Systematic data analysis and data mining in CatWalk gait analysis by heat mapping exemplified in rodent models for neurodegenerative diseases

Timotius I, Canneva F, Minakaki G, Moceri S, Plank AC, Casadei N, Riess O, Winkler J, Klucken J, Eskofier B, von Hörsten S (2019)

Publication Language: English

Publication Type: Journal article, Original article

Publication year: 2019

Journal

Journal of Neuroscience Methods Elsevier

Book Volume: 326

Article Number: 108367

URI: https://www.sciencedirect.com/science/article/pii/S0165027019302249

DOI: 10.1016/j.jneumeth.2019.108367

Open Access Link: https://www.sciencedirect.com/science/article/pii/S0165027019302249

Abstract

Background

Motor impairment appears as a characteristic symptom of several diseases and injuries. Therefore, tests for analyzing motor dysfunction are widely applied across preclinical models and disease stages. Among those, gait analysis tests are commonly used, but they generate a huge number of gait parameters. Thus, complications in data analysis and reporting raise, which often leads to premature parameter selection.

New methods

In order to avoid arbitrary parameter selection, we present here a systematic initial data analysis by utilizing heat-maps for data reporting. We exemplified this approach within an intervention study, as well as applied it to two longitudinal studies in rodent models related to Parkinson’s disease (PD) and Huntington disease (HD).

Results

The systematic initial data analysis (IDA) is feasible for exploring gait parameters, both in experimental and longitudinal studies. The resulting heat maps provided a visualization of gait parameters within a single chart, highlighting important clusters of differences.

Comparison with existing method

Often, premature parameter selection is practiced, lacking comprehensiveness. Researchers often use multiple separated graphs on distinct gait parameters for reporting. Additionally, negative results are often not reported.

Conclusions

Heat mapping utilized in initial data analysis is advantageous for reporting clustered gait parameter differences in one single chart and improves data mining.

Authors with CRIS profile

Ivanna Timotius Lehrstuhl für Informatik 14 (Bild- und Sprachverarbeitung) Georgia Minakaki Professur für Molekulare Neurologie Sandra Moceri Professur für Experimentelle Biomedizin Anne-Christine Plank Professur für Experimentelle Biomedizin Jürgen Winkler Professur für Molekulare Neurologie Jochen Klucken Department of Molecular Neurology Björn Eskofier Lehrstuhl für Informatik 14 (Machine Learning and Data Analytics) Stephan von Hörsten Professur für Experimentelle Biomedizin

Additional Organisation(s)

Machine Learning and Data Analytics Lab Interdisziplinäres Zentrum für Klinische Forschung

Involved external institutions

Eberhard Karls Universität Tübingen DE Germany (DE)

How to cite

APA:

Timotius, I., Canneva, F., Minakaki, G., Moceri, S., Plank, A.-C., Casadei, N.,... von Hörsten, S. (2019). Systematic data analysis and data mining in CatWalk gait analysis by heat mapping exemplified in rodent models for neurodegenerative diseases. Journal of Neuroscience Methods, 326. https://dx.doi.org/10.1016/j.jneumeth.2019.108367

MLA:

Timotius, Ivanna, et al. "Systematic data analysis and data mining in CatWalk gait analysis by heat mapping exemplified in rodent models for neurodegenerative diseases." Journal of Neuroscience Methods 326 (2019).

BibTeX: Download