1
|
Smyth B, Lawlor A, Berndsen J, Feely C. Recommendations for marathon runners: on the application of recommender systems and machine learning to support recreational marathon runners. USER MODELING AND USER-ADAPTED INTERACTION 2021; 32:787-838. [PMID: 36452939 PMCID: PMC9701182 DOI: 10.1007/s11257-021-09299-3] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 06/19/2020] [Accepted: 07/22/2021] [Indexed: 06/17/2023]
Abstract
Every year millions of people, from all walks of life, spend months training to run a traditional marathon. For some it is about becoming fit enough to complete the gruelling 26.2 mile (42.2 km) distance. For others, it is about improving their fitness, to achieve a new personal-best finish-time. In this paper, we argue that the complexities of training for a marathon, combined with the availability of real-time activity data, provide a unique and worthwhile opportunity for machine learning and for recommender systems techniques to support runners as they train, race, and recover. We present a number of case studies-a mix of original research plus some recent results-to highlight what can be achieved using the type of activity data that is routinely collected by the current generation of mobile fitness apps, smart watches, and wearable sensors.
Collapse
Affiliation(s)
- Barry Smyth
- Insight SFI Centre for Data Analytics, University College Dublin, Dublin, Ireland
| | - Aonghus Lawlor
- Insight SFI Centre for Data Analytics, University College Dublin, Dublin, Ireland
| | - Jakim Berndsen
- Insight SFI Centre for Data Analytics, University College Dublin, Dublin, Ireland
| | - Ciara Feely
- Insight SFI Centre for Data Analytics, University College Dublin, Dublin, Ireland
| |
Collapse
|
2
|
O'Reilly MA, Slevin P, Ward T, Caulfield B. A Wearable Sensor-Based Exercise Biofeedback System: Mixed Methods Evaluation of Formulift. JMIR Mhealth Uhealth 2018; 6:e33. [PMID: 29386171 PMCID: PMC5812980 DOI: 10.2196/mhealth.8115] [Citation(s) in RCA: 13] [Impact Index Per Article: 2.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/30/2017] [Revised: 08/29/2017] [Accepted: 09/28/2017] [Indexed: 01/09/2023] Open
Abstract
Background Formulift is a newly developed mobile health (mHealth) app that connects to a single inertial measurement unit (IMU) worn on the left thigh. The IMU captures users’ movements as they exercise, and the app analyzes the data to count repetitions in real time and classify users’ exercise technique. The app also offers feedback and guidance to users on exercising safely and effectively. Objective The aim of this study was to assess the Formulift system with three different and realistic types of potential users (beginner gym-goers, experienced gym-goers, and qualified strength and conditioning [S&C] coaches) under a number of categories: (1) usability, (2) functionality, (3) the perceived impact of the system, and (4) the subjective quality of the system. It was also desired to discover suggestions for future improvements to the system. Methods A total of 15 healthy volunteers participated (12 males; 3 females; age: 23.8 years [SD 1.80]; height: 1.79 m [SD 0.07], body mass: 78.4 kg [SD 9.6]). Five participants were beginner gym-goers, 5 were experienced gym-goers, and 5 were qualified and practicing S&C coaches. IMU data were first collected from each participant to create individualized exercise classifiers for them. They then completed a number of nonexercise-related tasks with the app. Following this, a workout was completed using the system, involving squats, deadlifts, lunges, and single-leg squats. Participants were then interviewed about their user experience and completed the System Usability Scale (SUS) and the user version of the Mobile Application Rating Scale (uMARS). Thematic analysis was completed on all interview transcripts, and survey results were analyzed. Results Qualitative and quantitative analysis found the system has “good” to “excellent” usability. The system achieved a mean (SD) SUS usability score of 79.2 (8.8). Functionality was also deemed to be good, with many users reporting positively on the systems repetition counting, technique classification, and feedback. A number of bugs were found, and other suggested changes to the system were also made. The overall subjective quality of the app was good, with a median star rating of 4 out of 5 (interquartile range, IQR: 3-5). Participants also reported that the system would aid their technique, provide motivation, reassure them, and help them avoid injury. Conclusions This study demonstrated an overall positive evaluation of Formulift in the categories of usability, functionality, perceived impact, and subjective quality. Users also suggested a number of changes for future iterations of the system. These findings are the first of their kind and show great promise for wearable sensor-based exercise biofeedback systems.
Collapse
Affiliation(s)
- Martin Aidan O'Reilly
- Insight Centre for Data Analytics, University College Dublin, Dublin, Ireland.,School of Public Health, Physiotherapy and Sports Science, University College Dublin, Dublin, Ireland
| | - Patrick Slevin
- Insight Centre for Data Analytics, University College Dublin, Dublin, Ireland.,School of Public Health, Physiotherapy and Sports Science, University College Dublin, Dublin, Ireland
| | - Tomas Ward
- Insight Centre for Data Analytics, Dublin City University, Dublin, Ireland.,School of Computing, Dublin City University, Dublin, Ireland
| | - Brian Caulfield
- Insight Centre for Data Analytics, University College Dublin, Dublin, Ireland.,School of Public Health, Physiotherapy and Sports Science, University College Dublin, Dublin, Ireland
| |
Collapse
|
3
|
Lee BC, Thrasher TA, Fisher SP, Layne CS. The effects of different sensory augmentation on weight-shifting balance exercises in Parkinson's disease and healthy elderly people: a proof-of-concept study. J Neuroeng Rehabil 2015; 12:75. [PMID: 26329918 PMCID: PMC4557900 DOI: 10.1186/s12984-015-0064-y] [Citation(s) in RCA: 19] [Impact Index Per Article: 2.1] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/21/2015] [Accepted: 08/18/2015] [Indexed: 11/22/2022] Open
Abstract
Background Earlier versions of biofeedback systems for balance-related applications were intended primarily to provide “alarm” signals about body tilt rather than to guide rehabilitation exercise motion. Additionally, there have been few attempts to evaluate guidance modalities for balance rehabilitation exercises. The purpose of this proof-of-concept study is to evaluate the effects of guidance modalities during common dynamic weight-shifting exercises used in clinical settings. Methods A motion guidance system providing visual biofeedback, vibrotactile biofeedback, or both, was used during weight-shifting exercises. Eleven people with idiopathic Parkinson’s disease (PD) and nine healthy elderly people participated. Each participant wore a six-degree-of-freedom inertial measurement unit (IMU) located near the sacrum and four linear vibrating actuators (Tactors) attached to the skin over the front, back, and right and left sides of the abdomen. The IMU measured angular displacements and velocities of body tilt in anterior-posterior (A/P) and medial-lateral (M/L) directions. Participants were instructed to follow a slow moving target by shifting their weight in either the A/P or M/L direction up to 90 % of their limits of stability (LOS). Real-time position error was provided to participants in one of three sensory modalities: visual, vibrotactile, or both. Participants performed 5 trials for each biofeedback modality and movement direction (A/P and M/L) for a total of 30 trials in a random order. To characterize performance, position error was defined as the average absolute difference between the target and participant movements in degrees. Results Simultaneous delivery of visual and vibrotactile biofeedback resulted in significantly lower position error compared to either visual or vibrotactile biofeedback alone regardless of the movement direction for both participant cohorts. The pairwise comparisons were not significantly different between visual and vibrotactile biofeedback. Conclusion The study is the first attempt to assess the effects of guidance modalities on common balance rehabilitation exercises in people with PD and healthy elderly people. The results suggest that combined visual and vibrotactile biofeedback can improve volitional responses during postural tracking tasks. Index Terms – sensory augmentation, weight-shifting balance exercise, guidance modality, vibrotactile biofeedback, visual biofeedback, Parkinson’s disease.
Collapse
Affiliation(s)
- Beom-Chan Lee
- Department of Health and Human Performance, University of Houston, Houston, TX, USA. .,Center for Neuromotor and Biomechanics Research, University of Houston, Houston, TX, USA.
| | - Timothy A Thrasher
- Department of Health and Human Performance, University of Houston, Houston, TX, USA. .,Center for Neuromotor and Biomechanics Research, University of Houston, Houston, TX, USA.
| | - Stanley P Fisher
- Center for Neuromotor and Biomechanics Research, University of Houston, Houston, TX, USA. .,Movement Disorders and Neurorehabilitation Center, Houston Methodist Neurological Institute, Houston, TX, USA.
| | - Charles S Layne
- Department of Health and Human Performance, University of Houston, Houston, TX, USA. .,Center for Neuromotor and Biomechanics Research, University of Houston, Houston, TX, USA. .,Center for Neuro-Engineering and Cognitive Science, University of Houston, Houston, TX, USA.
| |
Collapse
|
4
|
Gatica-Rojas V, Méndez-Rebolledo G. Virtual reality interface devices in the reorganization of neural networks in the brain of patients with neurological diseases. Neural Regen Res 2014; 9:888-96. [PMID: 25206907 PMCID: PMC4146258 DOI: 10.4103/1673-5374.131612] [Citation(s) in RCA: 55] [Impact Index Per Article: 5.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 01/21/2014] [Indexed: 11/04/2022] Open
Abstract
Two key characteristics of all virtual reality applications are interaction and immersion. Systemic interaction is achieved through a variety of multisensory channels (hearing, sight, touch, and smell), permitting the user to interact with the virtual world in real time. Immersion is the degree to which a person can feel wrapped in the virtual world through a defined interface. Virtual reality interface devices such as the Nintendo® Wii and its peripheral nunchuks-balance board, head mounted displays and joystick allow interaction and immersion in unreal environments created from computer software. Virtual environments are highly interactive, generating great activation of visual, vestibular and proprioceptive systems during the execution of a video game. In addition, they are entertaining and safe for the user. Recently, incorporating therapeutic purposes in virtual reality interface devices has allowed them to be used for the rehabilitation of neurological patients, e.g., balance training in older adults and dynamic stability in healthy participants. The improvements observed in neurological diseases (chronic stroke and cerebral palsy) have been shown by changes in the reorganization of neural networks in patients' brain, along with better hand function and other skills, contributing to their quality of life. The data generated by such studies could substantially contribute to physical rehabilitation strategies.
Collapse
Affiliation(s)
- Valeska Gatica-Rojas
- Laboratory of Human Motor Control, Faculty of Health Sciences, University of Talca, Talca, Chile
| | - Guillermo Méndez-Rebolledo
- Laboratory of Human Motor Control, School of Kinesiology, Faculty of Health Sciences, University of Talca, Talca, Chile
| |
Collapse
|
5
|
van Diest M, Lamoth CJC, Stegenga J, Verkerke GJ, Postema K. Exergaming for balance training of elderly: state of the art and future developments. J Neuroeng Rehabil 2013; 10:101. [PMID: 24063521 PMCID: PMC3851268 DOI: 10.1186/1743-0003-10-101] [Citation(s) in RCA: 135] [Impact Index Per Article: 12.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/21/2012] [Accepted: 09/19/2013] [Indexed: 12/26/2022] Open
Abstract
Fall injuries are responsible for physical dysfunction, significant disability, and loss of independence among elderly. Poor postural control is one of the major risk factors for falling but can be trained in fall prevention programs. These however suffer from low therapy adherence, particularly if prevention is the goal. To provide a fun and motivating training environment for elderly, exercise games, or exergames, have been studied as balance training tools in the past years. The present paper reviews the effects of exergame training programs on postural control of elderly reported so far. Additionally we aim to provide an in-depth discussion of technologies and outcome measures utilized in exergame studies. Thirteen papers were included in the analysis. Most of the reviewed studies reported positive results with respect to improvements in balance ability after a training period, yet few reached significant levels. Outcome measures for quantification of postural control are under continuous dispute and no gold standard is present. Clinical measures used in the studies reviewed are well validated yet only give a global indication of balance ability. Instrumented measures were unable to detect small changes in balance ability as they are mainly based on calculating summary statistics, thereby ignoring the time-varying structure of the signals. Both methods only allow for measuring balance after the exergame intervention program. Current developments in sensor technology allow for accurate registration of movements and rapid analysis of signals. We propose to quantify the time-varying structure of postural control during gameplay using low-cost sensor systems. Continuous monitoring of balance ability leaves the user unaware of the measurements and allows for generating user-specific exergame training programs and feedback, both during one game and in timeframes of weeks or months. This approach is unique and unlocks the as of yet untapped potential of exergames as balance training tools for community dwelling elderly.
Collapse
Affiliation(s)
- Mike van Diest
- INCAS3, Dr, Nassaulaan 9, 9401 HJ Assen, The Netherlands.
| | | | | | | | | |
Collapse
|
6
|
Lee BC, Kim J, Chen S, Sienko KH. Cell phone based balance trainer. J Neuroeng Rehabil 2012; 9:10. [PMID: 22316167 PMCID: PMC3340298 DOI: 10.1186/1743-0003-9-10] [Citation(s) in RCA: 103] [Impact Index Per Article: 8.6] [Reference Citation Analysis] [Abstract] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/06/2011] [Accepted: 02/08/2012] [Indexed: 11/10/2022] Open
Abstract
BACKGROUND In their current laboratory-based form, existing vibrotactile sensory augmentation technologies that provide cues of body motion are impractical for home-based rehabilitation use due to their size, weight, complexity, calibration procedures, cost, and fragility. METHODS We have designed and developed a cell phone based vibrotactile feedback system for potential use in balance rehabilitation training in clinical and home environments. It comprises an iPhone with an embedded tri-axial linear accelerometer, custom software to estimate body tilt, a "tactor bud" accessory that plugs into the headphone jack to provide vibrotactile cues of body tilt, and a battery. Five young healthy subjects (24 ± 2.8 yrs, 3 females and 2 males) and four subjects with vestibular deficits (42.25 ± 13.5 yrs, 2 females and 2 males) participated in a proof-of-concept study to evaluate the effectiveness of the system. Healthy subjects used the system with eyes closed during Romberg, semi-tandem Romberg, and tandem Romberg stances. Subjects with vestibular deficits used the system with both eyes-open and eyes-closed conditions during semi-tandem Romberg stance. Vibrotactile feedback was provided when the subject exceeded either an anterior-posterior (A/P) or a medial-lateral (M/L) body tilt threshold. Subjects were instructed to move away from the vibration. RESULTS The system was capable of providing real-time vibrotactile cues that informed corrective postural responses. When feedback was available, both healthy subjects and those with vestibular deficits significantly reduced their A/P or M/L RMS sway (depending on the direction of feedback), had significantly smaller elliptical area fits to their sway trajectory, spent a significantly greater mean percentage time within the no feedback zone, and showed a significantly greater A/P or M/L mean power frequency. CONCLUSION The results suggest that the real-time feedback provided by this system can be used to reduce body sway. Its advantages over more complex laboratory-based and commercial balance training systems in terms of cost, size, weight, functionality, flexibility, and accessibility make it a good candidate for further home-based balance training evaluation.
Collapse
Affiliation(s)
- Beom-Chan Lee
- Department of Mechanical Engineering, University of Michigan, Ann Arbor, USA
| | | | | | | |
Collapse
|