Publications

KA Ingraham, NL Zaino, C Feddema, ME Hoffman, L Gijbels, A Sinclair, AN Meltzoff, PK Kuhl, HA Feldner, KM Steele (2025) “Quantifying Joystick Interactions and Movement Patterns of Toddlers With Disabilities Using Powered Mobility With an Instrumented Explorer Mini”

Katie Landwehr EarlyMobility, JournalArticle, News, Publications , , , , ,

Journal Article in IEEE Transactions on Neural Systems and Rehabilitation Engineering

Powered mobility technology can be a powerful tool to facilitate self-initiated exploration and play for toddlers with motor disabilities. The joystick-controlled Permobil Explorer Mini is currently the only commercially available powered mobility device for children ages 1-3 years in the United States. However, many open questions persist regarding how joystick-based mobility technologies should be designed to optimally suit the developmental needs of toddlers.

The instrumented Explorer Mini measures joystick position in (x,y) coordinates and the number of wheel rotations for the left and right wheels at a sampling frequency of 100 Hz. Wheel displacement is calculated by multiplying the number of rotations by the measured wheel circumference. Representative raw data collected from the device are shown here for 100 seconds.Aim: The purpose of this study was to quantify how toddlers with motor disabilities use the Explorer Mini during free exploration and play.

Methods: For this work, we developed a custom-instrumented Explorer Mini with embedded sensors to measure joystick interactions and wheel rotations. Nine children with motor disabilities (ages 12-36 months) participated in 12 in-lab visits, and during each visit they engaged in two 15-20 minute play sessions. For each session, we calculated several quantitative outcome metrics, including the time spent using the joystick, distance traveled, and the number, duration, and complexity of joystick interactions.

Results: Every participant independently interacted with the joystick and moved the Explorer Mini during every session. Over 12 visits, participants significantly increased their distance traveled and the time spent with the joystick active. Surprisingly, we found that only 48% of joystick interactions resulted in device movement, which has important implications for learning.

Interpretation: These results can serve as a benchmark for caregivers and clinicians to understand early device use patterns. Furthermore, this knowledge can be used to inform the design of new powered mobility technologies for toddlers with disabilities or support the refinement of existing devices.

SR Shrivastav, CR DeVol, VM Landrum, KF Bjornson, D Roge, KM Steele, CT Moritz (2024) “Transcutaneous Spinal Stimulation and Short-burst Interval Treadmill Training in Children with Cerebral Palsy: A Pilot Study”

Katie Landwehr JournalArticle, News, Publications, SpinalStim , , ,

Journal Article in IEEE Transactions on Biomedical Engineering

Non-invasive neuromodulation may be an alternative approach that can improve outcomes in CP when combined with physical therapy. Transcutaneous spinal cord stimulation (tSCS) is a novel, non-invasive neuromodulation technique that can modulate spinal and supraspinal circuits especially when implemented with physical therapy.

A) Short-burst interval locomotor treadmill training (SBLTT) with contact guard assist. B) Investigative spinal cord neuromodulation device (SpineX, Inc.) with stimulating electrodes on the T11 and L1 dorsal spinous processes and two ground electrodes on the anterior superior iliac spine (ASIS - not visible). C) Spinal stimulation waveform with 10 kHz carrier frequency. D) Protocol timeline including the assessments before and after each intervention and after 8-weeks of follow-up. tSCS = transcutaneous spinal cord stimulationAim: The purpose of this pilot study was to evaluate the effects of transcutaneous spinal cord stimulation (tSCS) and short-burst interval locomotor treadmill training (SBLTT) on spasticity and mobility in children with cerebral palsy (CP).

Methods: We employed a single-arm design with two interventions: SBLTT only, and tSCS + SBLTT, in four children with CP. Children received 24-sessions each of SBLTT only and tSCS + SBLTT. Spasticity, neuromuscular coordination, and walking function were evaluated before, immediately after, and 8- weeks following each intervention.

Results: Spasticity, measured via the Modified Ashworth Scale (MAS), reduced in four lower extremity muscles after tSCS + SBLTT (1.40 ± 0.22,) more than following SBLTT only (0.43 ± 0.39). One-minute walk test (1-MWT) distance was maintained during both interventions. tSCS + SBLTT led to improvements in peak hip and knee peak extension (4.9 ± 7.3° and 6.5 ± 7.7°), that drove increases in joint dynamic range of 4.3 ± 2.4° and 3.8 ± 8.7° at the hip and knee, respectively. Children and parents reported reduction in fatigue and improved gait outcomes after tSCS + SBLTT. Improvements in spasticity and walking function were sustained for 8-weeks after tSCS + SBLTT.

Interpretation: These preliminary results suggest that tSCS + SBLTT may improve spasticity while simultaneously maintaining neuromuscular coordination and walking function in ambulatory children with CP. This work provides preliminary evidence on the effects of tSCS and the combination of tSCS + SBLTT in children with CP.

KA Ingraham, HA Feldner, KM Steele (2024) “Forward first: Joystick interactions of toddlers during digital play”

Katie Landwehr JournalArticle, News, Publications , , , ,

Journal Article in PLoS ONE

Computers and technology are essential tools for supporting the development of toddlers with and without disabilities. Developmentally appropriate access to technology can support toddlers in learning and play. While touch screens are a popular interaction modality for children under the age of three, they may not be appropriate for all children or all tasks.

A child is pictured sitting in a Explorer Mini while using the joystick to play a directional cause-and-effect game on a computer screen, toddlers demonstrated a strong preference for initiating movement in the forward direction, irrespective of the target on the screen.Aim: We know comparatively little about how toddlers interact with joystick-based technology, and more fundamental research is required to understand joystick interactions at different ages and developmental stages.

Methods: We quantified how 36 nondisabled toddlers used a joystick to play a cause-and-effect game on a computer.

Results: Children demonstrated a strong preference for moving the joystick forward first, regardless of the target direction. On average, the oldest children navigated the joystick to the target 5 seconds faster than the youngest children, and were nearly twice as efficient in their joystick path.

Interpretation: These findings inform the design of assistive algorithms for joystick-enabled computer play and developmentally appropriate technologies for toddlers.

AM Spomer, BC Conner, MH Schwartz, ZF Lerner, KM Steele (2024) “Multi-session adaptation to audiovisual and sensorimotor biofeedback is heterogeneous among adolescents with cerebral palsy”

Katie Landwehr JournalArticle, MotorControl, News, Publications , , , , ,

Journal Article in PLoS ONE

There is growing interest in the use of biofeedback-augmented gait training in cerebral palsy (CP). Audiovisual, sensorimotor, and immersive biofeedback paradigms are commonly used to elicit short-term gait improvements; however, outcomes remain variable. Because biofeedback training requires that individuals have the capacity to both adapt their gait in response to feedback and retain improvements across sessions, changes in either capacity may affect outcomes. Yet, neither has been explored extensively in CP.

Experimental protocol used to evaluate multi-session adaptation to multimodal biofeedback. Participants completed a four-day protocol using combined audiovisual and sensorimotor biofeedback. Audiovisual biofeedback on soleus activity was provided unilaterally on the more-affected limb whereas sensorimotor biofeedback was administered bilaterally using a resistive ankle exoskeleton. Each session was separated into baseline (1 minute), biofeedback (2, 10-minute bouts), and washout (1 minute) phases. The nominal torque value of the ankle exoskeleton was set at 0.1 Nm/kg during the first bout of the first session and incrementally adjusted by 0.025 Nm/kg over the subsequent bouts, according to the schedule shown. Overground walking data were collected pre- and post-intervention. A licensed physical therapist also performed a full physical examination at the pre-intervention session. Motion capture data were collected during at the pre- and post-intervention sessions and electromyography (EMG) data were collected bilaterally from the vastus lateralis, semitendinosus, soleus, and tibialis anterior across all sessions.Aim: The aim of this study was to evaluate the extent to which individuals with CP adapt gait and retain improvements during multi-session practice with a multimodal biofeedback paradigm, designed to promote plantarflexor recruitment. Secondarily, we compared overground walking performance before and after biofeedback sessions to understand if any observed in-session improvements were transferred. 

Methods: In this study, we evaluated the extent to which adolescents with CP (7M/1F; 14 years (12.5,15.26)) could adapt gait and retain improvements across four, 20-minute sessions using combined audiovisual and sensorimotor biofeedback. Both systems were designed to target plantarflexor activity. Audiovisual biofeedback displayed real-time soleus activity and sensorimotor biofeedback was provided using a bilateral resistive ankle exoskeleton. We quantified the time-course of change in muscle activity within and across sessions and overground walking function before and after the four sessions.

Results: All individuals were able to significantly increase soleus activity from baseline using multimodal biofeedback (p < 0.031) but demonstrated heterogeneous adaptation strategies. In-session soleus adaptation had a moderate positive correlation with short-term retention of the adapted gait patterns (0.40 ≤ ρ ≤ 0.81), but generally weak correlations with baseline walking function (GMFCS Level) and motor control complexity (ρ ≤ 0.43). The latter indicates that adaptation capacity may be a critical and unique metric underlying response to biofeedback. Notably, in-session gains did not correspond to significant improvements in overground walking function (p > 0.11).

Interpretation: This work suggests that individuals with CP have the capacity to adapt their gait using biofeedback, but responses are highly variable. Characterizing the factors driving adaptation to biofeedback may be a promising avenue to understand the heterogeneity of existing biofeedback training outcomes and inform future system optimization for integration into clinical care.

 

CR DeVol, SR Shrivastav, AM Spomer, KF Bjornson, D Roge, CT Moritz, KM Steele (2024) “Effects of interval treadmill training on spatiotemporal parameters in children with cerebral palsy: A machine learning approach”

Katie Landwehr JournalArticle, Publications , , , ,

Journal Article in Journal of Biomechanics

Quantifying individualized rehabilitation responses and optimizing therapy for each person is challenging. For interventions like treadmill training, there are multiple parameters, such as speed or incline, that can be adjusted throughout sessions.

A) Pre-post effect of SBLTT on step length for the more affected side. B) BART results quantify direct effects of SBLTT on step length. Accumulated Local Effects (ALE) plots for each input variable show the effect of that variable on step length including session number, treadmill speed (Froude number), time within session, side, and treadmill incline. The size of the data point on each ALE plot depicts the relative number of data points in each bin. C) BART model fit (R2) for each participant. D) Direct effects of each input variable on the response variable, step length, calculated from the change in the ALE plots in B).Aim: This study evaluates if causal modeling and Bayesian Additive Regression Trees (BART) can be used to accurately track the direct effects of treadmill training on gait.

Methods: We developed a Directed Acyclic Graph (DAG) to specify the assumed relationship between training input parameters and spatiotemporal outcomes during Short Burst Locomotor Treadmill Training (SBLTT), a therapy designed specifically for children with cerebral palsy (CP). We evaluated outcomes after 24 sessions of SBLTT for simulated datasets of 150 virtual participants and experimental data from four children with CP, ages 4–13 years old. Individual BART models were created from treadmill data of each step.

Results: Simulated datasets demonstrated that BART could accurately identify specified responses to training, including strong correlations for step length progression (R2 = 0.73) and plateaus (R2 = 0.87). Model fit was stronger for participants with less step-to-step variability but did not impact model accuracy. For experimental data, participants’ step lengths increased by 26 ± 13 % after 24 sessions. Using BART to control for speed or incline, we found that step length increased for three participants (direct effect: 13.5 ± 4.5 %), while one participant decreased step length (−11.6 %). SBLTT had minimal effects on step length asymmetry and step width.

Interpretation: Tools such as BART can leverage step-by-step data collected during training for researchers and clinicians to monitor progression, optimize rehabilitation protocols, and inform the causal mechanisms driving individual responses.