Enhanced contralateral hand learning through virtual reality-induced sensorimotor mismatch training

1520 

Abstract Submission 

Elisabeth Jochmann¹, Alexander Schmidt¹, Thomas Jochmann², Maximilian Weber¹, Matthias Nürnberger¹, Carsten Klingner¹

¹Jena University Hospital, Jena, Germany, ²Technische Universität Ilmenau, Ilmenau, Germany

Elisabeth Jochmann  
Jena University Hospital
Jena, Germany

Alexander Schmidt  
Jena University Hospital
Jena, Germany

Thomas Jochmann  
Technische Universität Ilmenau
Ilmenau, Germany

Maximilian Weber  
Jena University Hospital
Jena, Germany

Matthias Nürnberger  
Jena University Hospital
Jena, Germany

Carsten Klingner  
Jena University Hospital
Jena, Germany

Current rehabilitation strategies often insufficiently mitigate motor deficits following a stroke. Early post-stroke, patients experience a sensitive period with enhanced motor learning capabilities, which is crucial for regaining motor functions (Zeiler, 2019). However, this period of rapid recovery diminishes substantially over time (Coleman, 2017).
During that sensitive period of enhanced motor learning, patients typically suffer from a mismatch between intended and executed movements. Our study posits that this mismatch may be a key driver in motor learning. We hypothesize that artificially inducing a similar sensorimotor mismatch could reactivate the sensitive period for motor learning, particularly in elderly subjects.
To explore this, we evaluated motor learning in volunteers subjected to a virtual reality (VR)-induced sensorimotor mismatch. We used resting state functional magnetic resonance imaging (rs-fMRI) to investigate how this mismatch affects the brain's functional networks responsible for motor learning.

We recruited 69 healthy right-handed participants, aged 19-82 years (mean = 54.1 ± 19.2 years). Participants were randomized into two groups: mismatch (intervention) and errorless (control) (Fig. 1). Participants underwent fMRI scans (3T, Siemens Trio, resting state) before and after a 10-minute VR ball-throwing task with their right hand (Oculus Rift S VR headset, in-house developed game software, performed outside the scanner). The task consisted of picking up a ping pong ball from a table and throwing it at a target. For the mismatch group, the VR system introduced varying sensorimotor discrepancies, such as axis inversion, amplitude changes, or simulated tremors. The target size was reduced or increased to constantly maintain a moderate success rate. The errorless group performed the task without artificial sensorimotor disturbances and the target size adaptation was programmed to a higher success rate. Afterwards, a second round of the VR task was performed, followed by a motor learning task on a laptop: This task required participants to click on transient round targets using a computer mouse, structured into 12 blocks of 30 seconds each, with 30-second rest intervals, first using the right hand and then the left. The objective was to maximize target hits.
Rs-fMRI pre- and post-intervention were compared using parametric multivariate statistics (Jafri, 2008), comparing ROI-to-ROI evaluations centered on the somatomotor regions from the Schaefer 2018 atlas (Schaefer, 2018), using the CONN toolbox for SPM (Whitfield-Gabrieli and Nieto-Castanon, 2012). Motor learning performance was derived from the scores in the computer mouse-based target clicking task using a generalized estimating equation (GEE) in R (Liang, 1986; Højsgaard, 2006).

Rs-fMRI analysis: Focusing on participants aged 60 and above (mismatch: n=15; errorless: n=16), the mismatch group displayed enhanced functional connectivity in comparison to the errorless group. These changes were primarily observed in the right somatomotor areas, between the left and right somatomotor regions interhemispherically, and within the left somatomotor regions (Fig. 2). Across the entire cohort, no significant changes in resting state connectivity were observed.
Behavioral task: The mismatch group demonstrated a significant increase in the learning rate of the left hand versus the right hand in the mismatch group compared to the errorless group for an adjusted age of 60 years (p=0.039).

VR-induced artificial sensorimotor mismatch training enhances the functional connectivity in both the contralateral and ipsilateral somatomotor cortex. This training method also improves the learning rate of the contralateral hand in subsequent motor tasks. The next phase involves applying these findings to clinical settings, specifically examining functional connectivity and learning improvements in chronic stroke patients after VR mismatch training.

Skill Learning ²

Connectivity (eg. functional, effective, structural) ¹

fMRI Connectivity and Network Modeling

Visuo-Motor Functions

ADULTS

FUNCTIONAL MRI

Learning

Motor

Neurological

Somatosensory

Other - Virtual Reality