Neurophysiological biomarkers for depression classification: Utilizing microstate k-mers and a bag-of-words model.

Microstates are analogous to characters in a language, and short fragments consisting of several microstates (k-mers) are analogous to words. We aimed to investigate whether microstate k-mers could be used as neurophysiological biomarkers to differentiate between depressed patients and normal controls. We utilized a bag-of-words model to process microstate sequences, using k-mers with a k range of 1-10 as terms, and the term frequency (TF) with or without inverse-document-frequency (IDF) as features. We performed nested cross-validation on Dataset 1 (27 patients and 26 controls) and Dataset 2 (34 patients and 30 controls) separately and then trained on one dataset and tested on the other. The best area under the curve (AUC) of 81.5% was achieved for the model with L1 regularization using the TF of 4-mers as features in Dataset 1, and the best AUC of 88.9% was achieved for the model with L1 regularization using the TF of 9-mers as features in Dataset 2. When Dataset 1 was used as the training set, the best AUC of predicting Dataset 2 was 74.1% for the model with L2 regularization using the TF-IDF of 9-mers as features, while the best AUC of predicting Dataset 1 was 70.2% for the model with L1 regularization using the TF of 8-mers as features. Our study provided novel insights into the potential of microstate k-mers as neurophysiological biomarkers for individual-level classification of depression. These may facilitate further exploration of microstate sequences using natural language processing techniques.

A resting-state electroencephalographic microstates study in depressed adolescents with non-suicidal self-injury.

Neuroimaging studies have revealed abnormal brain activities in depressed teenagers who engage in non-suicidal self-injury (NSSI). We used resting-state electroencephalography (EEG) microstate analysis, which indicates the brief overlap of brain network activation for exploring the characteristics of large-scale cortical activities in depressed adolescents engaged with NSSI to clarify the underlying temporal mechanism. A modified k-means cluster algorithm was used to segment 64-channel resting-state EEG data into microstates. Data from 27 healthy adolescents, 37 adolescents with major depressive disorder (MDD), and 53 adolescents with both MDD and NSSI were examined in this study. The resting-state microstate parameters were compared among groups using the one-way ANOVA and Spearman correlation. Then the associations between significantly different microstate parameters and the depressive severity and self-harming data in the patient groups were further analyzed. The MDD group had higher contribution (p < 0.01), occurrence (p < 0.01) of microstate A, and higher microstate E→A transition (p < 0.05) than the HC and the NSSI group. The MDD group showed a distinctly longer duration (p < 0.05) of microstate A and microstate A→C transition than the HC. The transition probability from B to C was increased in the NSSI group compared to the HC. In the MDD group, the HAMD correlated with the duration of microstate A (Spearman's rho = 0.34, p = 0.044), as the PHQ-9 correlated with its occurrence (Spearman's rho = 0.37, p = 0.028). This research revealed that whereas depressive adolescents with NSSI and MDD displayed similar patterns with healthy controls in EEG microstate, the MDD group did not. Additionally, the non-random transition from microstate E→A may protect against recent self-harm in adolescents with MDD.

Altered Neural Reactivity in Adolescents With Nonsuicidal Self-Injury During Exposure to Self-Injury Related Cues: Electrophysiological Evidence From a Two-Choice Oddball Paradigm.

Background: Nonsuicidal self-injury (NSSI) may be a type of addiction, that is characterized by cue reactivity. We aimed to explore the behavioral performance and neural reactivity during exposure to self-injury cues in adolescents with NSSI and major depressive disorder (MDD). Methods: Eighteen MDD patients, 18 MDD patients with NSSI, and 19 healthy controls (HC) were recruited to perform a two-choice oddball paradigm. All subjects were 12-18 years old. Neutral cues and self-injury related cues separately served as deviant stimuli. Difference waves in N2 and P3 (N2d and P3d) were derived from deviant waves minus standard waves. Accuracy cost and reaction time (RT) cost were used as behavioral indexes, while the N2d and P3d were used as electrophysiological indexes; the N2d reflects early conflict detection, and the P3d reflects the process of response inhibition. Results: No significant main effects of group or cue or an effect of their interaction were observed on accuracy cost and P3d latency. For RT cost, N2d amplitude, and N2d latency, there was a significant main effect of cue. For P3d amplitude, there was a significant main effect of cue and a significant group × cue interaction. In the NSSI group, the P3d amplitude with self-injury cues was significantly larger than that with neutral cues. However, there was no such effect in the MDD and HC groups. Conclusions: Adolescents with NSSI showed altered neural reactivity during exposure to self-injury cue. Further studies with larger sample sizes are needed to confirm our results.

Radiomics-based model for accurately distinguishing between severe acute respiratory syndrome associated coronavirus 2 (SARS-CoV-2) and influenza A infected pneumonia.

Clinicians have been faced with the challenge of differentiating between severe acute respiratory syndrome associated coronavirus 2 (SARS-CoV-2) infected pneumonia (NCP) and influenza A infected pneumonia (IAP), a seasonal disease that coincided with the outbreak. We aim to develop a machine-learning algorithm based on radiomics to distinguish NCP from IAP by texture analysis based on computed tomography (CT) imaging. Forty-one NCP and 37 IAP patients admitted from January to February 6, 2019 admitted to two hospitals in Wenzhou, China. All patients had undergone chest CT examination and blood routine tests prior to receiving medical treatment. NCP was diagnosed by real-time RT-PCR assays. Eight of 56 radiomic features extracted by LIFEx were selected by least absolute shrinkage and selection operator regression to develop a radiomics score and subsequently constructed into a nomogram to predict NCP with area under the operating characteristics curve of 0.87 (95% confidence interval: 0.77-0.93). The nomogram also showed excellent calibration with Hosmer-Lemeshow test yielding a nonsignificant statistic (P = .904). The novel nomogram may efficiently distinguish between NCP and IAP patients. The nomogram may be incorporated to existing diagnostic algorithm to effectively stratify suspected patients for SARS-CoV-2 pneumonia.