University of British Columbia, Vancouver, BC, Canada. 7. Clinical Research Support Unit, BC Children' Hospital Research Institute, Vancouver, BC, Canada. 8. Institut National de la Santé et de la Recherche Médicale (INSERM), Paris, France. 9. Department of Mechanical Engineering, Faculty of Applied Science, University of British Columbia, Vancouver, BC, Canada. 10. School of Kinesiology, Faculty of Education and Department of Computer Science, Faculty of Science, University of British Columbia, Vancouver, BC, Canada. Electronic address: calvin.kuo@ubc.ca. 11. H-Behaviours Research Lab, BC Children's Hospital Research Institute, Vancouver, BC, Canada; Department of Pediatrics, Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada. Electronic address: oipsiroglu@bcchr.ca.
Abstract
BACKGROUND: Behavioral observations support clinical in-depth phenotyping but phenotyping and pattern recognition are affected by training background. As Attention Deficit Hyperactivity Disorder, Restless Legs syndrome/Willis Ekbom disease and medication induced activation syndromes (including increased irritability and/or akathisia), present with hyperactive-behaviors with hyper-arousability and/or hypermotor-restlessness (H-behaviors), we first developed a non-interpretative, neutral pictogram-guided phenotyping language (PG-PL) for describing body-segment movements during sitting. METHODOLOGY & RESULTS: The PG-PL was applied for annotating 12 1-min sitting-videos (inter-observer agreements >85%->97%) and these manual annotations were used as a ground truth to develop an automated algorithm using OpenPose, which locates skeletal landmarks in 2D video. We evaluated the algorithm's performance against the ground truth by computing the area under the receiver operator curve (>0.79 for the legs, arms, and feet, but 0.65 for the head). While our pixel displacement algorithm performed well for the legs, arms, and feet, it predicted head motion less well, indicating the need for further investigations. CONCLUSION: This first automated analysis algorithm allows to start the discussion about distinct phenotypical characteristics of H-behaviors during structured behavioral observations and may support differential diagnostic considerations via in-depth phenotyping of sitting behaviors and, in consequence, of better treatment concepts.
BACKGROUND: Behavioral observations support clinical in-depth phenotyping but phenotyping and pattern recognition are affected by training background. As Attention Deficit Hyperactivity Disorder, Restless Legs syndrome/Willis Ekbom disease and medication induced activation syndromes (including increased irritability and/or akathisia), present with hyperactive-behaviors with hyper-arousability and/or hypermotor-restlessness (H-behaviors), we first developed a non-interpretative, neutral pictogram-guided phenotyping language (PG-PL) for describing body-segment movements during sitting. METHODOLOGY & RESULTS: The PG-PL was applied for annotating 12 1-min sitting-videos (inter-observer agreements >85%->97%) and these manual annotations were used as a ground truth to develop an automated algorithm using OpenPose, which locates skeletal landmarks in 2D video. We evaluated the algorithm's performance against the ground truth by computing the area under the receiver operator curve (>0.79 for the legs, arms, and feet, but 0.65 for the head). While our pixel displacement algorithm performed well for the legs, arms, and feet, it predicted head motion less well, indicating the need for further investigations. CONCLUSION: This first automated analysis algorithm allows to start the discussion about distinct phenotypical characteristics of H-behaviors during structured behavioral observations and may support differential diagnostic considerations via in-depth phenotyping of sitting behaviors and, in consequence, of better treatment concepts.
Authors: Melvin Chan; Emmanuel K Tse; Seraph Bao; Mai Berger; Nadia Beyzaei; Mackenzie Campbell; Heinrich Garn; Hebah Hussaina; Gerhard Kloesch; Bernhard Kohn; Boris Kuzeljevic; Yi Jui Lee; Khaola Safia Maher; Natasha Carson; Jecika Jeyaratnam; Scout McWilliams; Karen Spruyt; Hendrik F Machiel Van der Loos; Calvin Kuo; Osman Ipsiroglu Journal: Data Brief Date: 2021-01-17