Statistical Parametric Mapping (SPM) for alpha-based statistical analyses of multi-muscle EMG time-series

Area of Science:

• Biomechanics
• Neuroscience
• Statistics

Background:

• Electromyography (EMG) time-series data are inherently correlated and time-dependent.
• Traditional scalar analysis methods often fail to capture these complex dependencies in multi-muscle EMG data.
• This limitation can lead to missed findings or Type II errors in statistical analysis.

Purpose of the Study:

• To promote and demonstrate the utility of SPM vector-field analysis for generalized EMG time-series analysis.
• To contrast the effectiveness of SPM vector-field analysis with traditional scalar analysis on EMG gait data.
• To highlight the ability of SPM vector-field analysis to control for Type I and Type II errors.

Main Methods:

• Reanalysis of a publicly available dataset comparing young versus adult EMG gait data.
• Application of independent scalar analysis to specific EMG time-series segments (35%-45% stance phase).
• Application of SPM vector-field analysis to the same EMG gait dataset.

Main Results:

• Independent scalar analysis revealed no statistically significant differences between young and adult groups in the specified stance phase.
• SPM vector-field analysis identified statistically significant differences within the same time period.
• Scalar analysis exhibited Type II error due to failure to account for multi-muscle and time dependencies.

Conclusions:

• SPM vector-field analysis is highly applicable to EMG data, effectively accounting for multi-muscle and time dependencies.
• This method offers superior control for Type I and Type II errors compared to traditional scalar analysis.
• The generalizability of SPM vector-field analysis to various statistical models makes it valuable for electromyography and kinesiology research.