Methods Inf Med 2015; 54(03): 248-255
DOI: 10.3414/ME14-01-0082
Original Articles
Schattauer GmbH

Daily Life Activity Routine Discovery in Hemiparetic Rehabilitation Patients Using Topic Models

J. Seiter
1   Wearable Computing Lab., ETH Zurich, Zurich, Switzerland
,
A. Derungs
2   ACTLab, University of Passau, Passau, Germany
,
C. Schuster-Amft
3   Research Department, Reha Rheinfelden, Rheinfelden, Switzerland
,
O. Amft
2   ACTLab, University of Passau, Passau, Germany
,
G. Tröster
1   Wearable Computing Lab., ETH Zurich, Zurich, Switzerland
› Institutsangaben
Weitere Informationen

Publikationsverlauf

received: 08. August 2014

accepted: 31. Februar 2014

Publikationsdatum:
22. Januar 2018 (online)

Summary

Background: Monitoring natural behavior and activity routines of hemiparetic rehabilitation patients across the day can provide valuable progress information for therapists and patients and contribute to an optimized rehabilitation process. In particular, continuous patient monitoring could add type, frequency and duration of daily life activity routines and hence complement standard clinical scores that are assessed for particular tasks only. Machine learning methods have been applied to infer activity routines from sensor data. However, supervised methods require activity annotations to build recognition models and thus require extensive patient supervision. Discovery methods, including topic models could provide patient routine information and deal with variability in activity and movement performance across patients. Topic models have been used to discover characteristic activity routine patterns of healthy individuals using activity primitives recognized from supervised sensor data. Yet, the applicability of topic models for hemiparetic rehabilitation patients and techniques to derive activity primitives without supervision needs to be addressed.

Objectives: We investigate, 1) whether a topic model-based activity routine discovery framework can infer activity routines of rehabilitation patients from wearable motion sensor data. 2) We compare the performance of our topic model-based activity routine discovery using rule-based and clustering-based activity vocabulary.

Methods: We analyze the activity routine discovery in a dataset recorded with 11 hemiparetic rehabilitation patients during up to ten full recording days per individual in an ambulatory daycare rehabilitation center using wearable motion sensors attached to both wrists and the non-affected thigh. We introduce and compare rule-based and clustering-based activity vocabulary to process statistical and frequency acceleration features to activity words. Activity words were used for activity routine pattern discovery using topic models based on Latent Dirichlet Allocation. Discovered activity routine patterns were then mapped to six categorized activity routines.

Results: Using the rule-based approach, activity routines could be discovered with an average accuracy of 76% across all patients. The rule-based approach outperformed clustering by 10% and showed less confusions for predicted activity routines.

Conclusion: Topic models are suitable to discover daily life activity routines in hemiparetic rehabilitation patients without trained classifiers and activity annotations. Activity routines show characteristic patterns regarding activity primitives including body and extremity postures and movement. A patient-independent rule set can be derived. Including expert knowledge supports successful activity routine discovery over completely data-driven clustering.

 
  • References

  • 1 Gietzelt M, Wolf K, Kohlmann M, Marschollek M, Haux R. Measurement of Accelerometry-based Gait Parameters in People with and without Dementia in the Field. A Technical Feasibility Study. Methods Inf Med 2013; 52 (04) 319-325.
  • 2 Marschollek M. et al. Sensor-based fall risk assessment - an expert ‘to go’. Methods Inf Med 2011; 50 (05) 420-426.
  • 3 Antón D. et al. Exercise Recognition for Kinect-based Telerehabilitation. Methods Inf Med 2015; 54 epub ahead of print Oct 10, 2014.
  • 4 Del Din S. et al. Estimating Fugl-Meyer clinical scores in stroke survivors using wearable sensors. In: EMBC. IEEE 2011; 5839-5842.
  • 5 Haeuber E. et al. Accelerometer monitoring of home-and community-based ambulatory activity after stroke. Arch Phys Med Rehab 2004; 85 (12) 1997-2001.
  • 6 Uswatte G. et al. Ambulatory monitoring of arm movement using accelerometry: an objective measure of upper-extremity rehabilitation in persons with chronic stroke. Arch Phys Med Rehab 2005; 86 (07) 1498-1501.
  • 7 Wade E. et al. Automated administration of the wolf motor function test for post-stroke assessment. In: PervasiveHealth. IEEE 2010; 1-7.
  • 8 van der Pas SC. et al. Assessment of arm activity using triaxial accelerometry in patients with a stroke. Arch Phys Med Rehab 2011; 92 (09) 1437-1442.
  • 9 de Niet M. et al. The stroke upper-limb activity monitor: its sensitivity to measure hemiplegic upper-limb activity during daily life. Arch Phys Med Rehab 2007; 88 (09) 1121-1126.
  • 10 Fugl-Meyer AR. et al. The post-stroke hemiplegic patient. 1. A method for evaluation of physical performance. Scand J Rehabil Med 1974; 7 (01) 13-31.
  • 11 Gowland C. et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke 1993; 24: 58-63.
  • 12 Taub E. et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehab 1993; 74 (04) 347-354.
  • 13 Bao L, Intille SS. Activity recognition from user-annotated acceleration data. In: Pervasive computing. Springer 2004; 1-17.
  • 14 Blanke U, Schiele B. Daily routine recognition through activity spotting. In: Location and Context Awareness. Springer 2009; 192-206.
  • 15 Tapia EM. et al. Activity recognition in the home using simple and ubiquitous sensors. In: Proceedings of Pervasive. Springer 2004; 158-175.
  • 16 Mannini A. et al. Machine learning methods for classifying human physical activity from on-body accelerometers. Sensors 2010; 10 (02) 1154-1175.
  • 17 Parkka J. et al. Activity classification using realistic data from wearable sensors. IEEE Transactions on Information Technology in Biomedicine 2006; 10 (01) 119-128.
  • 18 Huynh T. et al. Discovery of activity patterns using topic models. In: UbiComp. ACM 2008; 10-19.
  • 19 Farrahi K, Gatica-Perez D. Discovering routines from large-scale human locations using probabilistic topic models. ACM Trans Intell Syst Technol 2011; 2 (01) 1-27.
  • 20 Begole JB. et al. Rhythm modeling, visualizations and applications. In: UIST. ACM 2003; 11-20.
  • 21 Barger TS. et al. Health-status monitoring through analysis of behavioral patterns. IEEE Transactions on Systems, Man and Cybernetics 2005; 35 (01) 22-27.
  • 22 Blei DM. et al. Latent Dirichlet allocation. JMLR 2003; 3: 993-1022.
  • 23 Seiter J. et al. Activity Routine Discovery in Stroke Rehabilitation Patients without Data Annotation. In: PervasiveHealth 2014; 270-273.
  • 24 Burns A. et al. SHIMMER - A wireless sensor platform for noninvasive biomedical research. Sensors Journal, IEEE 2010; 10 (09) 1527-1534.
  • 25 Selim SZ, Ismail MA. K-means-type algorithms: a generalized convergence theorem and characterization of local optimality. TPAMI 1984; 1: 81-87.
  • 26 Cunningham P. et al. k-Nearest neighbour classifiers. Multiple Classifier Systems. 2007: 1-17.
  • 27 Seiter J. et al. Discovery of activity composites using topic models: An analysis of unsupervised methods. Pervasive Mob Comput 2014; 15: 215-227.
  • 28 Ho TK. et al. Decision combination in multiple classifier systems. TPAMI 1994; 16 (01) 66-75.
  • 29 Teh YW. et al. Hierarchical Dirichlet processes. JASA 2006; 101: 1566-1581.