Oscillatory characteristics of resting-state magnetoencephalography reflect pathological and symptomatic conditions of cognitive impairment

Abstract

BackgroundDementia and mild cognitive impairment are characterised by symptoms of cognitive decline, which are typically assessed using neuropsychological assessments (NPAs), such as the Mini-Mental State Examination (MMSE) and Frontal Assessment Battery (FAB). Magnetoencephalography (MEG) is a novel clinical assessment technique that measures brain activities (summarised as oscillatory parameters), which are associated with symptoms of cognitive impairment. However, the relevance of MEG and regional cerebral blood flow (rCBF) data obtained using single-photon emission computed tomography (SPECT) has not been examined using clinical datasets. Therefore, this study aimed to investigate the relationships among MEG oscillatory parameters, clinically validated biomarkers computed from rCBF, and NPAs using outpatient data retrieved from hospital records.MethodsClinical data from 64 individuals with mixed pathological backgrounds were retrieved and analysed. MEG oscillatory parameters, including relative power (RP) from delta to high gamma bands, mean frequency, individual alpha frequency, and Shannon’s spectral entropy, were computed for each cortical region. For SPECT data, three pathological parameters—‘severity’, ‘extent’, and ‘ratio’—were computed using an easy z-score imaging system (eZIS). As for NPAs, the MMSE and FAB scores were retrieved.ResultsMEG oscillatory parameters were correlated with eZIS parameters. The eZIS parameters associated with Alzheimer’s disease pathology were reflected in theta power augmentation and slower shift of the alpha peak. Moreover, MEG oscillatory parameters were found to reflect NPAs. Global slowing and loss of diversity in neural oscillatory components correlated with MMSE and FAB scores, whereas the associations between eZIS parameters and NPAs were sparse.ConclusionMEG oscillatory parameters correlated with both SPECT (i.e. eZIS) parameters and NPAs, supporting the clinical validity of MEG oscillatory parameters as pathological and symptomatic indicators. The findings indicate that various components of MEG oscillatory characteristics can provide valuable pathological and symptomatic information, making MEG data a rich resource for clinical examinations of patients with cognitive impairments. SPECT (i.e. eZIS) parameters showed no correlations with NPAs. The results contributed to a better understanding of the characteristics of electrophysiological and pathological examinations for patients with cognitive impairments, which will help to facilitate their co-use in clinical application, thereby improving patient care.

Dorsal brain activity reflects the severity of menopausal symptoms

“…The analysis pipeline mostly followed the strategy used in a previous study 18,19 . First, continuous MEG signals were cleaned using a dual signal subspace projection algorithm 20 available on vendor-provided software (RICOH MEG Analysis).…”

“…17 The analysis pipeline mostly followed the strategy used in a previous study. 18,19 First, continuous MEG signals were cleaned using a dual signal subspace projection algorithm 20 available on vendor-provided software (RICOH MEG Analysis). Next, to remove remaining artifacts, signals were decomposed by independent component analysis (ICA) using the FastICA algorithm implemented on Brainstorm.…”

“…24 This algorithm was selected for the following reasons: (i) it is recommended as a default option in Brainstorm; (ii) we used a template brain instead of individual magnetic resonance images, which yields only rough approximations of the forward model and is not suitable for the other inversion algorithms (eg, Beamforming), requiring better model approximation than wMNE; and (iii) wMNE is used in the context of MEG and electroencephalography (EEG) source localisation for studying oscillatory characteristics of pathological conditions. 18,19,[25][26][27] Neural sources were restricted to be normal to the cortex. 18,19 The resultant data of the source reconstruction process were continuous time-series signals in each of the 15,000 cortical vertices.…”

“…18,19,[25][26][27] Neural sources were restricted to be normal to the cortex. 18,19 The resultant data of the source reconstruction process were continuous time-series signals in each of the 15,000 cortical vertices. The high dimensional data were limited to 68 anatomical regions, defined using the Desikan-Killiany atlas 28 by averaging the signals of the vertices included…”

“…While averaging, the signs of the signals were flipped in the vertices where the normal orientation were opposite to the dominant orientation of the belonging region. 18,19 Finally, to extract the oscillatory power from regional time-series signals, power spectral density (PSD) was computed using the Blackman-Tukey approach 29 by using nonoverlapping 5-second segments. The cumulative sum of the power in each of the canonical frequency band (delta [1-3 Hz], theta [4-7 Hz], alpha1 [7-9 Hz], alpha2 [9][10][11], alpha3 [11][12][13], beta [13][14][15][16][17][18][19][20][21][22][23][24][25], and gamma [low gamma, 26-40 Hz; high gamma, 41-70 Hz])referred to as band-wise oscillatory power-was computed for use in the statistical analysis.…”