CURRENT PROJECTS

I. NSF CAREER ADAPTIVE BIOFEEDBACK TO IMPROVE BALANCE DURING EVERYDAY MOBILITY

Sonified biofeedback conveys biomechanics metrics via sound. We are developing systems with which people can interactively “tune” music biofeedback to tune their movement strategies. Recently, our lab has been awarded with an NSF CAREER award titled “Adaptive Biofeedback to Improve Balance during Everyday Mobility” to advance this mission.

There are two major objectives associated with this grant. One is to understand balance during turns of varied environmental contexts and the second is to use real-time sonified biofeedback to improve balance.

I.1 Balance during turns

In the first objective of the CAREER project, we are focused on understanding person-specific balance strategies used during turns of varied environmental contexts. For our first study, we selected the everyday mobility context of turning into an aisle in a store when the item of interest is known to be in a specific aisle (pre-planned), or when a visual cue is received by looking into the aisle while walking by the aisle (late-cued).

During the pre-planned turns, we can observe that the turn occurs over multiple footfalls. The participant does not demonstrate the traditional “step” or “spin” turn strategies. These strategies would be demonstrated if there were one footfall that accomplishes most of the change in linear trajectory and body facing direction.

• A spin turn is marked by the most rotation occurring during a footfall of the foot on the side of the body in the desired direction of travel (interior to the turn).

• A step turn is marked by the most rotation occurring during a footfall of the foot on the side of the body opposite of the desired direction of travel (exterior to the turn).

Instead, this participant, and others used multiple footfalls to gradually change their body’s horizontal trajectory and facing direction. The overhead view demonstrates the curved horizontal center of mass trajectory with the body gradually reorienting to the desired facing direction.

During these late-cued turn examples, we observe that when the participant started the trial with a left foot step, they use a small half-step to redirect their center of mass translation and rotate their body suddenly with a footfall that facilitated ground reaction forces in the desired direction of travel and a widened base of support (vs. turning when supported only by the previous footfall). In contrast, during the late-cued turns performed when the participant started the trial with a right foot step, they were able to complete the turn without this half-step. This was accomplished due to a combination of factors, one of which may be that the cue was received with enough time for the right foot fall’s role to quickly switch to accommodate the turn’s mechanical objectives.

These trials demonstrate a few strategies that young adults can use to turn during pre-planned and late-cued conditions. The fact that behavior was similar across trials started with each foot implies the importance of the footfall context relative to the location of the turn in strategy selection (strategies used depend partially on spatial configurations).

During these different turn strategies, we are starting to understand the coordination between changing linear momentum and angular momentum about different body axes. The newest analysis we are undertaking is to understand the coupling of the body’s mediolateral margin of stability (distance from the center of mass to the mediolateral edge of the base of support) along with whole body angular momenta in the frontal plane (body rotation side-to-side about a forward axis) and transverse plane (body rotation about vertical axes).

We are finding that during pre-planned turns, the mediolateral margin of stability is at its lowest in the mid-range of frontal and transverse plane angular momenta. However, during late-cued turns, the margin of stability is negative (meaning the center of mass is outside of the mediolateral bounds of the base of support) at times that the angular momenta are at rapidly changing or at their greatest values. This is suggesting that during the late-cued turns, people are challenged in ways that impose the need to rotate about multiple axes while barely supported by the base of support.

I.2 Real-Time Sonified Balance Biofeedback

Our sound biofeedback design is currently in redesign phase. Please see below for preliminary studies of balance using our original sound design prototype.

Our Ph.D. student Mitchell Tillman recently presented a preliminary study related to this research virtually at ASB2020! Please click here for a pdf of Mitchell’s poster and watch the video by clicking the picture of the poster.

Methods Details for initial study of stationary single leg balance:

The current sonification design sound generation consists of two oscillators whose outputs are each sent through a band-pass filter, then summed and followed by a gain. The sound can be varied according to three parameters: pulse speed, brightness, and dissonance. Pulse speed varies the periodic rate at which the gain and filters are modulated; brightness controls the average center frequency of the filters; dissonance controls the relative tuning of the two oscillators. Pulse speed and brightness mappings were scaled from 0 to 1 such that the entire range was achievable during the single leg stance task.

Margin of stability (MOS) is mapped to pulse speed such that as the total body center of mass (TBCM) moved closer to the Base of Support (BOS) convex hull edge (lower MOS) the pulse speed increases closer to 1 (faster pulse speed). This mapping aims to impart a sense of urgency about moving away from the BOS boundaries, while the very slow pulses when the TBCM was at the center of the BOS indicates relative safety from falling.

Next, BOS convex hull area is mapped inversely to brightness; larger BOS areas resulted in lower brightness, smaller BOS areas resulted in higher brightness (Fig. 2). Higher brightness is less pleasant to listen to, therefore the sound emphasized maintaining a larger base of support, which we think will be more conducive to balance in static tasks.  Finally, the Boolean value of whether TBCM was within (1) or outside (0) of the BOS convex hull was mapped to the dissonance of the system. If the Boolean is 0, there is no dissonance. But if TBCM is outside the BOS boundary (Boolean=1), dissonance is introduced to the audio.

audiovisual examples:

II. Sports Biomechanics Projects (Please check for updates periodically!)

Pitching Biomechanics

Among many research questions, we are interested in understanding how each leg contributes to whole body angular and linear momentum during the baseball pitch. We have many other research questions, but we are eager to share preliminary findings on this page! Our doctoral student, Sam Liu, has uncovered how each leg works to generate linear and angular impulse, which initiate the linear and angular momenta respectively.

Professional pitchers (n=4) volunteered for the study and pitched six to eleven fastballs from the stretch position on an instrumented mound while kinematic and kinetic data were collected . We excluded pitching trials that were rated unrepresentative/poor by the pitchers. The variables of interest were “X GRF”, the forward directed GRF, and “My”, the moment applied about the Y-axis passing through the TBCM (forward rotating moment) (see next Figure). These two variables were integrated to find the X Linear Impulse and Y Angular Impulse, respectively, for each leg and both legs (net) from the beginning of the push phase until ball release. The push phase began the last time Z GRF decreased before increasing to its peak.

Volleyball spike biomechanics

III. Shoulder Orthopaedic Research

Reverse total shoulder arthroplasty