Deep learning based automatic detection and dipole estimation of epileptic discharges in MEG: a multi-center study.

Magnetoencephalography (MEG) provides crucial information in diagnosing focal epilepsy. However, dipole estimation, a commonly used analysis method for MEG, can be time-consuming since it necessitates neurophysiologists to manually identify epileptic spikes. To reduce this burden, we developed the automatic detection of spikes using deep learning in single center. In this study, we performed a multi-center study using six MEG centers to improve the performance of the automated detection of neuromagnetically recorded epileptic spikes, which we previously developed using deep learning. Data from four centers were used for training and evaluation (internal data), and the remaining two centers were used for evaluation only (external data). We used a five-fold subject-wise cross-validation technique to train and evaluate the models. A comparison showed that the multi-center model outperformed the single-center model in terms of performance. The multi-center model achieved an average ROC-AUC of 0.9929 and 0.9426 for the internal and external data, respectively. The median distance between the neurophysiologist-analyzed and automatically analyzed dipoles was 4.36 and 7.23 mm for the multi-center model for internal and external data, respectively, indicating accurate detection of epileptic spikes. By training data from multiple centers, automated analysis can improve spike detection and reduce the analysis workload for neurophysiologists. This study suggests that the multi-center model has the potential to detect spikes within 1 cm of a neurophysiologist's analysis. This multi-center model can significantly reduce the number of hours required by neurophysiologists to detect spikes.

Two Distinct Neural Mechanisms Underlying Acupuncture Analgesia.

Acupuncture analgesia is a traditional treatment with a long history, although it lacks scientific evidence. It is reportedly associated with the central nervous system, including various brain regions, from the cortices to the brain stem. However, it remains unclear whether the distributed regions behave as a single unit or consist of multiple sub-units playing different roles. Magnetoencephalography is a neuroimaging technique that can measure the oscillatory frequency of neural signals and brain regions. The frequency band of neural signals allows further understanding of the characteristics of the acupuncture-related neural systems. This study measured resting-state brain activity using magnetoencephalography in 21 individuals with chronic pain before and after acupuncture treatment. The subjective level of pain was assessed using a visual analog scale, and brain activity was compared to identify the brain regions and the frequencies associated with acupuncture analgesia. Here, we categorized the changes in resting-state brain activity into two groups: low-frequency oscillatory activity (<3 Hz) in the left middle occipital and right superior partial lobule and high-frequency oscillatory activity (81-120 Hz) on both sides of the prefrontal, primary sensory, and right fusiform gyri. These findings suggest that acupuncture analgesia influences two or more sub-units of the neural systems, which helps us understand the neural mechanisms underlying acupuncture analgesia.

Non-pharmacological treatment changes brain activity in patients with dementia.

Non-pharmacological treatment (NPT) improves cognitive functions and behavioural disturbances in patients with dementia, but the underlying neural mechanisms are unclear. In this observational study, 21 patients with dementia received NPTs for several months. Patients were scanned using magnetoencephalography twice during the NPT period to evaluate NPT effects on resting-state brain activity. Additionally, cognitive functions and behavioural disturbances were measured using the Mini-Mental State Examination (MMSE-J) and a short version of the Dementia Behaviour Disturbance Scale (DBD-13) at the beginning and the end of the NPT period. In contrast to the average DBD-13 score, the average MMSE-J score improved after the NPT period. Magnetoencephalography data revealed a reduced alpha activity in the right temporal lobe and fusiform gyrus, as well as an increased low-gamma activity in the right angular gyrus. DBD-13 score changes were correlated with beta activity in the sensorimotor area. These findings corroborate previous studies confirming NPT effects on brain activity in healthy participants and people at risk of dementia. Our results provide additional evidence that brains of patients with dementia have the capacity for plasticity, which may be responsible for the observed NPT effects. In dementia, NPT might lead to improvements in the quality of life.

MEG Frequency Analysis Depicts the Impaired Neurophysiological Condition of Ischemic Brain.

PURPOSE: Quantitative imaging of neuromagnetic fields based on automated region of interest (ROI) setting was analyzed to determine the characteristics of cerebral neural activity in ischemic areas. METHODS: Magnetoencephalography (MEG) was used to evaluate spontaneous neuromagnetic fields in the ischemic areas of 37 patients with unilateral internal carotid artery (ICA) occlusive disease. Voxel-based time-averaged intensity of slow waves was obtained in two frequency bands (0.3-4 Hz and 4-8 Hz) using standardized low-resolution brain electromagnetic tomography (sLORETA) modified for a quantifiable method (sLORETA-qm). ROIs were automatically applied to the anterior cerebral artery (ACA), anterior middle cerebral artery (MCAa), posterior middle cerebral artery (MCAp), and posterior cerebral artery (PCA) using statistical parametric mapping (SPM). Positron emission tomography with 15O-gas inhalation (15O-PET) was also performed to evaluate cerebral blood flow (CBF) and oxygen extraction fraction (OEF). Statistical analyses were performed using laterality index of MEG and 15O-PET in each ROI with respect to distribution and intensity. RESULTS: MEG revealed statistically significant laterality in affected MCA regions, including 4-8 Hz waves in MCAa, and 0.3-4 Hz and 4-8 Hz waves in MCAp (95% confidence interval: 0.020-0.190, 0.030-0.207, and 0.034-0.213), respectively. We found that 0.3-4 Hz waves in MCAp were highly correlated with CBF in MCAa and MCAp (r = 0.74, r = 0.68, respectively), whereas 4-8 Hz waves were moderately correlated with CBF in both the MCAa and MCAp (r = 0.60, r = 0.63, respectively). We also found that 4-8 Hz waves in MCAp were statistically significant for misery perfusion identified on 15O-PET (p<0.05). CONCLUSIONS: Quantitatively imaged spontaneous neuromagnetic fields using the automated ROI setting enabled clear depiction of cerebral ischemic areas. Frequency analysis may reveal unique neural activity that is distributed in the impaired vascular metabolic territory, in which the cerebral infarction has not yet been completed.

Disturbed resting functional inter-hemispherical connectivity of the ventral attentional network in alpha band is associated with unilateral spatial neglect.

Unilateral spatial neglect (USN) is one of the most common symptoms of right hemisphere damage; its classical symptom is that patients fail to respond to information on their left side. It has been postulated that disturbance of 2 separate attentional networks relates to the occurrence of USN. However, little is known about the underlying mechanism and neuronal substrates. In this study, we measured spontaneous neural activity by means of magnetoencephalography in 13 patients with brain damage and 5 control subjects. To study the relationship between functional connectivity at rest and severity of USN symptoms, we determined the imaginary coherence values relating to the inter-hemispherical ventral and dorsal attentional networks, as well as the clinical severity of USN using neuropsychological tests and behavioral rating scales. The present results showed that inter-hemispherical connectivity in the ventral attentional network, especially between the left and right angular gyri, detected in the alpha band is associated with the severity of USN symptoms. This may suggest that connectivity of inter-hemispherical homologous regions of the ventral attentional network in the alpha band could be one of the biomarkers of attentional network imbalance occurring in patients with USN.