Samuel D Butensky1, Andrew P Sloan2, Eric Meyers3, Jason B Carmel4. 1. Burke Medical Research Institute, White Plains, NY, 10605, USA. Electronic address: samuel.butensky@gmail.com. 2. Texas Biomedical Center, The University of Texas at Dallas, Richardson, TX, 75080, USA; Erik Jonsson School of Engineering and Computer Science, The University of Texas at Dallas, Richardson, TX, 75080, USA. Electronic address: drew@vulintus.com. 3. Texas Biomedical Center, The University of Texas at Dallas, Richardson, TX, 75080, USA; Erik Jonsson School of Engineering and Computer Science, The University of Texas at Dallas, Richardson, TX, 75080, USA. Electronic address: ericmeyers55@gmail.com. 4. Burke Medical Research Institute, White Plains, NY, 10605, USA; Brain and Mind Research Institute, Weill Cornell Medical College, New York, NY, 10065, USA; Departments of Neurology and Pediatrics, Weill Cornell Medical College, New York, NY, USA. Electronic address: jason.carmel@med.cornell.edu.
Abstract
BACKGROUND: Hand function is critical for independence, and neurological injury often impairs dexterity. To measure hand function in people or forelimb function in animals, sensors are employed to quantify manipulation. These sensors make assessment easier and more quantitative and allow automation of these tasks. While automated tasks improve objectivity and throughput, they also produce large amounts of data that can be burdensome to analyze. We created software called Dexterity that simplifies data analysis of automated reaching tasks. NEW METHOD: Dexterity is MATLAB software that enables quick analysis of data from forelimb tasks. Through a graphical user interface, files are loaded and data are identified and analyzed. These data can be annotated or graphed directly. Analysis is saved, and the graph and corresponding data can be exported. For additional analysis, Dexterity provides access to custom scripts created by other users. RESULTS: To determine the utility of Dexterity, we performed a study to evaluate the effects of task difficulty on the degree of impairment after injury. Dexterity analyzed two months of data and allowed new users to annotate the experiment, visualize results, and save and export data easily. COMPARISON WITH EXISTING METHOD(S): Previous analysis of tasks was performed with custom data analysis, requiring expertise with analysis software. Dexterity made the tools required to analyze, visualize and annotate data easy to use by investigators without data science experience. CONCLUSIONS: Dexterity increases accessibility to automated tasks that measure dexterity by making analysis of large data intuitive, robust, and efficient.
BACKGROUND: Hand function is critical for independence, and neurological injury often impairs dexterity. To measure hand function in people or forelimb function in animals, sensors are employed to quantify manipulation. These sensors make assessment easier and more quantitative and allow automation of these tasks. While automated tasks improve objectivity and throughput, they also produce large amounts of data that can be burdensome to analyze. We created software called Dexterity that simplifies data analysis of automated reaching tasks. NEW METHOD: Dexterity is MATLAB software that enables quick analysis of data from forelimb tasks. Through a graphical user interface, files are loaded and data are identified and analyzed. These data can be annotated or graphed directly. Analysis is saved, and the graph and corresponding data can be exported. For additional analysis, Dexterity provides access to custom scripts created by other users. RESULTS: To determine the utility of Dexterity, we performed a study to evaluate the effects of task difficulty on the degree of impairment after injury. Dexterity analyzed two months of data and allowed new users to annotate the experiment, visualize results, and save and export data easily. COMPARISON WITH EXISTING METHOD(S): Previous analysis of tasks was performed with custom data analysis, requiring expertise with analysis software. Dexterity made the tools required to analyze, visualize and annotate data easy to use by investigators without data science experience. CONCLUSIONS: Dexterity increases accessibility to automated tasks that measure dexterity by making analysis of large data intuitive, robust, and efficient.
Authors: Lucia Friedli; Ephron S Rosenzweig; Quentin Barraud; Martin Schubert; Nadia Dominici; Lea Awai; Jessica L Nielson; Pavel Musienko; Yvette Nout-Lomas; Hui Zhong; Sharon Zdunowski; Roland R Roy; Sarah C Strand; Rubia van den Brand; Leif A Havton; Michael S Beattie; Jacqueline C Bresnahan; Erwan Bézard; Jocelyne Bloch; V Reggie Edgerton; Adam R Ferguson; Armin Curt; Mark H Tuszynski; Grégoire Courtine Journal: Sci Transl Med Date: 2015-08-26 Impact factor: 17.956
Authors: Anil Sindhurakar; Asht M Mishra; Disha Gupta; Jennifer F Iaci; Tom J Parry; Jason B Carmel Journal: Neurorehabil Neural Repair Date: 2017-01-20 Impact factor: 3.919
Authors: Seth A Hays; Navid Khodaparast; Andrew M Sloan; Daniel R Hulsey; Maritza Pantoja; Andrea D Ruiz; Michael P Kilgard; Robert L Rennaker Journal: J Neurosci Methods Date: 2012-11-23 Impact factor: 2.390
Authors: Eric Meyers; Anil Sindhurakar; Rachel Choi; Ruby Solorzano; Taylor Martinez; Andrew Sloan; Jason Carmel; Michael P Kilgard; Robert L Rennaker; Seth Hays Journal: J Neurosci Methods Date: 2016-03-11 Impact factor: 2.390
Authors: Rachel P Allred; DeAnna L Adkins; Martin T Woodlee; Lincoln C Husbands; Mónica A Maldonado; Jacqueline R Kane; Timothy Schallert; Theresa A Jones Journal: J Neurosci Methods Date: 2008-02-01 Impact factor: 2.390
Authors: Andrew M Sloan; Melyssa K Fink; Amber J Rodriguez; Adam M Lovitz; Navid Khodaparast; Robert L Rennaker; Seth A Hays Journal: PLoS One Date: 2015-10-27 Impact factor: 3.240