Hideaki Touyama

Electroencephalography (EEG)-based emotion recognition remains challenging under subject-independent evaluation because of substantial inter-subject variability. This study investigates the feasibility of state-fidelity-based quantum-kernel learning for subject-independent valence and arousal recognition using differential entropy (DE) and phase-locking value (PLV) features. Using the DEAP dataset, DE and PLV features were extracted and evaluated individually and in combination, with subject-wise normalization and F-score feature selection applied within the evaluation framework. Six features were retained for valence and eight for arousal. The selected features were encoded using a Qiskit ZFeatureMap, and the fidelity-based quantum kernel was evaluated through statevector simulation on a classical computer. Quantum-kernel, Linear, and RBF SVMs were compared using nested participant-grouped cross-validation across DE, PLV, and DE–PLV feature representations. For valence, the quantum-kernel SVM using DE features achieved the highest mean balanced accuracy of 59.18%. For arousal, the Linear SVM using DE features achieved the highest mean balanced accuracy of 54.16%. Overall, the three classifiers showed comparable performance across the evaluated feature representations. These results support the feasibility of state-fidelity-based quantum similarity for subject-independent EEG emotion recognition.

During the advantages of shorter training and higher information throughput, steady-state visual evoked potential (SSVEP) is widely used in brain–computer interface (BCI) research. Recently, collecting EEG signals from the ear area (ear-EEG) has gained increasing attention because it is more comfortable and convenient than scalp electrodes. The ear-EEG-based BCI system based on ear electrodes has weaker signals and more noise components because the electrodes are located far away from the top of the head. In this study, the RandOm Convolutional KErnel Transform (ROCKET) algorithm integrated with the Morlet wavelet transform (Morlet-ROCKET) was proposed to solve this issue. This study compared the performence of Morlet-ROCKET with two established methods: canonical correlation analysis-based (FBCCA) and Transformer methods. The proposed Morlet-ROCKET model demonstrated superior performance across multiple measures, including increased classification accuracy in 1 s, 3 s, and 4 s time windows and higher area under the curve (AUC) values in receiver operating characteristic (ROC) analysis. The analysis result proved that with efficient data processing algorithms, ear-EEG-based BCI systems can also have good performance, and providing support for the popularization of BCI.

NeuroFeedback (NF) is a method of altering the brain information of users, and it has been widely applied in the fields of medical and welfare. Recently, the research which aimed at improving cognitive function in healthy subjects has attracted a lot of attention. Many studies used EEG frequency bands and behavioral indices to improve and evaluate cognitive function. However, the evaluation of cognitive function using brain information indices mainly on the discussion of EEG frequency band. In addition, the frequency band has been mainly treated as a fixed band (alpha band et al.). In this study, we trained Upper Alpha (UA) based on individual alpha frequency using a proposed NF method, and not only confirm the changes in UA, but also examines the changes of P300. We found that training UA significantly reduced the latency of P300. In the future, we will use a control group to investigate whether the reduction of P300 latency is caused by NF and how the number of NF affects the latency of P300.

Aldosterone-producing adenomas (APAs) have different steroid profiles in serum, depending on the causative genetic mutation. Ion mobility is a separation technique for gas-phase ions based on their m/z values, shapes, and sizes. Human serum (100 µL) was purified by liquid–liquid extraction using tert-butyl methyl ether/ethyl acetate at 1/1 (v/v) and mixed with deuterium-labeled steroids as the internal standard. The separated supernatant was dried, re-dissolved in water containing 20% methanol, and injected into a liquid chromatography–ion mobility–mass spectrometer (LC/IM/MS). We established a highly sensitive assay system by separating 20 steroids based on their retention time, m/z value, and drift time. Twenty steroids were measured in the serum of patients with primary aldosteronism, essential hypertension, and healthy subjects and were clearly classified using principal component analysis. This method was also able to detect phosphatidylcholine and phosphatidylethanolamine, which were not targeted. LC/IM/MS has a high selectivity for known compounds and has the potential to provide information on unknown compounds. This analytical method has the potential to elucidate the pathogenesis of APA and identify unknown steroids that could serve as biomarkers for APA with different genetic mutations.

Grazing-incidence small-angle X-ray scattering (GISAXS) patterns have multiple superimposed contributions from the shape of the nanoscale structure, the coupling between the particles, the partial pair correlation, and the layer geometry. Therefore, it is not easy to identify the model manually from the huge amounts of combinations. The convolutional neural network (CNN), which is one of the artificial neural networks, can find regularities to classify patterns from large amounts of combinations. CNN was applied to classify GISAXS patterns, focusing on the shape of the nanoparticles. The network found regularities from the GISAXS patterns and showed a success rate of about 90% for the classification. This method can efficiently classify a large amount of experimental GISAXS patterns according to a set of model shapes and their combinations.