fMRI brain activation during a novel VR-based motor imagery and observation task: comparison with a conventional motor imagery task

Introduction: We have previously developed a novel VR-based motor imagery (MI) and motor observation (MO) task targeting the upper limbs for stroke rehabilitation-NeuRow; it consists of imagining the kinaesthetic experience of rowing using the left or right arm [1]. Here, we evaluate brain activation elicited by NeuRow and compare it with an abstract MI task using the conventional Graz paradigm. 

Subjects/methods: Two groups of healthy right-handed subjects participated: 10 young adults (27 ± 4 years); and 10 older subjects (51 ± 6 years). Imaging was performed on a 3T GE Signa HDxt MRI 123 S52 ESMRMB Congress (2020) 33 (Suppl 1): S47–S67scanner using a 12-channel RF coil. fMRI used 2D EPI (TR/TE = 2500/30 ms, 3.75 9 3.75 9 3.00 mm3 voxel size). Subjects underwent four fMRI runs (each consisting of 8 cycles, alternating 20 s blocks of task and baseline, total 5.33 min) under the following conditions (pseudo-randomized order): (1) left Graz-MI; (2) right Graz-MI; (3) left NeuRow; and (4) right NeuRow. for Graz-MI, an arrow indicated the arm movement (left/right) to imagine, while for NeuRow, a two-armed virtual avatar executed the intended movement (left/right). Data analysis was performed using FSL (fsl.fmrib.ox.ac.uk). After pre-processing (motion correction, spatial smoothing, highpass temporal filtering) and registration (to MNI space), a general linear model including the task and six motion parameters as confounds was fitted to the data. For each task, a two-sample unpaired t-test was used to assess differences between groups. Since no significant differences were found, a repeated-measures ANOVA was used to assess differences between sides (left/right arm) and condition (Graz-MI/ NeuRow) for the whole subject cohort. All statistical maps were converted into Z-Scores and submitted to cluster thresholding (voxel Z >3.1; cluster p <0.05) to correct multiple comparisons. 

Results/discussion: Left and right arms showed significant differences as expected (not shown). Compared to Graz-MI, NeuRow activated more occipital and parietal regions. Our findings are in line with the literature of MI and MO tasks, which report activation of parietal and occipital areas more consistently kindred with MO than MI [2]. Besides, we found no age-related differences, in concordance with a previous study demonstrating that MI ability is not affected up to 70 years [3]. Overall, our results indicate that NeuRow may be more powerful than conventional MI in the context of neurorehabilitation, by more strongly engaging brain areas associated with the mirror neuron and sensorimotor systems [4].

References

1. Vourvopoulos, Front Hum Neurosci, 3:1–17, 2019 

2. Hardwick, Neurosci Biobehav Rev, 94:31–44, 2018 

3. Schott, Exp Aging Res, 38(5):559–583, 2012 

4. Garrison, Neurorehabil Neural Repair, 24(5):404–412, 2010

Publication:

Nunes, J., Vourvopoulos, A., Blanco-Mora, D., Jorge, C., Fernandes, J.-C., Fernandes, S., Bermúdez i Badia, S., & Figueiredo, P. (2020). fMRI brain activation during a novel VR-based motor imagery and observation task: comparison with a conventional motor imagery task. https://doi.org/10.13140/RG.2.2.21185.17768