Self-supervised pretraining improves the performance of classification of task functional magnetic resonance imaging.

Introduction: Decoding brain activities is one of the most popular topics in neuroscience in recent years. And deep learning has shown high performance in fMRI data classification and regression, but its requirement for large amounts of data conflicts with the high cost of acquiring fMRI data. Methods: In this study, we propose an end-to-end temporal contrastive self-supervised learning algorithm, which learns internal spatiotemporal patterns within fMRI and allows the model to transfer learning to datasets of small size. For a given fMRI signal, we segmented it into three sections: the beginning, middle, and end. We then utilized contrastive learning by taking the end-middle (i.e., neighboring) pair as the positive pair, and the beginning-end (i.e., distant) pair as the negative pair. Results: We pretrained the model on 5 out of 7 tasks from the Human Connectome Project (HCP) and applied it in a downstream classification of the remaining two tasks. The pretrained model converged on data from 12 subjects, while a randomly initialized model required 100 subjects. We then transferred the pretrained model to a dataset containing unpreprocessed whole-brain fMRI from 30 participants, achieving an accuracy of 80.2 ± 4.7%, while the randomly initialized model failed to converge. We further validated the model's performance on the Multiple Domain Task Dataset (MDTB), which contains fMRI data of 26 tasks from 24 participants. Thirteen tasks of fMRI were selected as inputs, and the results showed that the pre-trained model succeeded in classifying 11 of the 13 tasks. When using the 7 brain networks as input, variations of the performance were observed, with the visual network performed as well as whole brain inputs, while the limbic network almost failed in all 13 tasks. Discussion: Our results demonstrated the potential of self-supervised learning for fMRI analysis with small datasets and unpreprocessed data, and for analysis of the correlation between regional fMRI activity and cognitive tasks.

Cerebellar Structural Abnormality in Autism Spectrum Disorder: A Magnetic Resonance Imaging Study.

OBJECTIVE: This study uses structural magnetic resonance imaging to explore changes in the cerebellar lobules in patients with autism spectrum disorder (ASD) and further analyze the correlation between cerebellar structural changes and clinical symptoms of ASD. METHODS: A total of 75 patients with ASD and 97 typically developing (TD) subjects from Autism Brain Imaging Data Exchange dataset were recruited. We adopted an advanced automatic cerebellar lobule segmentation technique called CEREbellum Segmentation to segment each cerebellar hemisphere into 12 lobules. Normalized cortical thickness of each lobule was recorded, and group differences in the cortical measures were evaluated. Correlation analysis was also performed between the normalized cortical thickness and the score of Autism Diagnostic Interview-Revised. RESULTS: Results from analysis of variance showed that the normalized cortical thickness of the ASD group differed significantly from that of the TD group; specifically, the ASD group had lower normalized cortical thickness than the TD group. Post-hoc analysis revealed that the differences were more predominant in the left lobule VI, left lobule Crus I and left lobule X, and in the right lobule VI and right lobule Crus I. Lowered normalized cortical thickness in the left lobule Crus I in the ASD patients correlated positively with the abnormality of development evident at or before 36 months subscore. CONCLUSION: These results suggest abnormal development of cerebellar lobule structures in ASD patients, and such abnormality might significantly influence the pathogenesis of ASD. These findings provide new insights into the neural mechanisms of ASD, which may be clinically relevant to ASD diagnosis.

Attention module improves both performance and interpretability of four-dimensional functional magnetic resonance imaging decoding neural network.

Decoding brain cognitive states from neuroimaging signals is an important topic in neuroscience. In recent years, deep neural networks (DNNs) have been recruited for multiple brain state decoding and achieved good performance. However, the open question of how to interpret the DNN black box remains unanswered. Capitalizing on advances in machine learning, we integrated attention modules into brain decoders to facilitate an in-depth interpretation of DNN channels. A four-dimensional (4D) convolution operation was also included to extract temporo-spatial interaction within the fMRI signal. The experiments showed that the proposed model obtains a very high accuracy (97.4%) and outperforms previous researches on the seven different task benchmarks from the Human Connectome Project (HCP) dataset. The visualization analysis further illustrated the hierarchical emergence of task-specific masks with depth. Finally, the model was retrained to regress individual traits within the HCP and to classify viewing images from the BOLD5000 dataset, respectively. Transfer learning also achieves good performance. Further visualization analysis shows that, after transfer learning, low-level attention masks remained similar to the source domain, whereas high-level attention masks changed adaptively. In conclusion, the proposed 4D model with attention module performed well and facilitated interpretation of DNNs, which is helpful for subsequent research.

Inhibition of RhoA/Rho kinase signaling pathway by fasudil protects against kainic acid-induced neurite injury.

AIM: RhoA/Rho kinase pathway is essential for regulating cytoskeletal structure. Although its effect on normal neurite outgrowth has been demonstrated, the role of this pathway in seizure-induced neurite injury has not been revealed. The research examined the phosphorylation level of RhoA/Rho kinase signaling pathway and to clarify the effect of fasudil on RhoA/Rho kinase signaling pathway and neurite outgrowth in kainic acid (KA)-treated Neuro-2A cells and hippocampal neurons. METHOD: Western blotting analysis was used to investigate the expression of key proteins of RhoA/Rho kinase signaling pathway and the depolymerization of actin. After incubated without serum to induce neurite outgrowth, Neuro-2A cells were fixed, and immunofluorescent assay of rhodamine-phalloidin was applied to detect the cellular morphology and neurite length. The influence of KA on neurons was detected in primary hippocampal neurons. Whole-cell patch clamp was conducted in cultured neurons or hippocampal slices to record action potentials. RESULT: KA at the dose of 100-200 µmol/L induced the increase in phosphorylation of Rho-associated coiled-coil-containing protein kinase and decrease in phosphorylation of Lin11, Isl-1 and Mec-3 kinase and cofilin. The effect of 200 µmol/L KA was peaked at 1-2 hours, and then gradually returned to baseline after 8 hours. Pretreatment with Rho kinase inhibitor fasudil reversed KA-induced activation of RhoA/Rho kinase pathway and increase in phosphorylation of slingshot and 14-3-3, which consequently reduced the ratio of G/F-actin. KA treatment induced inhibition of neurite outgrowth and decrease in spines both in Neuro-2a cells and in cultured hippocampal neurons, and pretreatment with fasudil alleviated KA-induced neurite outgrowth inhibition and spine loss. CONCLUSION: These data indicate that inhibiting RhoA/Rho kinase pathway might be a potential treatment for seizure-induced injury.

Chemogenetic manipulation of astrocytic activity: Is it possible to reveal the roles of astrocytes?

Astrocytes are the major glial cells in the central nervous system, but unlike neurons, they do not produce action potentials. For many years, astrocytes were considered supporting cells in the central nervous system (CNS). Technological advances over the last two decades are changing the face of glial research. Accumulating data from recent investigations show that astrocytes display transient calcium spikes and regulate synaptic transmission by releasing transmitters called gliotransmitters. Many new powerful technologies are used to interfere with astrocytic activity, in order to obtain a better understanding of the roles of astrocytes in the brain. Among these technologies, chemogenetics has recently been used frequently. In this review, we will summarize new functions of astrocytes in the brain that have been revealed using this cutting-edge technique. Moreover, we will discuss the possibilities and challenges of manipulating astrocytic activity using this technology.

Astroglial adrenoreceptors modulate synaptic transmission and contextual fear memory formation in dentate gyrus.

Astrocytes perform various supporting functions, including ion buffering, metabolic supplying and neurotransmitter clearance. They can also sense neuronal activity owing to the presence of specific receptors for neurotransmitters. In turn, astrocytes can regulate synaptic activity through the release of gliotransmitters. Evidence has shown that astrocytes are very sensitive to the locus coeruleus (LC) afferents. However, little is known about how LC neuromodulatory norepinephrine (NE) modulates synaptic transmission through astrocytic activity. In mouse dentate gyrus (DG), we demonstrated an increase in the frequency of miniature excitatory postsynaptic currents (mEPSC) in response to NE, which required the release of glutamate from astrocytes. The rise in glutamate release probability is likely due to the activation of presynaptic GluN2B-containing NMDA receptors. Moreover, we showed that the activation of NE signaling in DG is necessary for the formation of contextual learning memory. Thus, NE signaling activation during fear conditioning training contributed to enduring changes in the frequency of mEPSC in DG. Our results strongly support the physiological neuromodulatory role of NE signaling, which is derived from activation of astrocytes.

[High dose vitamin C inhibits proliferation of breast cancer cells through reducing glycolysis and protein synthesis].

OBJECTIVE: To investigate the effects of high dose vitamin C (VC) on proliferation of breast cancer cells and to explore its mechanisms. METHODS: Human breast cancer cells Bcap37 and MDA-MB-453 were treated with VC at low dose (0.01 mmol/L), medium dose (0.10 mmol/L) and high dose (2.00 mmol/L). Cell proliferation was determined with CCK-8 assay, protein expression was evaluated by Western blot, and the secretion of lactic acid in tumor cells was detected by colorimetric method. Bcap37 cells were inoculated in nude mice, and tumor baring nude mice were intraperitoneally injected with high VC(4 g/kg, VC group, n=5)or normal saline (control group, n=5) for 24 d. Tumor weight and body weight were calculated. RESULTS: In vitro experiments demonstrated that high dose VC significantly inhibited cell proliferation in Bcap37 and MDA-MB-453 cells (all P<0.01); the expressions of Glut1 and mTOR signaling pathway-related proteins were decreased (all P<0.05); and the secretion of lactic acid was also markedly reduced (all P<0.05). In vivo experiment showed that the tumor weight was decreased in mice treated with high-dose VC as compared with control group (P<0.05), but no difference in body weights between two groups was observed. CONCLUSIONS: High dose VC may inhibit proliferation of breast cancer cells both in vitro and in vivo through reducing glycolysis and protein synthesis.