Neural Correlates of Finger Tapping Performance in Alzheimer's Disease

328 

Abstract Submission 

Vincent Koppelmans¹, Sarah Cote², Kevin Duff³

¹University of Utah, Salt Lake City, UT, ²Yeshiva University, New York, NY, ³Oregon Health & Science University, Portland, OR

Vincent Koppelmans, PhD  
University of Utah
Salt Lake City, UT

Sarah Cote  
Yeshiva University
New York, NY

Kevin Duff  
Oregon Health & Science University
Portland, OR

Finger tapping performance and fine motor skill can be impaired in Mild Cognitive Impairment (MCI) and Alzheimer's Disease (AD). However, the neural mechanisms behind these impairments are largely unknown. We examined if unimanual and bimanual finger tapping performance relates to white matter microstructure using diffusion weighted imaging.

One hundred and three subjects (57 females; mean age 74.4 years; 50 cognitively intact, 29 MCI, and 24 AD) completed a computerized finger tapping test and an MRI scan. The computerized finger tapping test comprised unimanual (dominant and non-dominant hand) tapping, synchronous bimanual tapping, and alternate bimanual tapping. Outcome measures included initial reaction time, tapping speed, and variance. A T1-weighted MP2RAGE scan (1mm isotropic) and a diffusion weighted scan (one b=0 s/mm2 volume, 64 volumes with b=3000 s/mm2, 1.5mm isotropic) were collected on a 3T Siemens Prisma scanner with a 64 channel head coil.

MRI data were converted to BIDS format using BIDSkit and were subsequently pre-processed with fastsurfer and qsiprep v0.19.0, which was set to perform the following steps: brainmask creation, T1 to MNI registration, diffusion denoising, intensity normalization, B1 field inhomogeneity correction, Eddy current and head motion correction, outlier replacement, and resampling to ACPC-space with 1.2mm isotropic voxels.

Reconstruction was performed using MRtrix3 in the framework of qsiprep: Multi-tissue fiber response functions were estimated using the Dhollander algorithm. Fiber orientation distributions were estimated via constrained spherical deconvolution using an unsupervised multi-tissue method. A single-shell-optimized multi-tissue constrained spherical deconvolution was performed. Fiber orientation distributions were intensity-normalized. Whole brain connectivity was then tracked using the fiber orientation distribution with the fastsurfer gray-matter white-matter boundary as a constraint. Finally, ROI-to-ROI connectivity was obtained using the brainnetome atlas with 246 atlas. Because of lack of coverage of the cerebellum in our diffusion scans for the majority of subjects, cerebellar regions were excluded from analyses. Here, we analyze the connectivity between ROIs defined as the apparent Fiber density scaled by the size of the ROIs.

Connectivity matrices were fed into the Network Based Statistics toolbox to a) analyze differences in the extent of network connectivity between cognitively normal, MCI and AD subjects; and b) analyze associations between structural connectivity and finger tapping performance collapse across the three experimental groups. A significant threshold of T=2.5 was set for selecting individual edges to be included in the network analysis. Analyses were adjusted for age and sex and significance of networks was adjusted for using family-wise error correction.

Significant network differences were observed between cognitively intact and AD subjects (p<.001) and between MCI and AD subjects (p=.015), but no differences were observed between MCI subjects and the other two groups. These networks spanned almost the entire brain, including 95% and 88% of the ROIs in the atlas respectively. Network connectivity strength was significantly associated with finger tapping speed for all four tapping conditions (p=.002-.014), with networks spanning 79% (for the non-dominant hand) to 96% (for the dominant hand) of the ROIs. No associations with tapping variability in tapping speed or initial reaction time were found.

Network structural connectivity is affected in AD, but not yet in MCI. Additionally, it is related to finger tapping speed, but not variability or reaction time. These results suggest that one explanation for impaired fine motor skill in AD is global reductions in network strength due to white matter pathology. These preliminary findings deserve further investigation into the neural mechanisms of this motor impairment.

Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) ¹

Motor Behavior Other ²

Degenerative Disease

Motor

Tractography

White Matter

Other - Alzheimer