The current study put forward a multi-feature kernel discriminant dictionary learning algorithm for face recognition. It was based on the supervised within-class-similar discriminative dictionary learning algorithm (SCDDL) we introduced previously. The proposed new algorithm was thus named as multi-feature kernel SCDDL (MKSCDDL). In contrast to the weighted combination or the constraint of representation coefficients for the feature combination used by some popular methods, MKSCDDL introduced the multiple kernel learning technique into the dictionary learning scheme. The experimental results on three large well-known face databases suggested that combination multiple features in MKSCDDL improved the recognition rate compared with SCDDL. In addition, adopting multiple kernel learning technique resulted in an excellent multi-feature dictionary learning approach when compared with some state-of-the-art multi-feature algorithms such as multiple kernel learning and multi-task joint sparse representation methods, indicating the effectiveness of the multiple kernel learning technique in the combination of multiple features for classification.

The emotional Stroop task (EST) is among the most influential paradigms used to probe attention-related or cognitive control-related emotional processing in healthy subjects and clinical populations. The neuropsychological mechanism underlying the emotional Stroop effect has attracted extensive and long-lasting attention in both cognitive and clinical psychology and neuroscience; however, a precise characterization of the neural substrates underlying the EST in healthy and clinical populations remains elusive. Here, we implemented a coordinate-based meta-analysis covering functional imaging studies that employed the emotion-word or emotional counting Stroop paradigms to determine the underlying neural networks in healthy subjects and the trans-diagnostic alterations across clinical populations. Forty-six publications were identified that reported relevant contrasts (negative > neutral; positive > neutral) for healthy or clinical populations as well as for hyper- or hypo-activation of patients compared to controls. We demonstrate consistent involvement of the vlPFC and dmPFC in healthy subjects and consistent involvement of the vlPFC in patients. We further identify a trans-diagnostic pattern of hyper-activation in the prefrontal and parietal regions. These findings underscore the critical roles of cognitive control processes in the EST and implicate trans-diagnostic cognitive control deficits. Unlike the current models that emphasize the roles of the amygdala and rACC, our findings implicate novel mechanisms underlying the EST for both healthy and clinical populations.

Objective. Depression is a severe mental disorder. However, the neural mechanisms underlying affective interference (difficulties in directing attention away from negative distractors) in depression patients are still not well-understood. In particular, the connections between brain regions remain unclear. Using the emotional face-word Stroop task, we aimed to reveal the altered electroencephalography (EEG) functional connectivity in patients with depression, using concepts from event-related potentials (ERPs) and time series clustering. Approach. In this study, the EEG signals of ten healthy participants and ten depression patients were collected from a 64-sensor cap. Subsequently, EEG signals were segmented into temporal windows corresponding to the ERPs. For each duration, the dynamic time warping algorithm was used to calculate the similarities between EEG signals from different electrodes, and differences of these similarities were compared between the groups. Finally, hierarchical clustering was used to identify functionally connected regions and examine changes in depression. Main results. It was observed that during the time interval of 400–600 ms (N450 components), depression patients had more long-range connections than did healthy control patients and exhibited abnormal functional connectivity via the superior and middle frontal gyrus, specifically, the dorsolateral prefrontal cortex (DL-PFC, Brodmann's area 8 and 9), which is related to the control and resolution of affective interference. Moreover, the functionally connected region of depression patients was much larger than that of healthy participants, which is caused by brain resource reorganization. Significance. These findings thus provide new insights into the neural mechanisms of depression and further identify the DL-PFC and connections between certain electrodes as quantitative indicators of depression.

Narcissism is one of the most fundamental personality traits in which individuals in general population exhibit a large heterogeneity. Despite a surge of interest in examining behavioral characteristics of narcissism in the past decades, the neurobiological substrates underlying narcissism remain poorly understood. Here, we addressed this issue by applying a machine learning approach to decode trait narcissism from whole‐brain resting‐state functional connectivity (RSFC). Resting‐state functional MRI (fMRI) data were acquired for a large sample comprising 155 healthy adults, each of whom was assessed for trait narcissism. Using a linear prediction model, we examined the relationship between whole‐brain RSFC and trait narcissism. We demonstrated that the machine‐learning model was able to decode individual trait narcissism from RSFC across multiple neural systems, including functional connectivity between and within limbic and prefrontal systems as well as their connectivity with other networks. Key nodes that contributed to the prediction model included the amygdala, prefrontal and anterior cingulate regions that have been linked to trait narcissism. These findings remained robust using different validation procedures. Our findings thus demonstrate that RSFC among multiple neural systems predicts trait narcissism at the individual level.

Anxiety is a complex emotional state that has a great impact on people's physical and mental health. Effectively identifying different anxiety states is very important. By inducing various anxiety states with electroencephalograph (EEG) recording, comprehensive EEG features (frequency and time domain features, statistical and nonlinear features) were extracted from different EEG bands and brain locations. Next, correlation analysis was performed for feature selection. And different classifiers were applied to classify four anxiety levels using different features alone or together to explore their anxiety recognition ability. Based on our dataset, the highest accuracy of identifying four anxiety states reached approximately 62.56% using the Support Vector Machine (SVM), which improved the classification accuracy compared with previous studies. The results also revealed the importance of EEG linear features (especially for features including total power, mean square and variance) in anxiety recognition. Furthermore, it suggested that EEG features in the beta band and the frontal lobe contributed to anxiety recognition more than the features in the other bands or other brain locations. In short, this study improves the accuracy of multi-level anxiety recognition and helps in choosing better features for anxiety recognition, which lay the foundation for the detection of continuous anxiety changes.

In this paper, we investigate brain directed connectivity (BDC) networks for emotion recognition using electroencephalogram (EEG) source signals that were estimated from high-density sensor EEG signals, for the first time. Currently, a variety of features extracted from sensor EEG signals are used for emotion recognition. However, they cannot unambiguously describe the location of emotions associated with neural activities and information propagation or the interaction between brain regions. In addition, most current studies use low-density sensor EEG signals. Moreover, source signals estimated from high-density sensor EEG signal have not been employed for emotion recognition to date. We designed a BDC network-based framework using EEG source signals to investigate emotion recognition. The global cortex factor-based multivariate autoregressive (GCF-MVAR) method was utilized to extract emotion-related BDC features. Our study revealed that the combined BDC and DE features facilitated a recognition accuracy of up to 89.58 %, which is higher than the rate obtained from BDC features and DE features alone. The sensor features derived from high-density EEG signals also exhibited higher recognition accuracy compared to low-density EEG signals. These findings suggest that BDC features derived from EEG source signals can better characterize human emotional states and are meaningful for emotion recognition.

The connectivity hub property of the hippocampus (HIP) and the medial prefrontal cortex (MPFC) is essential for their widespread involvement in cognition; however, the cooperation mechanism between them is far from clear. Herein, using resting-state functional MRI and Gaussian Bayesian network to describe the directed organizing architecture of the HIP–MPFC pathway with regions in the brain, we demonstrated that the HIP and the MPFC have central roles as the driving hub and aggregating hub, respectively. The status of the HIP and the MPFC is dominant in communications between the HIP and the default-mode network, between the HIP and core neurocognitive networks, including the default-mode, frontoparietal, and salience networks, and between brain-wide representative regions, suggesting a strong and robust central position of the two regions in regulating the dynamics of large-scale brain activity. Furthermore, we found that the directed connectivity and flow from the right HIP to the MPFC is significantly linked to fluid intelligence. Together, these results clarify the different roles of the HIP and the MPFC that jointly contribute to network dynamics and cognitive ability from a data-driven insight via the use of the directed connectivity method.

Effective features can improve the performance of a model and help us understand the characteristics and underlying structure of complex data. Previously proposed feature selection methods usually cannot retain more discriminative information. To address this shortcoming, we propose a novel supervised orthogonal least square regression model with feature weighting for feature selection. The optimization problem of the objective function can be solved by employing generalized power iteration and augmented Lagrangian multiplier methods. Experimental results show that the proposed method can more effectively reduce feature dimensionality and obtain better classification results than traditional feature selection methods. The convergence of our iterative method is also proved. Consequently, the effectiveness and superiority of the proposed method are verified both theoretically and experimentally.

Brain functional connectivity (FC) has shown great potential in becoming biomarkers of brain status. However, the problem of accurately estimating FC from complex-noisy fMRI time series remains unsolved. Usually, a regularization function is more appropriate in fitting the real inherent properties of the brain function activity pattern, which can further limit noise interference to improve the accuracy of the estimated result. Recently, the neuroscientists widely suggested that the inherent brain function activity pattern indicates sparse, modular and overlapping topology. However, previous studies have never considered this factual characteristic. Thus, we propose a novel method by the integration of these inherent brain function activity pattern priors to estimate FC. Extensive experiments on synthetic data demonstrate that our method can more accurately estimate the FC than previous. Then, we applied the estimated FC to predict the symptom severity of depressed patients, the symptom severity is related to subtle abnormal changes in the brain function activity, a more accurate FC can more effectively capture the subtle abnormal brain function activity changes. As results, our method better than others with a higher correlation coefficient of 0.4201. Moreover, the overlapping probabilistic of each brain region can be further explored by the proposed method.

Autism Spectrum Disorder (ASD) is a pervasive neurodevelopmental disorder characterized by restricted interests and repetitive behaviors. Non-invasive measurements of brain activity with functional magnetic resonance imaging (fMRI) have demonstrated that the abnormality in the default mode network (DMN) is a crucial neural basis of ASD, but the time-frequency characteristic of the network has not yet been revealed. Hilbert-Huang transform (HHT) is conducive to feature extraction of biomedical signals and has recently been suggested as an effective method to explore the time-frequency feature of the brain activity and mechanism. In this study, the resting-state fMRI dataset of 105 subjects including 59 ASD participants and 46 healthy control (HC) participants were involved in the original time-frequency clustering analysis based on improved HHT and modified k-means clustering with label-replacement. Compared with HC, ASD selectively showed enhanced Hilbert Weight Frequency (HWF) in high frequency bands in crucial regions of the DMN, including the medial prefrontal cortex (MPFC), posterior cingulate cortex (PCC) and anterior cingulate cortex (ACC). Time-frequency clustering analysis revealed altered DMN organization in ASD. In the posterior DMN, the PCC and bilateral precuneus were separated for HC but clustered for ASD; in the anterior DMN, the clusters of ACC, dorsal MPFC, and ventral MPFC were relatively scattered for ASD. This study paves a promising way to uncover the alteration in the DMN of ASD and identify a potential neuroimaging biomarker for diagnostic reference.