Page 90 - MEGIN Book Of Abstracts - 2023

(UNMC), Omaha, NE, USA; Memory Disorders and Behavioral Detection of Mild Cognitive Impairment with MEG
Neurology Program, UNMC, Omaha, NE, USA Functional Connectivity Using Wavelet-Based
Neuromarkers (2021)
BACKGROUND Alzheimer's disease (AD) is generally
thought to spare primary sensory function; however, Yang, Su; Bornot, Jose Miguel Sanchez; Fernandez,
such interpretations have drawn from a literature that Ricardo Bruña; Deravi, Farzin; Hoque, Sanaul; Wong-Lin,
has rarely taken into account the variable cognitive KongFatt; Prasad, Girijesh
declines seen in patients with AD. As these cognitive
domains are now known to modulate cortical somato- Department of Computer Science, Swansea University, Swan-
sensory processing, it remains possible that abnormali- sea SA1 8EN, UK; Intelligent Systems Research Centre, School
ties in somatosensory function in patients with AD of Computing, Engineering and Intelligent Systems, Ulster
have been suppressed by neuropsychological variabil- University, Londonderry BT48 7JL, UK; Centre for Biomedical
ity in previous research. Technology, Technical University of Madrid, 28223 Madrid,
Spain; School of Engineering, University of Kent, Canterbury
METHODS In this study, we combine magnetoenceph- CT2 7NZ, UK
alographic (MEG) brain imaging during a paired-pulse
somatosensory gating task with an extensive battery ABSTRACT Studies on developing effective neuro-
of neuropsychological tests to investigate the influence markers based on magnetoencephalographic (MEG)
of cognitive variability on estimated differences in so- signals have been drawing increasing attention in the
matosensory function between biomarker-confirmed neuroscience community. This study explores the idea
patients on the AD spectrum and cognitively-normal of using source-based magnitude-squared spectral
older adults. coherence as a spatial indicator for effective regions of
interest (ROIs) localization, subsequently discriminating
FINDINGS We show that patients on the AD spectrum the participants with mild cognitive impairment (MCI)
exhibit largely non-significant differences in somato- from a group of age-matched healthy control (HC)
sensory function when cognitive variability is not elderly participants. We found that the cortical regions
considered (p-value range: .020-.842). However, once could be divided into two distinctive groups based
attention and processing speed abilities are considered, on their coherence indices. Compared to HC, some
robust differences in gamma-frequency somatosensory ROIs showed increased connectivity (hyper-connected
response amplitude (p < .001) and gating (p = .004) ROIs) for MCI participants, whereas the remaining ROIs
emerge, accompanied by significant statistical suppres- demonstrated reduced connectivity (hypo-connected
sion effects. ROIs). Based on these findings, a series of wavelet-
based source-level neuromarkers for MCI detection are
INTERPRETATION These findings suggest that patients proposed and explored, with respect to the two distinc-
with AD exhibit insults to functional somatosensory tive ROI groups. It was found that the neuromarkers
processing in primary sensory cortices, but these extracted from the hyper-connected ROIs performed
effects are masked by variability in cognitive decline significantly better for MCI detection than those from
across individuals. the hypo-connected ROIs. The neuromarkers were clas-
sified using support vector machine (SVM) and k-NN
FUNDING National Institutes of Health, USA; Fremont classifiers and evaluated through Monte Carlo cross-
Area Alzheimer's Fund, USA. validation. An average recognition rate of 93.83% was
obtained using source-reconstructed signals from the
Keywords: Amyloid-β, Gamma oscillations, Magnetoen- hyper-connected ROI group. To better conform to clini-
cephalography, Neuropsychology, Sensory gating cal practice settings, a leave-one-out cross-validation
(LOOCV) approach was also employed to ensure that
EBioMedicine (2021), Vol. 73 (34689085) (5 citations) the data for testing was from a participant that the clas-

 ontents  Index 69
C