Files and links (1)
PDFOpen Access (License Unspecified), Open Access
Thesis
Different forms of modularity in trunk muscles in the rat revealed by various statistical methods
Master of Science (M.S.), Drexel University
May 2007
DOI:
https://doi.org/10.17918/etd-1563
Abstract
Motor responses are governed by the central nervous system at various levels viz., cerebrum (highly complex responses ), extending into brain stem (responses are less complicated) and in the spinal cord where the responses include simple motor reflexes. One of the widely accepted views is that the nervous system can orchestrate a huge repertoire of movements because of its ability to incorporate modularity in its control strategies at various levels. In the current context, modularity in the coordination of trunk muscles in rat motor behavior is examined in the spinal cord (of an intact rat). Modularity at the level of spinal cord can be examined by identifying co-activation of trunk muscles. In other words, existence of modularity in spinal cord is validated by extracting synergies or pre-motor drives activating a chosen group of trunk muscles. Extraction of synergies from Electromyogram (EMG) recordings have been performed using dimension reduction techniques such as ICA (Independent Component Analysis), NNMF (Non-negative matrix factorization), cluster analysis, and factor analysis (FA) and Principal Component Analysis (PCA) [Hart and Giszter. 2004, Drew and Krouchev. 2006, Ivanenko et al. 2005], etc. Each of these techniques assumes different aspects of the data as a basis for dimension reduction. In the current project extraction of synergies from leg and trunk EMG was performed using ICA. This technique was preferred because it uses information separation rather than variance as the basis for dimension reduction [Hart and Giszter. 2004, Bell and Sejnowski. 1995]. When ICA was applied to 16 leg muscle recordings (leg EMG) in the rat, it revealed the existence of modularity. ICA was able to perform dimension reduction of leg EMG and preserved the onset information of bursts within 30 ms. This indicates the efficiency with which significant components yielded by ICA capture data and retain the onset information associated with each burst. When ICA was applied to 12 trunk muscles recordings (trunk EMG) in the rat, it did not indicate the existence of modularity. Since ICA did not reveal modularity in trunk muscles, an examination of modularity in phase was performed using various clustering techniques (K-means, Hierarchical and Associative clustering). K-means and Hierarchical clustering techniques examined the peak onset shifts (in all channels w.r.t the onset of every event in the reference channel, the hip flexor) and the corresponding phase information in an n-dimensional space (n =number of bursts in the EMG recording in the hip flexor). These techniques indicated consistent groupings comprising of the hip flexor and at least one of the ipsi-lateral abdominal muscles. Coupling of the hip flexor with the extensor of the contra-lateral forelimb was also observed, but usually on one side. Associative clustering was used to examine the onset and offset phase of every burst in a two dimensional phase plane. It incorporated both the onset and offset phase information associated with every burst. This technique showed results consistent with K-means and Hierarchical clustering. Also the distribution of two dimensional phase data was mostly robust for the same animal across different recording sessions (for at least 8 out of 12 channels). The results of Associative clustering were further used to construct direct components. Direct components could be used to compare the muscle groupings indicated by Associative clustering with that of ICA. But this comparison was not significant since ICA did not indicate modularity. The various clustering techniques applied to trunk muscle recordings indicate the existence of modularity in phase, by revealing consistent groupings of muscles in the n-dimensional (n >=2) phase domain. An examination of the existence of groupings in trunk muscles in phase provides an insight into the strategies used by the spinal cord in the control and co-ordination of locomotion.
Metrics
18 File views/ downloads
16 Record Views
Details
- Title
- Different forms of modularity in trunk muscles in the rat revealed by various statistical methods
- Creators
- Vidyaangi Patil - DU
- Contributors
- Simon F. Giszter (Advisor) - Drexel University (1970-)
- Awarding Institution
- Drexel University
- Degree Awarded
- Master of Science (M.S.)
- Publisher
- Drexel University; Philadelphia, Pennsylvania
- Resource Type
- Thesis
- Language
- English
- Academic Unit
- School of Biomedical Engineering, Science, and Health Systems (1997-2026); Drexel University
- Other Identifier
- 1563; 991014632937804721