Region-of-interest analyses of one-dimensional biomechanical trajectories: bridging 0D and 1D theory, augmenting statistical power

Abstract

One-dimensional (1D) kinematic, force, and EMG trajectories are often analyzed using zero-dimensional (0D) metrics like local extrema. Recently whole-trajectory 1D methods have emerged in the literature as alternatives. Since 0D and 1D methods can yield qualitatively different results, the two approaches may appear to be theoretically distinct. The purposes of this paper were (a) to clarify that 0D and 1D approaches are actually just special cases of a more general region-of-interest (ROI) analysis framework, and (b) to demonstrate how ROIs can augment statistical power. We first simulated millions of smooth, random 1D datasets to validate theoretical predictions of the 0D, 1D and ROI approaches and to emphasize how ROIs provide a continuous bridge between 0D and 1D results. We then analyzed a variety of public datasets to demonstrate potential effects of ROIs on biomechanical conclusions. Results showed, first, that a priori ROI particulars can qualitatively affect the biomechanical conclusions that emerge from analyses and, second, that ROIs derived from exploratory/pilot analyses can detect smaller biomechanical effects than are detectable using full 1D methods. We recommend regarding ROIs, like data filtering particulars and Type I error rate, as parameters which can affect hypothesis testing results, and thus as sensitivity analysis tools to ensure arbitrary decisions do not influence scientific interpretations. Last, we describe open-source Python and MATLAB implementations of 1D ROI analysis for arbitrary experimental designs ranging from one-sample t tests to MANOVA.

Negative Influence of Motor Impairments on Upper Limb Movement Patterns in Children with Unilateral Cerebral Palsy. A Statistical Parametric Mapping Study

“…Furthermore, our analysis identified specific deviant parts of the movement pattern, i.e., clusters, which may help establishing the basis for further studies. Combining these identified clusters (regions of interest, Pataky et al, 2016a ), possibly together with spatiotemporal parameters and extracted scalars will permit further hypothesis driven research. Moreover, the implementation of dimensionality reduction tool [Principal Component Analysis (PCA), Independent Component Analysis (ICA), or kernel Principal Component Analysis (kPCA)] and/or machine learning tools [Artificial Neural Network (ANN), Support Vector Machines (SVM), or Self-Organizing Maps (SOM)] would be of interest to classify movement patterns.…”

Physiological Responses to Firefighting in Extreme Temperatures Do Not Compare to Firefighting in Temperate Conditions

“…Basically, a FFT method was used, where the original vector was transformed to the Fourier domain and then transformed back with desired data points (here 100, representing 100% of completion time). In addition, data was processed with the open source SPM code by Pataky ( 2016 ) for two-tailed paired t -tests. The critical test statistic threshold that retained a family-wise error rate of α = 0.05 was calculated as described by Pataky et al ( 2015 ).…”

A force profile analysis comparison between functional data analysis, statistical parametric mapping and statistical non-parametric mapping in on-water single sculling

“…in the example of running gait, high frequency content in ground reaction force at foot-strike and then lower frequency content throughout the rest of the stance phase). Additionally, as SPM and SnPM treats data as a vector of points, these approaches are more flexible for a priori selection of important regions of interest (ROI) on a waveform 33 , and allows for hypothesis testing to take place on these ROIs (thus influencing statistical power), rather than the whole vector. This is not possible within FDA.…”

“…As a cautionary point for the potential user of SPM or SnPM however, choosing smaller regions within a waveform will change the critical t-threshold for a given alpha level (as the smoothness of the curve will change, thus influencing the outcomes of RFT). Researchers, therefore, must have a strong rationale for pre-selection of a section of a movement to be analysed, as the results of an ROI analysis will likely change, when compared to SPM or SnPM being applied to the entire time-series 33 .…”