Cerebrospinal fluid amyloid beta and response of cognition to a tap test in idiopathic normal pressure hydrocephalus: a case-control study.

OBJECTIVES: To examine the relationship between cerebrospinal fluid (CSF) biomarkers of Alzheimer's disease (AD) and tap test response to elucidate the effects of comorbidity of AD in idiopathic normal-pressure hydrocephalus (iNPH). DESIGN: Case-control study. SETTING: Osaka University Hospital. PARTICIPANTS: Patients with possible iNPH underwent a CSF tap test. MEASUREMENTS: Concentrations of amyloid beta (Aß) 1-40, 1-42, and total tau in CSF were measured. The response of tap test was judged using Timed Up and Go test (TUG), 10-m reciprocation walking test (10MWT), Mini-Mental State Examination (MMSE), and iNPH grading scale. The ratio of Aß1-42 to Aß1-40 (Aß42/40 ratio) and total tau concentration was compared between tap test-negative (iNPH-nTT) and -positive (iNPH-pTT) patients. RESULTS: We identified 27 patients as iNPH-nTT and 81 as iNPH-pTT. Aß42/40 ratio was significantly lower (mean [SD] = 0.063 [0.026] vs. 0.083 [0.036], p = 0.008), and total tau in CSF was significantly higher (mean [SD] = 385.6 [237.2] vs. 293.6 [165.0], p = 0.028) in iNPH-nTT than in iNPH-pTT. Stepwise logistic regression analysis revealed that low Aß42/40 ratio was significantly associated with the negativity of the tap test. The response of cognition was significantly related to Aß42/40 ratio. The association between Aß42/40 ratio and tap test response, especially in cognition, remained after adjusting for disease duration and severity at baseline. CONCLUSIONS: A low CSF Aß42/40 ratio is associated with a poorer cognitive response, but not gait and urinary response, to a tap test in iNPH. Even if CSF biomarkers suggest AD comorbidity, treatment with iNPH may be effective for gait and urinary dysfunction.

Special Issue "Machine Learning Methods for Biomedical Data Analysis".

Machine learning is an effective method for developing automatic algorithms for analysing sophisticated biomedical data [...].

Hybrid Convolutional Neural Network for Localization of Epileptic Focus Based on iEEG.

Epileptic focus localization by analysing intracranial electroencephalogram (iEEG) plays a critical role in successful surgical therapy of resection of the epileptogenic lesion. However, manual analysis and classification of the iEEG signal by clinicians are arduous and time-consuming and excessively depend on the experience. Due to individual differences of patients, the iEEG signal from different patients usually shows very diverse features even if the features belong to the same class. Accordingly, automatic detection of epileptic focus is required to improve the accuracy and to shorten the time for treatment. In this paper, we propose a novel feature fusion-based iEEG classification method, a deep learning model termed Time-Frequency Hybrid Network (TF-HybridNet), in which short-time Fourier transform (STFT) and 1d convolution layers are performed on the input iEEG in parallel to extract features of the time-frequency domain and feature maps. And then, the time-frequency features and feature maps are fused and fed to a 2d convolutional neural network (CNN). We used the Bern-Barcelona iEEG dataset for evaluating the performance of TF-HybridNet, and the experimental results show that our approach is able to differentiate the focal from nonfocal iEEG signal with an average classification accuracy of 94.3% and demonstrates an improved accuracy rate compared to the model using only STFT or one-dimensional convolutional layers as feature extraction.

Graph Eigen Decomposition-Based Feature-Selection Method for Epileptic Seizure Detection Using Electroencephalography.

Epileptic seizure is a sudden alteration of behavior owing to a temporary change in the electrical functioning of the brain. There is an urgent demand for an automatic epilepsy detection system using electroencephalography (EEG) for clinical application. In this paper, the EEG signal is divided into short time frames. Discrete wavelet transform is used to decompose each frame into a number of subbands. Different entropies as well as a group of features with which to characterize the spike events are extracted from each subband signal of an EEG frame. The features extracted from individual subbands are concatenated, yielding a high-dimensional feature vector. A discriminative subset of features is selected from the feature vector using a graph eigen decomposition (GED)-based approach. Thus, the reduced number of features obtained is effective for differentiating the underlying characteristics of EEG signals that indicate seizure events and those that indicate nonseizure events. The GED method ranks the features according to their contribution to correct classification. The selected features are used to classify seizure and nonseizure EEG signals using a feedforward neural network (FfNN). The performance of the proposed method is evaluated by conducting various experiments with a standard dataset obtained from the University of Bonn. The experimental results show that the proposed seizure-detection scheme achieves a classification accuracy of 99.55%, which is higher than that of state-of-the-art methods. The efficiency of FfNN is compared with linear discriminant analysis and support vector machine classifiers, which have classification accuracies of 98.72% and 99.39%, respectively. Hence, the proposed method is confirmed as a potential marker for EEG-based seizure detection.

Statistical Features in High-Frequency Bands of Interictal iEEG Work Efficiently in Identifying the Seizure Onset Zone in Patients with Focal Epilepsy.

The design of a computer-aided system for identifying the seizure onset zone (SOZ) from interictal and ictal electroencephalograms (EEGs) is desired by epileptologists. This study aims to introduce the statistical features of high-frequency components (HFCs) in interictal intracranial electroencephalograms (iEEGs) to identify the possible seizure onset zone (SOZ) channels. It is known that the activity of HFCs in interictal iEEGs, including ripple and fast ripple bands, is associated with epileptic seizures. This paper proposes to decompose multi-channel interictal iEEG signals into a number of subbands. For every 20 s segment, twelve features are computed from each subband. A mutual information (MI)-based method with grid search was applied to select the most prominent bands and features. A gradient-boosting decision tree-based algorithm called LightGBM was used to score each segment of the channels and these were averaged together to achieve a final score for each channel. The possible SOZ channels were localized based on the higher value channels. The experimental results with eleven epilepsy patients were tested to observe the efficiency of the proposed design compared to the state-of-the-art methods.

Apathy and right caudate perfusion in idiopathic normal pressure hydrocephalus: A case-control study.

OBJECTIVES: Apathy is prevalent in patients with idiopathic normal pressure hydrocephalus (iNPH), a treatable disorder resulting from ventricular enlargement. We assessed the relationship between apathy and regional cerebral blood flow (rCBF) in patients with iNPH. METHODS: Before lumbo-peritoneal shunt surgery (LPS), 56 iNPH patients were evaluated on apathy and dysphoria subscales of the Neuropsychiatric inventory (NPI), and were divided into two groups according to NPI apathy score: 15 without apathy (iNPH-APA) and 41 with apathy (iNPH+APA). Among iNPH+APA, 29 patients were evaluated for apathy and dysphoria 3 months after LPS, and were divided into two groups on the basis of the change in NPI apathy score: 13 with improvement (iNPH+ImpAPA) and 16 without improvement in apathy (iNPH-ImpAPA). N-isopropyl-p-iodoamphetamine single photon emission computed tomography using the autoradiography method was performed before and after LPS, and rCBF was calculated in 22 regions of interest in the frontal cortex, basal ganglia, and limbic system. RESULTS: In iNPH+APA, rCBF in the right caudate nuclei before LPS was significantly lower than that in iNPH-APA (P = 0.004; two-sample t test). Between iNPH-ImpAPA and iNPH+ImpAPA, a significant group-by-shunt interaction was observed for rCBF in only the right caudate nuclei (F1, 28  = 11.75, P = 0.002; two-way repeated-measures analysis of variance), with increased rCBF in iNPH+ImpAPA but not in iNPH-ImpAPA. The significant group-by-shunt interaction persisted if change in NPI dysphoria scores was used as a covariate (F1, 27  = 8.33, P = 0.008). CONCLUSIONS: Our findings suggest that right caudate dysfunction might cause apathy in iNPH patients.

Cerebral blood flow and Alzheimer's disease-related biomarkers in cerebrospinal fluid in idiopathic normal pressure hydrocephalus.

AIM: Alzheimer's disease (AD) pathology is highly prevalent in patients with idiopathic normal pressure hydrocephalus (iNPH), and the presence of AD pathology may involve regional cerebral blood flow (rCBF). In this study, we examined the relationship between rCBF and AD-related biomarkers in the cerebrospinal fluid of iNPH patients. METHODS: Patients with iNPH (n = 39) were classified into groups with (iNPH/AD+) (n=15) and without (iNPH/AD-) (n=24) high biomarker probability of AD (i.e. combined low amyloid ß 42 and high total tau in the cerebrospinal fluid). rCBF was quantified in 17 regions of interest by N-isopropyl-p-[123 I]iodoamphetamine single-photon emission computed tomography with the autoradiography method. We compared rCBF between the iNPH/AD- and iNPH/AD+ groups at baseline using a t-test and then compared changes in rCBF after shunt surgery between the groups using a paired t-test and two-way repeated measures ANOVA. RESULTS: At baseline, there were no significant differences in rCBF between the groups in most regions apart from the putamen. After shunt surgery, a significant increase in rCBF in the putamen, amygdala, hippocampus, and parahippocampal gyrus was observed in iNPH/AD- patients. In iNPH/AD+ patients, no significant improvement in rCBF was observed in any region. In repeated measures analysis of variance, a significant group × shunt interaction was observed in the parietal lobe, frontal lobe, posterior cingulate cortex, precuneus, lateral temporal lobe, amygdala, hippocampus, parahippocampal gyrus, and putamen. CONCLUSIONS: Improvement in rCBF after shunt surgery in iNPH/AD+ patients may be poorer than that in iNPH AD- patients.

TSH ratio as a novel diagnostic method for Cushing's syndrome.

Circadian variations impact thyrotropin (TSH) secretion; in Cushing's syndrome (CS) patients, the nocturnal serum TSH surge is abolished. The aim of this prospective study is to examine whether serum TSH surge may be a useful diagnostic method for CS. This prospective study recruited 136 inpatients for differential diagnosis of CS or subclinical CS (SCS), and 21 inpatients with depression at Osaka University Hospital. Serum TSH surge was assessed by the midnight-to-morning serum TSH ratio (2300-2400 h to 0800-0900 h). The diagnostic accuracy (sensitivity and specificity) between TSH ratio and ordinary screening tests [low-dose dexamethasone suppression test (LDDST), late-night serum cortisol and urine free cortisol (UFC)] were compared. Twenty-two patients were diagnosed as CS (12 overt CS and 10 SCS) and the remaining 120 patients were excluded for CS. The diagnostic accuracy of TSH ratio (cutoff value 1.0) yielded sensitivity 90.9% [95% confidence interval (CI) 70.8-98.9], specificity 95.0% (95% CI 89.4-98.1), and a high positive and low negative likelihood ratio [18.2 (95% CI 8.2-40.1) and 0.096 (95% CI 0.026-0.359), respectively]. The specificity of TSH ratio was significantly higher than LDDST and midnight serum cortisol test. The sensitivity of TSH ratio was significantly higher than UFC. TSH ratio showed more than 1.0 in all patients with depression and CYP3A4 inducer users. TSH ratio is a novel supportive diagnostic method with higher specificity than the current diagnostic methods for CS.

Usefulness of carer-held records to support informal caregivers of patients with dementia who live at home.

AIM: It is unclear whether carer-held records (CHR) are useful for patients with dementia. In this study, we evaluated the usefulness of the CHR for patients with dementia at the municipal level. METHODS: Candidates for CHR use in this study were informal caregivers of patients with dementia who lived at home in Kawanishi, Japan. CHR users were those who are involved in the patient's care and treatment, such as informal caregivers, family physicians, dementia specialists, care professionals, and care service coordinators, known as ?care managers' in Japan. Collaborative meetings were held every month mainly to help users, especially care managers, learn how to effectively use CHR. We surveyed informal caregivers before and 1.5 years after the start of CHR use to evaluate whether CHR improved collaboration and information provision. The Zarit Caregiver Burden Interview and Dementia Behaviour Disturbance Scale were also administered. We divided the informal caregivers who continued CHR use for 1.5 years into two subgroups based on whether their care manager attended the collaborative meetings at least twice. In addition, we divided informal caregivers into three subgroups depending on their relationship to the patient: spouse, child, or daughter-in-law. RESULTS: The study initially consisted of 201 informal caregivers. Among them, 74 informal caregivers continued CHR use for 1.5 years. The information provision score significantly improved after CHR use for all informal caregivers. The collaboration score significantly improved after CHR use only for informal caregivers whose care managers attended at least two collaborative meetings. The Zarit Caregiver Burden Interview score significantly improved after CHR use for daughter-in-law caregivers. The Dementia Behaviour Disturbance Scale scores did not significantly improve after CHR use. CONCLUSIONS: CHR were useful for informal caregivers of patients with dementia. However, care managers need to teach informal caregivers how to properly use CHR.

Chemical and structural characterization of α-N-acetylgalactosaminidase I and II from starfish, asterina amurensis.

BACKGROUND: The marine invertebrate starfish was found to contain a novel α-N-acetylgalactosaminidase, α-GalNAcase II, which catalyzes removal of terminal α-N-acetylgalactosamine (α-GalNAc), in addition to a typical α-N-acetylgalactosaminidase, α-GalNAcase I, which catalyzes removal of terminal α-N-acetylgalactosamine (α-GalNAc) and, to a lesser extent, galactose. The interrelationship between α-GalNAcase I and α-GalNAcase II and the molecular basis of their differences in substrate specificity remain unknown. RESULTS: Chemical and structural comparisons between α-GalNAcase I and II using immunostaining, N-terminal amino acid sequencing and peptide analysis showed high homology to each other and also to other glycoside hydrolase family (GHF) 27 members. The amino acid sequence of peptides showed conserved residues at the active site as seen in typical α-GalNAcase. Some substitutions of conserved amino acid residues were found in α-GalNAcase II that were located near catalytic site. Among them G171 and A173, in place of C171 and W173, respectively in α-GalNAcase were identified to be responsible for lacking intrinsic α-galactosidase activity of α-GalNAcase II. Chemical modifications supported the presence of serine, aspartate and tryptophan as active site residues. Two tryptophan residues (W16 and W173) were involved in α-galactosidase activity, and one (W16) of them was involved in α-GalNAcase activity. CONCLUSIONS: The results suggested that α-GalNAcase I and II are closely related with respect to primary and higher order structure and that their structural differences are responsible for difference in substrate specificities.

Robust Averaging of Covariances for EEG Recordings Classification in Motor Imagery Brain-Computer Interfaces.

The estimation of covariance matrices is of prime importance to analyze the distribution of multivariate signals. In motor imagery-based brain-computer interfaces (MI-BCI), covariance matrices play a central role in the extraction of features from recorded electroencephalograms (EEGs); therefore, correctly estimating covariance is crucial for EEG classification. This letter discusses algorithms to average sample covariance matrices (SCMs) for the selection of the reference matrix in tangent space mapping (TSM)-based MI-BCI. Tangent space mapping is a powerful method of feature extraction and strongly depends on the selection of a reference covariance matrix. In general, the observed signals may include outliers; therefore, taking the geometric mean of SCMs as the reference matrix may not be the best choice. In order to deal with the effects of outliers, robust estimators have to be used. In particular, we discuss and test the use of geometric medians and trimmed averages (defined on the basis of several metrics) as robust estimators. The main idea behind trimmed averages is to eliminate data that exhibit the largest distance from the average covariance calculated on the basis of all available data. The results of the experiments show that while the geometric medians show little differences from conventional methods in terms of classification accuracy in the classification of electroencephalographic recordings, the trimmed averages show significant improvement for all subjects.

Sleep disturbances are key symptoms of very early stage Alzheimer disease with behavioral and psychological symptoms: a Japan multi-center cross-sectional study (J-BIRD).

BACKGROUND: Sleep disturbances in Alzheimer disease (AD) may affect behavioral and psychological symptoms of dementia (BPSD). Our aim was to elucidate the associations between sleep disturbances and other BPSD at different stages of AD. METHODS: This investigation was part of a multicenter-retrospective study in Japan (J-BIRD). Eligible for final analyses were 684 AD patients. Global severity of dementia was estimated using the Clinical Dementia Rating (CDR) scale. BPSD were assessed using the Neuropsychiatric Inventory (NPI). We analyzed the relationships between sleep disturbances and BPSD at different stages of AD according to the CDR score. RESULTS: Among the 684 AD patients, 146 (21.3%) had sleep disturbances. Patients with very early AD (CDR 0.5) and sleep disturbances had significantly more BPSD than those without sleep disturbances, as indicated by the higher prevalence of the following four NPI items: anxiety, euphoria, disinhibition, and aberrant motor behavior. In AD at CDR 2, (moderate AD) only one NPI item (irritability) was affected, while none was affected at CDR 1 (mild AD) and 3 (severe AD). Multiple regression analyses were performed in those with AD having various CDR scores. At CDR 0.5, the presence of sleep disturbances was associated with a high total NPI score (ß = 0.32, p < 0.001). However, other factors, including cognitive decline, age, gender, and years of education, were not significantly associated with the NPI score. At CDR 1 and 2, no factor was significantly related to BPSD. CONCLUSION: Sleep disturbances were strongly associated with other BPSD in the very early stage of AD. Copyright © 2016 John Wiley & Sons, Ltd.

Transcriptome analysis of distinct mouse strains reveals kinesin light chain-1 splicing as an amyloid-ß accumulation modifier.

Alzheimer's disease (AD) is characterized by the accumulation of amyloid-ß (Aß). The genes that govern this process, however, have remained elusive. To this end, we combined distinct mouse strains with transcriptomics to directly identify disease-relevant genes. We show that AD model mice (APP-Tg) with DBA/2 genetic backgrounds have significantly lower levels of Aß accumulation compared with SJL and C57BL/6 mice. We then applied brain transcriptomics to reveal the genes in DBA/2 that suppress Aß accumulation. To avoid detecting secondarily affected genes by Aß, we used non-Tg mice in the absence of Aß pathology and selected candidate genes differently expressed in DBA/2 mice. Additional transcriptome analysis of APP-Tg mice with mixed genetic backgrounds revealed kinesin light chain-1 (Klc1) as an Aß modifier, indicating a role for intracellular trafficking in Aß accumulation. Aß levels correlated with the expression levels of Klc1 splice variant E and the genotype of Klc1 in these APP-Tg mice. In humans, the expression levels of KLC1 variant E in brain and lymphocyte were significantly higher in AD patients compared with unaffected individuals. Finally, functional analysis using neuroblastoma cells showed that overexpression or knockdown of KLC1 variant E increases or decreases the production of Aß, respectively. The identification of KLC1 variant E suggests that the dysfunction of intracellular trafficking is a causative factor of Aß pathology. This unique combination of distinct mouse strains and model mice with transcriptomics is expected to be useful for the study of genetic mechanisms of other complex diseases.

Association of cerebrospinal fluid tap-related oscillatory activity and shunt outcome in idiopathic normal-pressure hydrocephalus.

BACKGROUND: Idiopathic normal-pressure hydrocephalus (iNPH) is a neuropsychiatric syndrome characterized by the clinical triad of gait disturbance, urinary dysfunction, and cognitive impairment. The aim of the present study was to find specific EEG patterns associated with shunt response in iNPH. METHODS: Twenty five iNPH patients (10 shunt responders and 15 non-responders) were enrolled in this study. We performed current source density (CSD) analysis in several frequency bands (delta: 2-4 Hz, theta: 4-8 Hz, alpha: 8-13 Hz, beta: 13-30 Hz, gamma: 30-60 Hz) using exact Low Resolution Brain Electromagnetic Tomography (eLORETA). CSD distribution was compared between shunt responders and non-responders for each frequency band before and after CSF tap test. RESULTS: Shunt responders showed increased gamma CSD in the left temporal cortex before CSF tapping relative to non-responders. However, after CSF tapping, shunt response was associated with significantly higher CSDs in several frequency bands, specifically theta, alpha, beta and gamma, involving mainly the frontal and temporal areas. Using eLORETA analysis, we were able to identify cortical oscillatory activity before and after CSF tap test related to clinical recovery due to shunt operation in iNPH. CONCLUSION: Our findings support and extend the results of previous studies examining the effects of CSF tap test and shunt operation in patients with iNPH, possibly indicating electrophysiological features of shunt response in this disease. These findings warrant future studies to use EEG for prediction of shunt response in iNPH.

In vitro acetylcholinesterase inhibitory activity and the antioxidant properties of Aegle marmelos leaf extract: implications for the treatment of Alzheimer's disease.

BACKGROUND: Alzheimer's disease (AD) is a progressive neurodegenerative disorder clinically characterized by loss of memory and cognition. The effective therapeutic options for AD are limited and thus there is a demand for new drugs. Aegle marmelos (Linn.) (A. marmelos) leaves have been used in traditional medicine to promote intellect and enhance memory. In this study, we evaluated A. marmelos for its acetylcholinesterase (AChE) inhibitory activity and antioxidant property in vitro in the treatment of AD. METHODS: A crude methanol extract and four fractions (petroleum ether, chloroform, ethyl acetate and aqueous) were prepared from the leaves of A. marmelos. The preparations were assessed for AChE inhibitory activity by the Ellman method, and their antioxidant properties were assessed by several assays: reducing power, scavenging of 1,1-diphenyl-2-picrylhydrazyl free radical and hydroxyl radical, and inhibition of lipid peroxidation. Qualitative and quantitative analyses of endogenous substances in A. marmelos were performed by the standard phytochemical methods. RESULTS: Among the different extracts tested, the ethyl acetate fraction exhibited the highest inhibition of AChE activity. In the same way, ethyl acetate fraction showed the highest reducing activity and radical scavenging ability towards the 1,1-diphenyl-2-picrylhydrazyl (half maximal inhibitory concentration = 3.84 µg/mL) and hydroxyl free radicals (half maximal inhibitory concentration = 5.68 µg/mL). The antiradical activity of the ethyl acetate fraction appeared to be similar to that of the reference standard butylated hydroxytoluene and catechin used in this study. In addition, the ethyl acetate fraction displayed higher inhibition of brain lipid peroxidation. Phytochemical screening of different extractives of A. marmelos showed the presence of phenols and flavonoids, alkaloid, saponin, glycoside, tannin and steroids. Quantitative analysis revealed higher contents of phenolics (58.79-mg gallic acid equivalent/g dried extract) and flavonoids (375.73-mg gallic acid equivalent/g dried extract) in the ethyl acetate fraction. CONCLUSION: The results suggest that the ethyl acetate fraction of A. marmelos is a significant source of polyphenolic compounds with potential AChE inhibitory property and antioxidant activity and, thus, may be useful in the treatment of AD.

Nonparametric tensor ring decomposition with scalable amortized inference.

Multi-dimensional data are common in many applications, such as videos and multi-variate time series. While tensor decomposition (TD) provides promising tools for analyzing such data, there still remains several limitations. First, traditional TDs assume multi-linear structures of the latent embeddings, which greatly limits their expressive power. Second, TDs cannot be straightforwardly applied to datasets with massive samples. To address these issues, we propose a nonparametric TD with amortized inference networks. Specifically, we establish a non-linear extension of tensor ring decomposition, using neural networks, to model complex latent structures. To jointly model the cross-sample correlations and physical structures, a matrix Gaussian process (GP) prior is imposed over the core tensors. From learning perspective, we develop a VAE-like amortized inference network to infer the posterior of core tensors corresponding to new tensor data, which enables TDs to be applied to large datasets. Our model can be also viewed as a kind of decomposition of VAE, which can additionally capture hidden tensor structure and enhance the expressiveness power. Finally, we derive an evidence lower bound such that a scalable optimization algorithm is developed. The advantages of our method have been evaluated extensively by data imputation on the Healing MNIST dataset and four multi-variate time series data.

Alpha band modulation caused by selective attention to music enables EEG classification.

Humans are able to pay selective attention to music or speech in the presence of multiple sounds. It has been reported that in the speech domain, selective attention enhances the cross-correlation between the envelope of speech and electroencephalogram (EEG) while also affecting the spatial modulation of the alpha band. However, when multiple music pieces are performed at the same time, it is unclear how selective attention affects neural entrainment and spatial modulation. In this paper, we hypothesized that the entrainment to the attended music differs from that to the unattended music and that spatial modulation in the alpha band occurs in conjunction with attention. We conducted experiments in which we presented musical excerpts to 15 participants, each listening to two excerpts simultaneously but paying attention to one of the two. The results showed that the cross-correlation function between the EEG signal and the envelope of the unattended melody had a more prominent peak than that of the attended melody, contrary to the findings for speech. In addition, the spatial modulation in the alpha band was found with a data-driven approach called the common spatial pattern method. Classification of the EEG signal with a support vector machine identified attended melodies and achieved an accuracy of 100% for 11 of the 15 participants. These results suggest that selective attention to music suppresses entrainment to the melody and that spatial modulation of the alpha band occurs in conjunction with attention. To the best of our knowledge, this is the first report to detect attended music consisting of several types of music notes only with EEG.

Causal Connectivity Network Analysis of Ictal Electrocorticogram With Temporal Lobe Epilepsy Based on Dynamic Phase Transfer Entropy.

Temporallobe epilepsy (TLE) has been conceptualized as a brain network disease, which generates brain connectivity dynamics within and beyond the temporal lobe structures in seizures. The hippocampus is a representative epileptogenic focus in TLE. Understanding the causal connectivity in terms of brain network during seizures is crucial in revealing the triggering mechanism of epileptic seizures originating from the hippocampus (HPC) spread to the lateral temporal cortex (LTC) by ictal electrocorticogram (ECoG), particularly in high-frequency oscillations (HFOs) bands. In this study, we proposed the unified-epoch dynamic causality analysis method to investigate the causal influence dynamics between two brain regions (HPC and LTC) at interictal and ictal phases in the frequency range of 1-500 Hz by introducing the phase transfer entropy (PTE) out/in-ratio and sliding window. We also proposed PTE-based machine learning algorithms to identify epileptogenic zone (EZ). Nine patients with a total of 26 seizures were included in this study. We hypothesized that: 1) HPC is the focus with the stronger causal connectivity than that in LTC in the ictal state at gamma and HFOs bands. 2) Causal connectivity in the ictal phase shows significant changes compared to that in the interictal phase. 3) The PTE out/in-ratio in the HFOs band can identify the EZ with the best prediction performance.

Modeling Complex EEG Data Distribution on the Riemannian Manifold Toward Outlier Detection and Multimodal Classification.

OBJECTIVE: The usage of Riemannian geometry for Brain-computer interfaces (BCIs) has gained momentum in recent years. Most of the machine learning techniques proposed for Riemannian BCIs consider the data distribution on a manifold to be unimodal. However, the distribution is likely to be multimodal rather than unimodal since high-data variability is a crucial limitation of electroencephalography (EEG). In this paper, we propose a novel data modeling method for considering complex data distributions on a Riemannian manifold of EEG covariance matrices, aiming to improve BCI reliability. METHODS: Our method, Riemannian spectral clustering (RiSC), represents EEG covariance matrix distribution on a manifold using a graph with proposed similarity measurement based on geodesic distances, then clusters the graph nodes through spectral clustering. This allows flexibility to model both a unimodal and a multimodal distribution on a manifold. RiSC can be used as a basis to design an outlier detector named outlier detection Riemannian spectral clustering (odenRiSC) and a multimodal classifier named multimodal classifier Riemannian spectral clustering (mcRiSC). All required parameters of odenRiSC/mcRiSC are selected in data-driven manner. Moreover, there is no need to pre-set a threshold for outlier detection and the number of modes for multimodal classification. RESULTS: The experimental evaluation revealed odenRiSC can detect EEG outliers more accurately than existing methods and mcRiSC outperformed the standard unimodal classifier, especially on high-variability datasets. CONCLUSION: odenRiSC/mcRiSC are anticipated to contribute to making real-life BCIs outside labs and neuroergonomics applications more robust. SIGNIFICANCE: RiSC can work as a robust EEG outlier detector and multimodal classifier.

Feasibility of decoding covert speech in ECoG with a Transformer trained on overt speech.

Several attempts for speech brain-computer interfacing (BCI) have been made to decode phonemes, sub-words, words, or sentences using invasive measurements, such as the electrocorticogram (ECoG), during auditory speech perception, overt speech, or imagined (covert) speech. Decoding sentences from covert speech is a challenging task. Sixteen epilepsy patients with intracranially implanted electrodes participated in this study, and ECoGs were recorded during overt speech and covert speech of eight Japanese sentences, each consisting of three tokens. In particular, Transformer neural network model was applied to decode text sentences from covert speech, which was trained using ECoGs obtained during overt speech. We first examined the proposed Transformer model using the same task for training and testing, and then evaluated the model's performance when trained with overt task for decoding covert speech. The Transformer model trained on covert speech achieved an average token error rate (TER) of 46.6% for decoding covert speech, whereas the model trained on overt speech achieved a TER of 46.3% ( p > 0.05 ; d = 0.07 ) . Therefore, the challenge of collecting training data for covert speech can be addressed using overt speech. The performance of covert speech can improve by employing several overt speeches.