Visual Observation Skills in Therapists, The Physiology Behind Building Muscle, and The Two Pains of Life

🦾 Biomechanics, motor control, and (re)training paper recommendations

How accurate are visual assessments by physical therapists of lumbopelvic movements during the squat and deadlift?

Anyone with experience in strength and conditioning, physical therapy, chiropractic care, or other applied human movement professionals have experience with using their eyes to evaluate how people coordinate actions. A joint motion that practitioners may visually observe is lumbar spine motion during squatting and deadlifting activities. This paper assessed six physical therapists' ability to observe lumbo-pelvic motion in 14 powerlifters and 10 Olympilifters. They found that a posterior pelvic tilt of ~34deg was required for trained therapists to detect the motion. Although one could argue that this population may have relatively large erector spine muscles that could make it particularly challenging to observe lumbar spine flexion, this is still a considerable amount of pelvic tilt to begin detecting this motion. To put that into perspective, here is a visual representation of 34 degrees:

Although our eyes may be helpful tools for some purposes, it's perhaps important to exercise caution when interpreting how people move!

Estimating 3D joint kinematics from video sequences of running and cutting maneuvers—assessing the accuracy of simple visual inspection

Now you might wonder, "what if we trained people to observe these motions using the motion capture data?" Part of this paper addressed that exact question- after an initial test to evaluate therapists' accuracy in assessing knee valgus motion, they received 35 more videos with feedback on the knee valgus angle. Then, they re-took the test, and the researchers evaluated whether the therapists' visual observation skills improved but found that there was minimal, if any, improvement. However, it is crucial to consider that 1) therapists' interpretations of joint motions didn't align with the mathematics of how joint angles are computed in biomechanics, and 2) the training program wasn't very long, so it's possible that a more extended intervention could have seen better results. Nevertheless, I think it's essential for practitioners to be aware of the limitations of their visual observation skills when interpreting human motion.

Association Between Passive Hip Range of Motion and Pitching Kinematics in High School Baseball Pitchers

Slight change of pace from the studies above, but I always find it interesting to see across various applications how well data from our movement assessments can inform how people may move during sporting activities. The findings from this paper further illustrate that assessing individuals' joint constraints alone is perhaps not informative about how movement may emerge in sport-specific contexts.

🤖 Statistics and machine learning paper recommendations

ChatGPT3 Results for Writing Biomechanics and Motor Control Functions

I've been messing around a lot with ChatGPT3 recently and need to share how well it seems to write code with such simple prompts. For example, the only question I asked it was: "Can you write me a Python function for uncontrolled manifold analyses," and this was the code it provided:

import numpy as np

def ucm_analysis(data):
    """
    Performs uncontrolled manifold (UCM) analysis on a given dataset.

    Parameters:
    data (numpy.ndarray): A matrix of shape (num_samples, num_variables) representing the dataset.

    Returns:
    dict: A dictionary containing the UCM analysis results, including the UCM matrix, variance explained by the UCM, and variance explained by the non-UCM directions.
    """

    # Calculate the Jacobian matrix for the dataset
    jacobian = np.gradient(data, axis=0)

    # Calculate the average Jacobian
    avg_jacobian = np.mean(jacobian, axis=0)

    # Calculate the deviation of each Jacobian from the average Jacobian
    jacobian_deviation = jacobian - avg_jacobian

    # Calculate the UCM by finding the null space of the covariance matrix of the jacobian deviation
    cov = np.cov(jacobian_deviation.T)
    eigenvalues, eigenvectors = np.linalg.eig(cov)
    ucm = eigenvectors[:, np.where(eigenvalues == 0)[0][0]]

    # Calculate the variance explained by the UCM and non-UCM directions
    ucm_var = np.sum(np.dot(jacobian_deviation, ucm)**2)
    non_ucm_var = np.sum(np.dot(jacobian_deviation, np.delete(eigenvectors, np.where(eigenvalues == 0)[0][0], axis=1))**2)

    # Create a dictionary containing the UCM analysis results
    results = {
        'ucm': ucm,
        'ucm_var': ucm_var,
        'non_ucm_var': non_ucm_var
    }

    return results

I'm just blown away by how, with relatively simple and uninformative prompts, it can provide you with a very good skeleton of code from which to work with.

The Utility of Mixed Models in Sport Science: A Call for Further Adoption in Longitudinal Data Sets

Great reading for anyone interested in implementing more mixed modelling into their analyses!

Failing Grade: 89% of Introduction-to-Psychology Textbooks That Define or Explain Statistical Significance Do So Incorrectly

https://journals.sagepub.com/doi/10.1177/2515245919858072

Kind of an overly critical newsletter at this point, but I think this is an interesting paper for anyone doing any statistical analyses to read and be familiar with.

🎙 Podcast recommendation

This Podcast with Dr. Andy Galpin is an excellent physiology refresher. Catching up on and re-reviewing physiology is somewhat of an impromptu and late New Year's resolution for me as I've neglected these papers for so long, so I'll be listening to this one a few more times as well!

🗣 Quote of the month

"We must all suffer from one of two pains: the pain of discipline or the pain of regret. The difference is discipline weighs ounces while regret weighs tons."

- Jim Rohn