ADFCNN: Attention-Based Dual-Scale Fusion Convolutional Neural Network for Motor Imagery Brain-Computer Interface.

Convolutional neural networks (CNNs) have been successfully applied to motor imagery (MI)-based brain-computer interface (BCI). Nevertheless, single-scale CNN fail to extract abundant information over a wide spectrum from EEG signals, while typical multi-scale CNNs cannot effectively fuse information from different scales with concatenation-based methods. To overcome these challenges, we propose a new scheme equipped with attention-based dual-scale fusion convolutional neural network (ADFCNN), which jointly extracts and fuses EEG spectral and spatial information at different scales. This scheme also provides novel insight through self-attention for effective information fusion from different scales. Specifically, temporal convolutions with two different kernel sizes identify EEG µ and ß rhythms, while spatial convolutions at two different scales generate global and detailed spatial information, respectively, and the self-attention mechanism performs feature fusion based on the internal similarity of the concatenated features extracted by the dual-scale CNN. The proposed scheme achieves the superior performance compared with state-of-the-art methods in subject-specific motor imagery recognition on BCI Competition IV dataset 2a, 2b and OpenBMI dataset, with the cross-session average classification accuracies of 79.39% and significant improvements of 9.14% on BCI-IV2a, 87.81% and 7.66% on BCI-IV2b, 65.26% and 7.2% on OpenBMI dataset, and the within-session average classification accuracies of 86.87% and significant improvements of 10.89% on BCI-IV2a, 87.26% and 8.07% on BCI-IV2b, 84.29% and 5.17% on OpenBMI dataset, respectively. What is more, ablation experiments are conducted to investigate the mechanism and demonstrate the effectiveness of the dual-scale joint temporal-spatial CNN and self-attention modules. Visualization is also used to reveal the learning process and feature distribution of the model.

Compact Artificial Neural Network Based on Task Attention for Individual SSVEP Recognition With Less Calibration.

OBJECTIVE: Recently, artificial neural networks (ANNs) have been proven effective and promising for the steady-state visual evoked potential (SSVEP) target recognition. Nevertheless, they usually have lots of trainable parameters and thus require a significant amount of calibration data, which becomes a major obstacle due to the costly EEG collection procedures. This paper aims to design a compact network that can avoid the over-fitting of the ANNs in the individual SSVEP recognition. METHOD: This study integrates the prior knowledge of SSVEP recognition tasks into the attention neural network design. First, benefiting from the high model interpretability of the attention mechanism, the attention layer is applied to convert the operations in conventional spatial filtering algorithms to the ANN structure, which reduces network connections between layers. Then, the SSVEP signal models and the common weights shared across stimuli are adopted to design constraints, which further condenses the trainable parameters. RESULTS: A simulation study on two widely-used datasets demonstrates the proposed compact ANN structure with proposed constraints effectively eliminates redundant parameters. Compared to existing prominent deep neural network (DNN)-based and correlation analysis (CA)-based recognition algorithms, the proposed method reduces the trainable parameters by more than 90% and 80% respectively, and boosts the individual recognition performance by at least 57% and 7% respectively. CONCLUSION: Incorporating the prior knowledge of task into the ANN can make it more effective and efficient. The proposed ANN has a compact structure with less trainable parameters and thus requires less calibration with the prominent individual SSVEP recognition performance.

Enhancing Detection of Multi-Frequency-Modulated SSVEP Using Phase Difference Constrained Canonical Correlation Analysis.

OBJECTIVE: Multi-frequency-modulated visual stimulation scheme has been shown effective for the steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) recently, especially in increasing the visual target number with less stimulus frequencies and mitigating the visual fatigue. However, the existing calibration-free recognition algorithms based on the traditional canonical correlation analysis (CCA) cannot provide the merited performance. APPROACH: To improve the recognition performance, this study proposes a phase difference constrained CCA (pdCCA), which assumes that the multi-frequency-modulated SSVEPs share a common spatial filter over different frequencies and have a specified phase difference. Specifically, during the CCA computation, the phase differences of the spatially filtered SSVEPs are constrained using the temporal concatenation of the sine-cosine reference signals with the pre-defined initial phases. MAIN RESULTS: We evaluate the performance of the proposed pdCCA-based method on three representative multi-frequency-modulated visual stimulation paradigms (i.e., based on the multi-frequency sequential coding, the dual-frequency, and the amplitude modulation). The evaluation results on four SSVEP datasets (Dataset Ia, Ib, II, and III) show that the pdCCA-based method can significantly outperform the current CCA method in terms of recognition accuracy. It improves the accuracy by 22.09% in Dataset Ia, 20.86% in Dataset Ib, 8.61% in Dataset II, and 25.85% in Dataset III. SIGNIFICANCE: The pdCCA-based method, which actively controls the phase difference of the multi-frequency-modulated SSVEPs after spatial filtering, is a new calibration-free method for multi-frequency-modulated SSVEP-based BCIs.

Stimulus-Stimulus Transfer Based on Time-Frequency-Joint Representation in SSVEP-Based BCIs.

OBJECTIVE: Brain-computer interfaces (BCIs) based on steady-state visual evoked potential (SSVEP) require extensive and costly calibration to achieve high performance. Using transfer learning to re-use existing calibration data from old stimuli is a promising strategy, but finding commonalities in the SSVEP signals across different stimuli remains a challenge. METHOD: This study presents a new perspective, namely time-frequency-joint representation, in which SSVEP signals corresponding to different stimuli can be synchronized, and thus can emphasize common components. According to this time-frequency-joint representation, an adaptive decomposition technique based on the multi-channel adaptive Fourier decomposition (MAFD) is proposed to adaptively decompose SSVEP signals of different stimuli simultaneously. Then, common components can be identified and transferred across stimuli. RESULTS: A simulation study on public SSVEP datasets demonstrates that the proposed stimulus-stimulus transfer method has the ability to extract and transfer these common components across stimuli. By using calibration data from eight source stimuli, the proposed stimulus-stimulus transfer method can generate SSVEP templates of other 32 target stimuli. It boosts the ITR of the stimulus-stimulus transfer based recognition method from 95.966 bits/min to 123.684 bits/min. CONCLUSION: By extracting and transfer common components across stimuli in the proposed time-frequency-joint representation, the proposed stimulus-stimulus transfer method produces good classification performance without requiring calibration data of target stimuli. SIGNIFICANCE: This study provides a synchronization standpoint to analyze and model SSVEP signals. In addition, the proposed stimulus-stimulus method shortens the calibration time and thus improve comfort, which could facilitate real-world applications of SSVEP-based BCIs.

Thiolated Janus Silsesquioxane Tetrapod: New Precursors for Functional Materials.

Herein, we report synthetic strategies for the development of a bifunctional Janus T4 tetrapod (Janus ring), in which the orthogonal silsesquioxane and organic faces are independently functionalized. An all-cis T4 tetrasilanolate was functionalized to introduce thiol moieties on the silsesquioxane face and naphthyl groups on the organic face to introduce luminescent and self-organization properties. The stepwise synthesis conditions required to prepare such perfectly defined oligomers via a suite of well-defined intermediates and to avoid polymerization or reactions over all eight positions of the tetrapod are explored via 29Si, 13C and 1H NMR, FTIR and TOF-ESI mass spectroscopy. To the best of our knowledge, this is one of the few reports of Janus T4 tetrapods, with different functional groups located on both faces of the molecule, thus expanding the potential range of applications for these versatile precursors.

Mesoporous organosilicas with thiol functionalised pores: multifunctional dendrimers as sacrificial building block and template.

The synthesis of multifunctional poly(amidoamine) (PAMAM)-based dendrimers containing a cleavable disulfide linker within each arm of the dendrimer, together with condensable triethoxysilyl groups on the periphery of the dendrimer, is described. The dendrimers were mixed with 1,4-bis(triethoxysilyl)benzene and subsequently transformed into silsesquioxane gels or periodic mesoporous organosilicas (PMOs) to generate materials with dendrimers covalently embedded within the interior of the silsesquioxane networks. Subsequent treatment of the gels with dithiothreitol enabled the core of the dendrimers to be selectively cleaved at the disulfide site, thus generating thiol functions localised within the pores. The effect of different dendrimer generations on the reactivity of the pendant thiol functions was probed by impregnation with gold salts, which were reduced to obtain gold nanoparticles within the pore networks of the gels and PMOs. The gels yielded polydisperse gold nanoparticles (2 to 70 nm) with dimensions modulated by the generation of the dendrimer, together with well-defined gold/thiolate clusters with Au⋯S distances of 2.3 Å. Such clusters were also observed in the PMO system, together with monodispersed gold nanoparticles with diameters comparable to that of the organised pores in the PMO. The role of surface functionalisation in controlling the formation of gold clusters and/or nanoparticles is discussed.

Online Adaptation Boosts SSVEP-Based BCI Performance.

OBJECTIVE: A user-friendly steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI) prefers no calibration for its target recognition algorithm, however, the existing calibration-free schemes perform still far behind their calibration-based counterparts. To tackle this issue, learning online from the subject's unlabeled data is investigated as a potential approach to boost the performance of the calibration-free SSVEP-based BCIs. METHODS: An online adaptation scheme is developed to tune the spatial filters using the online unlabeled data from previous trials, and then developing the online adaptive canonical correlation analysis (OACCA) method. RESULTS: A simulation study on two public SSVEP datasets (Dataset I and II) with a total of 105 subjects demonstrated that the proposed online adaptation scheme can boost the CCA's averaged information transfer rate (ITR) from 94.60 to 158.87 bits/min in Dataset I and from 85.80 to 123.91 bits/min in Dataset II. Furthermore, in our online experiment it boosted the CCA's ITR from 55.81 bits/min to 95.73 bits/min. More importantly, this online adaptation scheme can be easily combined with any spatial filtering-based algorithms to achieve online learning. CONCLUSION: By online adaptation, the proposed OACCA performed much better than the calibration-free CCA, and comparable to the calibration-based algorithms. SIGNIFICANCE: This work provides a general way for the SSVEP-based BCIs to learn online from unlabeled data and thus avoid calibration.

Tunable high acoustic impedance alumina epoxy composite matching for high frequency ultrasound transducer.

Matching layer is a critical component that determines the performance of piezoelectric ultrasound transducer. For most piezoelectric materials, their acoustic impedances are significantly higher than human tissues and organs, so a tunable matching layer with a high acoustic impedance is required for optimizing the acoustic wave transmission. In this article, a high compression fabrication method is presented, with which the acoustic impedance of alumina-epoxy composite matching layer can be tuned from 6.50 to 9.47 MRayl by controlling the applied compression pressure and ratio of the components. The maximum acoustic impedance 9.47 MRayl can be achieved by compressing a mixture of 80% alumina weight ratio under a 62.4 MPa pressure. This enhancement mainly relies on the increased acoustic longitudinal velocity which enlarged the tolerance of high to ultra-high frequency transducer fabrication using quarter wavelength matching design. Furthermore, the attenuation of the matching layer developed by this method is only -10 dB/mm at 40 MHz. The very high acoustic impedance value and very low attenuation make this matching material superior than all reported matching materials, and therefore, can enhance the performance of the ultrasound transducers, especially for medical imaging applications.

Functionalised polyhedral oligomeric silsesquioxane with encapsulated fluoride - first observation of fluxional SiF interactions in POSS.

The synthesis of a styryl functionalised POSS incorporating an encapsulated fluoride ion within a (SiO1.5)8 cage (T8-F) is reported. It was characterised by single crystal XRD, MALDI-MS, FTIR, solution (29Si, 19F, 13C, 1H) and solid state (29Si, 19F) NMR. In the absence of 1H decoupling, the 29Si solution NMR spectrum exhibited a triplet of doublets. In contrast, 1H, 19F and 1H/19F double-decoupling resulted in two, three and one signal, respectively, being consistent with a single Si site whose 29Si NMR signal is modulated by both the proximal aromatic-ring protons and fluoride. The associated SiF coupling constant (2.5 Hz) is substantially lower than expected for a covalent Si-F bond and arises from a fluxional SiF covalent effect in which the F- interacts equivalently with all eight Si atoms. Additional variable temperature NMR studies demonstrated a threshold at -5 °C below which no SiF interactions are observed, and above which an increasing SiF covalent character occurs.

Large Polyhedral Oligomeric Silsesquioxane Cages: The Isolation of Functionalized POSS with an Unprecedented Si18 O27 Core.

The synthesis of organo-functionalized polyhedral oligomeric silsesquioxanes (POSS, (R-SiO1.5 )n , Tn ) is an area of significant activity. To date, T14 is the largest such cage synthesized and isolated as a single isomer. Herein, we report an unprecedented, single-isomer styryl-functionalized T18 POSS. Unambiguously identified among nine possible isomers by multinuclear solution NMR (1 H, 13 C, and 29 Si), MALDI-MS, FTIR, and computational studies, this is the largest single-isomer functionalized Tn compound isolated to date. A ring-strain model was developed to correlate the 29 Si resonances with the number of 6-, 5-, and/or 4-Si-atom rings that each non-equivalent Si atom is part of. The model successfully predicts the speciation of non-equivalent Si atoms in other families of Tn compounds, demonstrating its general applicability for assigning 29 Si resonances to Si atoms in cage silsesquioxanes and providing a useful tool for predicting Si-atom environments.

Common Spatial Pattern Reformulated for Regularizations in Brain-Computer Interfaces.

Common spatial pattern (CSP) is one of the most successful feature extraction algorithms for brain-computer interfaces (BCIs). It aims to find spatial filters that maximize the projected variance ratio between the covariance matrices of the multichannel electroencephalography (EEG) signals corresponding to two mental tasks, which can be formulated as a generalized eigenvalue problem (GEP). However, it is challenging in principle to impose additional regularization onto the CSP to obtain structural solutions (e.g., sparse CSP) due to the intrinsic nonconvexity and invariance property of GEPs. This article reformulates the CSP as a constrained minimization problem and establishes the equivalence of the reformulated and the original CSPs. An efficient algorithm is proposed to solve this optimization problem by alternately performing singular value decomposition (SVD) and least squares. Under this new formulation, various regularization techniques for linear regression can then be easily implemented to regularize the CSPs for different learning paradigms, such as the sparse CSP, the transfer CSP, and the multisubject CSP. Evaluations on three BCI competition datasets show that the regularized CSP algorithms outperform other baselines, especially for the high-dimensional small training set. The extensive results validate the efficiency and effectiveness of the proposed CSP formulation in different learning contexts.

Inter- and Intra-Subject Transfer Reduces Calibration Effort for High-Speed SSVEP-Based BCIs.

OBJECTIVE: Steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) that can deliver a high information transfer rate (ITR) usually require subject's calibration data to learn the class- and subject-specific model parameters (e.g. the spatial filters and SSVEP templates). Normally, the amount of the calibration data for learning is proportional to the number of classes (or visual stimuli), which could be huge and consequently lead to a time-consuming calibration. This study presents a transfer learning scheme to substantially reduce the calibration effort. METHODS: Inspired by the parameter-based and instance-based transfer learning techniques, we propose a subject transfer based canonical correlation analysis (stCCA) method which utilizes the knowledge within subject and between subjects, thus requiring few calibration data from a new subject. RESULTS: The evaluation study on two SSVEP datasets (from Tsinghua and UCSD) shows that the stCCA method performs well with only a small amount of calibration data, providing an ITR at 198.18±59.12 (bits/min) with 9 calibration trials in the Tsinghua dataset and 111.04±57.24 (bits/min) with 3 trials in the UCSD dataset. Such performances are comparable to those from using the multi-stimulus CCA (msCCA) and the ensemble task-related component analysis (eTRCA) methods with the minimally required calibration data (i.e., at least 40 trials in the Tsinghua dataset and at least 12 trials in the UCSD dataset), respectively. CONCLUSION: Inter- and intra-subject transfer helps the recognition method achieve high ITR with extremely little calibration effort. SIGNIFICANCE: The proposed approach saves much calibration effort without sacrificing the ITR, which would be significant for practical SSVEP-based BCIs.

Changes of EEG phase synchronization and EOG signals along the use of steady state visually evoked potential-based brain computer interface.

OBJECTIVE: The steady-state visual evoked potential (SSVEP)-based brain computer interface (BCI) has demonstrated relatively high performance with little user training, and thus becomes a popular BCI paradigm. However, due to the performance deterioration over time, its robustness and reliability appear not sufficient to allow a non-expert to use outside laboratory. It would be thus helpful to study what happens behind the decreasing tendency of the BCI performance. APPROACH: This paper explores the changes of brain networks and electrooculography (EOG) signals to investigate the cognitive capability changes along the use of the SSVEP-based BCI. The EOG signals are characterized by the blink amplitudes and the speeds of saccades, and the brain networks are estimated by the instantaneous phase synchronizations of electroencephalography signals. MAIN RESULTS: Experimental results revealed that the characteristics derived from EOG and brain networks have similar trends which contain two stages. At the beginning, the blink amplitudes and the saccade speeds start to reduce. Meanwhile, the global synchronizations of the brain networks are formed quickly. These observations implies that the cognitive decline along the use of the SSVEP-based BCI. Then, the EOG and the brain networks related characteristics demonstrate a slow recovery or relatively stable trend. SIGNIFICANCE: This study could be helpful for a better understanding about the depreciation of the BCI performance as well as its relationship with the brain networks and the EOG along the use of the SSVEP-based BCI.

Spatial Filtering in SSVEP-Based BCIs: Unified Framework and New Improvements.

OBJECTIVE: In the steady-state visual evoked potential (SSVEP)-based brain computer interfaces (BCIs), spatial filtering, which combines the multi-channel electroencephalography (EEG) signals in order to reduce the non-SSVEP-related component and thus enhance the signal-to-noise ratio (SNR), plays an important role in target recognition. Recently, various spatial filtering algorithms have been developed employing different prior knowledge and characteristics of SSVEPs, however how these algorithms interconnect and differ is not yet fully explored, leading to difficulties in further understanding, utilizing and improving them. METHODS: We propose a unified framework under which the spatial filtering algorithms can be formulated as generalized eigenvalue problems (GEPs) with four different elements: data, temporal filter, orthogonal projection and spatial filter. Based on the framework, we design new spatial filtering algorithms for improvements through the choice of different elements. RESULTS: The similarities, differences and relationships among nineteen mainstream spatial filtering algorithms are revealed under the proposed framework. Particularly, it is found that they originate from the canonical correlation analysis (CCA), principal component analysis (PCA), and multi-set CCA, respectively. Furthermore, three new spatial filtering algorithms are developed with enhanced performance validated on two public SSVEP datasets with 45 subjects. CONCLUSION: The proposed framework provides insights into the underlying relationships among different spatial filtering algorithms and helps the design of new spatial filtering algorithms. SIGNIFICANCE: This is a systematic study to explore, compare and improve the existing spatial filtering algorithms, which would be significant for further understanding and future development of high performance SSVEP-based BCIs.

Sequential delivery of synergistic drugs by silica nanocarriers for enhanced tumour treatment.

Herein hybrid silica nanoparticles have been engineered to direct the sequential delivery of multiple chemotherapeutic drugs in response to external stimuli such as variations in pH. The nanocarriers consist of conventional MCM-41-type nanoparticles, which have been functionalised with an organic ligand (or stalk) grafted onto the external surface. The stalk is designed to "recognise" a complementary molecule, which serves as a "cap" to block the pores of the nanoparticles. First, camptothecin is introduced into the pores by diffusion prior to capping the pore apertures via molecular recognition. The cap, which is a derivative of 5-fluorouracil, serves as a second cytotoxic drug for synergistic chemotherapy. In vitro tests revealed that negligible release of the drugs occurred at pH 7.4, thus avoiding toxic side effects in the blood stream. In contrast, the stalk/cap complex is destabilised within the endolysosomal compartment (pH 5.5) of cancer cells, where release of the drugs was demonstrated. Furthermore, this environmentally responsive system exhibited a synergistic effect of the two drugs, where the pH-triggered release of the cytotoxic cap followed by diffusion-controlled release of the drug cargo within the pores led to essentially complete elimination of breast cancer cells.

Learning across multi-stimulus enhances target recognition methods in SSVEP-based BCIs.

OBJECTIVE: Latest target recognition methods that are equipped with learning from the subject's calibration data, represented by the extended canonical correlation analysis (eCCA) and the ensemble task-related component analysis (eTRCA), can achieve extra high performance in the steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs), however their performance deteriorate drastically if the calibration trials are insufficient. This paper develops a new scheme to learn from limited calibration data. APPROACH: A learning across multiple stimuli scheme is proposed for the target recognition methods, which applies to learning the data corresponding to not only the target stimulus but also the other stimuli. The resulting optimization problems can be simplified and solved utilizing the prior knowledge and properties of SSVEPs across different stimuli. With the new learning scheme, the eCCA and the eTRCA can be extended to the multi-stimulus eCCA (ms-eCCA) and the multi-stimulus eTRCA (ms-eTRCA), respectively, as well as a combination of them (i.e. ms-eCCA+ms-eTRCA) that incorporates their merits. MAIN RESULTS: Evaluation and comparison using an SSVEP-BCI benchmark dataset with 35 subjects show that the ms-eCCA (or ms-eTRCA) performs significantly better than the eCCA (or eTRCA) method while the ms-eCCA+ms-eTRCA performs the best. With the learning across stimuli scheme, the existing target recognition methods can be further improved in terms of the target recognition performance and the ability against insufficient calibration. SIGNIFICANCE: A new learning scheme is proposed towards the efficient use of the calibration data, providing enhanced performance and saving calibration time in the SSVEP-based BCIs.

Modulating the Electrical Transport in the Two-Dimensional Electron Gas at LaAlO_{3}/SrTiO_{3} Heterostructures by Interfacial Flexoelectricity.

Thin film flexoelectricity is attracting more attention because of its enhanced effect and potential application in electronic devices. Here we find that a mechanical bending induced flexoelectricity significantly modulates the electrical transport properties of the interfacial two-dimensional electron gas (2DEG) at the LaAlO_{3}/SrTiO_{3} (LAO/STO) heterostructure. Under variant bending states, both the carrier density and mobility of the 2DEG are changed according to the flexoelectric polarization direction, showing an electric field effect modulation. By measuring the flexoelectric response of LAO, it is found that the effective flexoelectricity in the LAO thin film is enhanced by 3 orders compared to its bulk. These results broaden the horizon of study on the flexoelectricity effect in the hetero-oxide interface and more research on the oxide interfacial flexoelectricity may be stimulated.

Infrared and Raman spectroscopy of non-conventional hydrogen bonding between N,N'-disubstituted urea and thiourea groups: a combined experimental and theoretical investigation.

The variety of H bond (HB) interactions is a source of inspiration for bottom-up molecular engineering through self-aggregation. Non-conventional intermolecular HBs between N,N'-disubstituted urea and thiourea are studied in detail by vibrational spectroscopies and ab initio calculations. Raman and IR mode assignments are given. We show that it is possible to study selectively the different intermolecular bifurcated intra- and inter-dimer HBs with the two types of HB acceptors. Through the ab initio calculation, the thioamide I mode, a specific marker of N-HS[double bond, length as m-dash]C HB interactions, is unambiguously identified.

Porphyrin-based bridged silsesquioxane nanoparticles for targeted two-photon photodynamic therapy of zebrafish xenografted with human tumor.

BACKGROUND: Bridged silsesquioxane nanoparticles (BSNs) recently described represent a new class of nanoparticles exhibiting versatile applications and particularly a strong potential for nanomedicine. AIMS: In this work, we describe the synthesis of BSNs from an octasilylated functional porphyrin precursor (PORBSNs) efficiently obtained through a click reaction. These innovative and very small-sized nanoparticles were functionalized with PEG and mannose (PORBSNs-mannose) in order to target breast tumors in vivo. METHODS AND RESULTS: The structure of these nanoparticles is constituted of porphyrins J aggregates that allow two-photon spatiotemporal excitation of the nanoparticles. The therapeutic potential of such photoactivable nanoparticles was first studied in vitro, in human breast cancer cells in culture and then in vivo on zebrafish embryos bearing human tumors. These animal models were intravenously injected with 5 nL of a solution containing PORBSNs-mannose. An hour and half after the injection of photoactivable and targeted nanoparticles, the tumor areas were excited for few seconds with a two-photon beam induced focused laser. We observed strong tumor size decrease, with the involvement of apoptosis pathway activation. CONCLUSION: We demonstrated the high targeting, imaging, and therapeutic potential of PORBSNs-mannose injected in the blood stream of zebrafish xenografted with human tumors.

Eyes-Closed Resting EEG Predicts the Learning of Alpha Down-Regulation in Neurofeedback Training.

Neurofeedback training, which enables the trainee to learn self-control of the EEG activity of interest based on online feedback, has demonstrated benefits on cognitive and behavioral performance. Nevertheless, as a core mechanism of neurofeedback, learning of EEG regulation (i.e., EEG learning) has not been well understood. Moreover, a substantial number of non-learners who fail to achieve successful EEG learning have often been reported. This study investigated the EEG learning in alpha down-regulation neurofeedback, aiming to better understand the alpha learning and to early predict learner/non-learner. Twenty-nine participants received neurofeedback training to down-regulate alpha in two days, while eight of them were identified as non-learners who failed to reduce their alpha within sessions. Through a stepwise linear discriminant analysis, a prediction model was built based on participant's eyes-closed resting EEG activities in broad frequency bands including lower alpha, theta, sigma and beta 1 measured before training, which was validated in predicting learners/non-learners. The findings would assist in the early identification of the individuals who would not likely reduce their alpha during neurofeedback.