Non-primary Motor Cortical involvement in Reaching Behavior after Stroke: a TMS + MRI study
Poster No:
2067
Submission Type:
Abstract Submission
Authors:
George Wittenberg1, Jennifer Mak1, Golnaz Haddadshargh2, Fang Liu1, Jennifer Collinger1
Institutions:
1University of Pittsburgh, Pittsburgh, PA, 2University of Ptitsburgh, Pittsburgh, PA
First Author:
Co-Author(s):
Introduction:
Motor impairments following a stroke cause decreased quality of life. Our overall goal is to improve motor function after stroke by combining neuromodulation and motor task practice. First, we must understand the neural mechanisms underlying voluntary arm movements, particularly reaching. It remains unclear how (or whether) non-primary motor areas that support reaching change their connections to M1 after stroke. Here, we investigate premotor and parietal connectivity during reaching by measuring resting state and task-related BOLD responses, and by disrupting these areas with transcranial magnetic stimulation (TMS) in patients with small subcortical strokes and in healthy participants. The results will provide the foundations for neuromodulatory strategies to maximize therapeutic outcomes.
Methods:
Resting-state fMRI (rs-fMRI) and task-related fMRI (tr-fMRI focused on functional connectivity among bilateral dorsal premotor cortex (PMd), ventral premotor cortex (PMv), posterior parietal cortex (PP) and primary motor cortex (M1). FMRI used 3T Siemens Prisma scanners with a repetition time (TR) = 0.8 seconds on 21 subjects. The rs-fMRI data was acquired over two 9-minute runs while the subjects kept their eyes open, and the tr-fMRI data was recorded while the subjects performed 60 trials of 10-second planning and hand-movement tasks, which in total take around 10 minutes.
Preprocessing and statistical analysis were completed with FSL and SPM12 to estimate the functional connectivity among our regions of interest (ROIs). For rs-fMRI data, we utilized independent component analysis (ICA) and dual regression to explore the seed-based correlation and generate spatial maps at p < 0.001 statistical significance level. For tr-fMRI data, we developed generalized linear models (GLM) to extract the stimulus-induced signals and separated those signals based on events to create activation maps at a corrected cluster significance threshold of p = 0.05. Then we created dynamic causal models (DCM) to study the effective connectivity among the ROIs.
For TMS experiments, participants performed planar forward and backward reaches with their right or impaired hand inside a robotic exoskeleton. We delivered triple-pulse 10 Hz TMS 100ms before the reaction time at 80% resting motor threshold (RMT), 120% RMT, and no stimulation over 7 locations: left and right PMd, PPC and the post-central midline (control).
Preprocessing and statistical analysis were completed with FSL and SPM12 to estimate the functional connectivity among our regions of interest (ROIs). For rs-fMRI data, we utilized independent component analysis (ICA) and dual regression to explore the seed-based correlation and generate spatial maps at p < 0.001 statistical significance level. For tr-fMRI data, we developed generalized linear models (GLM) to extract the stimulus-induced signals and separated those signals based on events to create activation maps at a corrected cluster significance threshold of p = 0.05. Then we created dynamic causal models (DCM) to study the effective connectivity among the ROIs.
For TMS experiments, participants performed planar forward and backward reaches with their right or impaired hand inside a robotic exoskeleton. We delivered triple-pulse 10 Hz TMS 100ms before the reaction time at 80% resting motor threshold (RMT), 120% RMT, and no stimulation over 7 locations: left and right PMd, PPC and the post-central midline (control).
Results:
18 normal participants and 5 stroke-affected individuals (all internal capsule stroke) have been studied. Resting State Connectivity: Analysis of normal participants showed expected correlations among non-primary motor areas (including PP) and M1. Stroke participants, in general, showed reduced connectivity, particularly between hemispheres. Task-based DCM: We focused on each region's connection strength with M1. The method separated connectivity in different phases of the task: rest, plan, and move. The majority of significant connections were negative and in the rest phase. The stroke participants showed differences in connectivity as compared to the normal participants, and particularly in positive connections during the plan and move phases. But overall, the significance of the connections was low.
TMS: In healthy controls, PMd stimulation contralateral to the hand increased the length of the total movement, endpoint error, and initial angle and shortened the distance traveled at the time of maximum velocity. The stroke group had variable outcomes. In two patients, stimulation at RPPC, RPMv, and LPMd perturbed right hand movements and increased the initial movement burst.
TMS: In healthy controls, PMd stimulation contralateral to the hand increased the length of the total movement, endpoint error, and initial angle and shortened the distance traveled at the time of maximum velocity. The stroke group had variable outcomes. In two patients, stimulation at RPPC, RPMv, and LPMd perturbed right hand movements and increased the initial movement burst.
·MRI results
·TMS results
Conclusions:
Connectivity measures were overall consistent between RS-fMRI, DCM, and TMS, although the DCM method was problematic in terms of statistical significance. TMS data demonstrate that the PMd to M1 connection remains effective even into the execution phase of reaching and it, and other connections, are altered after even small subcortical strokes.
Brain Stimulation:
TMS 2
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
Motor Behavior:
Motor Planning and Execution 1
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
Cerebrovascular Disease
Motor
Transcranial Magnetic Stimulation (TMS)
1|2Indicates the priority used for review