BCAT1 alleviates early brain injury by inhibiting ferroptosis through PI3K/AKT/mTOR/GPX4 pathway after subarachnoid hemorrhage.

Regulation of the redox system by branched-chain amino acid transferase 1 (BCAT1) is of great significance in the occurrence and development of diseases, but the relationship between BCAT1 and subarachnoid hemorrhage (SAH) is still unknown. Ferroptosis, featured by iron-dependent lipid peroxidation accompanied by the depletion of glutathione peroxidase 4 (GPX4), has been implicated in the pathological process of early brain injury after subarachnoid hemorrhage. This study established SAH model by endovascular perforation and adding oxyhemoglobin (Hb) to HT22 cells and delved into the mechanism of BCAT1 in SAH-induced ferroptotic neuronal cell death. It was found that SAH-induced neuronal ferroptosis could be inhibited by BCAT1 overexpression (OE) in rats and HT22 cells, and BCAT1 OE alleviated neurological deficits and cognitive dysfunction in rats after SAH. In addition, the effect of BCAT1 could be reversed by the Ly294002, a specific inhibitor of the PI3K pathway. In summary, our present study indicated that BCAT1 OE alleviated early brain injury EBI after SAH by inhibiting neuron ferroptosis via activation of PI3K/AKT/mTOR pathway and the elevation of GPX4. These results suggested that BCAT1 was a promising therapeutic target for subarachnoid hemorrhage.

Multi-Semantic Decoding of Visual Perception with Graph Neural Networks.

Constructing computational decoding models to account for the cortical representation of semantic information plays a crucial role in understanding visual perception. The human visual system processes interactive relationships among different objects when perceiving the semantic contents of natural visions. However, the existing semantic decoding models commonly regard categories as completely separate and independent visually and semantically and rarely consider the relationships from prior information. In this work, a novel semantic graph learning model was proposed to decode multiple semantic categories of perceived natural images from brain activity. The proposed model was validated on the functional magnetic resonance imaging data collected from five normal subjects while viewing 2750 natural images comprising 52 semantic categories. The results showed that the Graph Neural Network-based decoding model achieved higher accuracies than other deep neural network models. Moreover, the co-occurrence probability among semantic categories showed a significant correlation with the decoding accuracy. Additionally, the results suggested that semantic content organized in a hierarchical way with higher visual areas was more closely related to the internal visual experience. Together, this study provides a superior computational framework for multi-semantic decoding that supports the visual integration mechanism of semantic processing.

Transcriptome-wide map of N6-methyladenosine (m6A) profiling in coronary artery disease (CAD) with clopidogrel resistance.

BACKGROUND: Clopidogrel resistance profoundly increases the risk of major cardiovascular events in coronary artery disease (CAD) patients. Here, we comprehensively analyse global m6A modification alterations in clopidogrel-resistant (CR) and non-CR patients. METHODS: After RNA isolation, the RNA transcriptome expression (lncRNA, circRNA, and mRNA) was analysed via RNA-seq, and m6A peaks were identified by MeRIP-seq. The altered m6A methylation sites on mRNAs, lncRNAs, and circRNAs were identified, and then, GO and KEGG pathway analyses were performed. Through joint analysis with RNA-seq and MeRIP-seq data, differentially expressed mRNAs harbouring differentially methylated sites were identified. The changes in m6A regulator levels and the abundance of differentially methylated sites were measured by RT-PCR. The identification of m6A-modified RNAs was verified by m6A-IP-qPCR. RESULTS: The expression of 2919 hypermethylated and 2519 hypomethylated mRNAs, 192 hypermethylated and 391 hypomethylated lncRNAs, and 375 hypermethylated and 546 hypomethylated circRNAs was shown to be altered in CR patients. The m6A peaks related to CR indicated lower mark density at the CDS region. Functional enrichment analysis revealed that inflammatory pathways and insulin signalling pathways might be involved in the pathological processes underlying CR. The expression of mRNAs (ST5, KDM6B, GLB1L2, and LSM14B), lncRNAs (MSTRG.13776.1 and ENST00000627981.1), and circRNAs (hsa_circ_0070675_CBC1, hsa-circRNA13011-5_CBC1, and hsa-circRNA6406-3_CBC1) was upregulated in CR patients, while the expression of mRNAs (RPS16 and CREG1), lncRNAs (MSTRG.9215.1), and circRNAs (hsa_circ_0082972_CBC1) was downregulated in CR patients. Moreover, m6A regulators (FTO, YTHDF3, and WTAP) were also differentially expressed. An additional combined analysis of gene expression and m6A peaks revealed that the expression of mRNAs (such as ST5, LYPD2, and RPS16 mRNAs) was significantly altered in the CR patients. CONCLUSION: The expression of m6A regulators, the RNA transcriptome, and the m6A landscape was altered in CR patients. These findings reveal epitranscriptomic regulation in CR patients, which might be novel therapeutic targets in future.

Melatonin alleviates early brain injury by inhibiting the NRF2-mediated ferroptosis pathway after subarachnoid hemorrhage.

Ferroptosis is a novel form of cell death that plays a critical role in the pathological and physiological processes of early brain injury following subarachnoid hemorrhage. Melatonin, as the most potent endogenous antioxidant, has shown strong protective effects against pathological changes following subarachnoid hemorrhage, but its impact on ferroptosis induced by subarachnoid hemorrhage remains unexplored. In our study, we established a subarachnoid hemorrhage model in male SD rats. We found that subarachnoid hemorrhage induced changes in ferroptosis-related indicators such as lipid peroxidation and iron metabolism, while intraperitoneal injection of melatonin (40 mg/kg) effectively ameliorated these changes to a certain degree. Moreover, in a subset of rats with subarachnoid hemorrhage who received pre-treatment via intravenous injection of the melatonin receptor antagonist Luzindole (1 mg/kg) and 4P-PDOT (1 mg/kg), we found that the protective effect of melatonin against subarachnoid hemorrhage includes inhibition of lipid peroxidation and reduction of iron accumulation depended on melatonin receptor 1B (MT2). Furthermore, our study demonstrated that melatonin inhibited neuronal ferroptosis by activating the NRF2 signaling pathway, as evidenced by in vivo inhibition of NRF2. In summary, melatonin acts through MT2 and activates NRF2 and downstream genes such as HO-1/NQO1 to inhibit ferroptosis in subarachnoid hemorrhage-induced neuronal injury, thereby improving neurological function in rats. These results suggest that melatonin is a promising therapeutic target for subarachnoid hemorrhage.

Comprehensive analysis of the RNA transcriptome expression profiles and construction of the ceRNA network in heart failure patients with sacubitril/valsartan therapeutic heterogeneity after acute myocardial infarction.

Sacubitril/valsartan has a noteworthy advantage in improving ventricular remodelling, as well as reducing cardiovascular mortality and the rate of heart failure (HF) readmission. However, clinically, some patients with HF still have low sensitivity to sacubitril/valsartan, indicating sacubitril/valsartan resistance (SVR). A total of 46 patients with HF after AMI (23 SVR and 23 non-sacubitril/valsartan resistance (NSVR)) were selected. Five SVR and 5 matched NSVR samples were screened for differentially expressed ncRNAs along with mRNAs. A total of 124 differentially expressed miRNAs, 137 circRNAs, 237 lncRNAs and 50 mRNAs were screened by RNA sequencing technology. After quantitative real-time PCR (qRT‒PCR) verification of selected biomarkers in 18 pairs of samples, we found that for patients with SVR, hsa-miR-543, hsa-miR-642b-5p, hsa-miR-760, hsa_circ_0137499, ENST00000474394, ENST00000528337, E2F1, NEAT1, and YTHDF2 were upregulated, and hsa-miR-424-5p, hsa-miR-21-3p, hsa_circRNA_0003275, hsa_circRNA_0004494, hsa_circ_0093522, ENST00000467951, ENST00000558177, ACTA2, ANPEP, and CAMP were downregulated. Then, with the help of our constructed ceRNA network and functional annotation enrichment, we speculated that inflammatory pathways (such as the apelin signalling pathway) and lipid metabolism pathways (such as fatty acid metabolism) may be involved in the regulation of SVR. These discoveries lay a foundation for further mechanistic research and provide a direction for individualized drug administration.

DNA methylation profile in the whole blood of acute coronary syndrome patients with aspirin resistance.

BACKGROUND: Aspirin resistance (AR) results in major adverse cardiovascular events, and DNA methylation might participate in the regulation of this pathological process. METHODS: In present study, a sum of 35 patients with AR and 35 non-AR (NAR) controls were enrolled. Samples from 5 AR and 5 NAR were evaluated in an 850 BeadChip DNA methylation assay, and another 30 AR versus 30 NAR were evaluated to validate the differentially methylated CpG loci (DML). Then, qRT-PCR was used to investigate the target mRNA expression of genes at CpG loci. Finally, Gene Ontology (GO) as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to reveal the enriched pathways. RESULTS: The AR and NAR groups displayed significant differences in DNA methylation at 7707 positions, with 270 hypermethylated sites (e.g., cg09555818 located in APOC2) and 7437 sites hypomethylated sites (e.g., cg26828689 located in SLC12A5). Six DML were validated by pyrosequencing, and it was confirmed that DNA methylation (cg16391727, cg21008208, cg21293749, and cg13945576) was related to the increasing risk of AR. The relative mRNA expression of the ROR1 gene was also associated with AR (p = 0.007), suggesting that the change of cg21293749 in DNA methylation might lead to differential ROR1 mRNA expression, ultimately resulting in AR. Furthermore, the identified differentially methylated sites were associated with the molecular pathways such as circadian rhythms and insulin secretion. CONCLUSION: Hence, the distinct DNA methylation might play a vital role in the biological regulation of AR through the pathways such as circadian rhythms.

Influence of Mg Doping on the Structure and Mechanical Properties of Al2Cu Precipitated Phase by First-Principles Calculations.

Al-Cu-Mg high-strength alloys are widely used in industrial production because of their excellent mechanical performance and good machining properties. In this study, first-principles calculations based on density functional theory were carried out to investigate the influence of Mg doping on the structural stability and mechanical properties of the Al2Cu (θ) precipitated phase in Al-Cu-Mg alloys. The results show that the structural stability, electronic structure, bulk modulus, mechanical anisotropy, and thermodynamic properties of the precipitated Al2CuMgX phase change with the concentration of Mg doping (X = 2, 4, 6, and 8). The cohesive energy calculation and electronic structure analysis show that Al2CuMg6 has a high structural stability. The criterion based on elastic constants indicates that Al2CuMg2, Al2CuMg4, and Al2CuMg8 have a brittle tendency and show strong anisotropy of mechanical properties, while Al2CuMg6 shows better comprehensive mechanical properties. The thermodynamic analysis results based on the quasi-harmonic Debye model show that the Al2CuMg6 precipitated phase has good stability at high temperatures and pressure.

Transfer RNA-derived small RNAs and their potential roles in the therapeutic heterogeneity of sacubitril/valsartan in heart failure patients after acute myocardial infarction.

Background: It has been reported that sacubitril/valsartan can improve cardiac function in acute myocardial infarction (AMI) patients complicated by heart failure (HF). However, a number of patients cannot be treated successfully; this phenomenon is called sacubitril/valsartan resistance (SVR), and the mechanisms remain unclear. Methods: In our present research, the expression profiles of transfer RNA (tRNA)-derived small RNAs (tsRNAs) in SVR along with no sacubitril/valsartan resistance (NSVR) patients were determined by RNA sequencing. Through bioinformatics, quantitative real-time PCR (qRT-PCR), and cell-based experiments, we identified SVR-related tsRNAs and confirmed their diagnostic value, predicted their targeted genes, and explored the enriched signal pathways as well as regulatory roles of tsRNAs in SVR. Results: Our research indicated that 36 tsRNAs were upregulated and that 21 tsRNAs were downregulated in SVR. Among these tsRNAs, the expression of tRF-59:76-Tyr-GTA-2-M3 and tRF-60:76-Val-AAC-1-M5 was upregulated, while the expression of tRF-1:29-Gly-GCC-1 was downregulated in the group of SVR. Receiver operating characteristic (ROC) curve analysis demonstrated that these three tsRNAs were potential biomarkers of the therapeutic heterogeneity of sacubitril/valsartan. Moreover, tRF-60:76-Val-AAC-1-M5 might target Tnfrsf10b and Bcl2l1 to influence the observed therapeutic heterogeneity through the lipid and atherosclerosis signaling pathways. Conclusion: Hence, tsRNA might play a vital role in SVR. These discoveries provide new insights for the mechanistic investigation of responsiveness to sacubitril/valsartan.

The DNAm levels of CREB5 (cg11301281) were associated with clopidogrel resistance.

PURPOSE: Clopidogrel resistance (CR) is mostly caused by interindividual variability of the platelet inhibition of clopidogrel, which may induce cardiovascular events. The aim of this research was to evaluate whether DNAm levels of CREB5 (cg01534253) are involved in CR among acute coronary syndrome (ACS) patients treated with clopidogrel. METHODS: 72 patients(36 CR and 36 non-CR) who underwent ACS were included in this study. The VerifyNow P2Y12 assay was selected to evaluate residual platelet reactivity, and bisulfite pyrosequencing methods was used to examine DNA methylation levels on cg01534253. Secondly, CREB5 mRNA expression was analyzed via quantitative real-time PCR. Last, we employed logistic regression to test the interaction between genetic factors of CREB5 methylation and multiple clinical variables in CR patients. RESULTS: Subunit analysis indicated that for patients whose HbA1c levels were ≥6.5% or whose GLU levels were ≥7 mmol/L, lower methylation of cg01534253 indicated a poorer clopidogrel response. In addition, CREB5 mRNA expression was increased in CR patients with GLU levels ≥7 mmol/L. Moreover, regression analysis indicated that the values of albumin and uric acid were correlated with the incidence of CR. CONCLUSIONS: Our findings were likely to provide fresh understanding for the new mechanism of platelet inhibition failure and promote individualized antiplatelet therapy.

Reconstructing rapid natural vision with fMRI-conditional video generative adversarial network.

Recent functional magnetic resonance imaging (fMRI) studies have made significant progress in reconstructing perceived visual content, which advanced our understanding of the visual mechanism. However, reconstructing dynamic natural vision remains a challenge because of the limitation of the temporal resolution of fMRI. Here, we developed a novel fMRI-conditional video generative adversarial network (f-CVGAN) to reconstruct rapid video stimuli from evoked fMRI responses. In this model, we employed a generator to produce spatiotemporal reconstructions and employed two separate discriminators (spatial and temporal discriminators) for the assessment. We trained and tested the f-CVGAN on two publicly available video-fMRI datasets, and the model produced pixel-level reconstructions of 8 perceived video frames from each fMRI volume. Experimental results showed that the reconstructed videos were fMRI-related and captured important spatial and temporal information of the original stimuli. Moreover, we visualized the cortical importance map and found that the visual cortex is extensively involved in the reconstruction, whereas the low-level visual areas (V1/V2/V3/V4) showed the largest contribution. Our work suggests that slow blood oxygen level-dependent signals describe neural representations of the fast perceptual process that can be decoded in practice.

A CNN-transformer hybrid approach for decoding visual neural activity into text.

BACKGROUND AND OBJECTIVE: Most studies used neural activities evoked by linguistic stimuli such as phrases or sentences to decode the language structure. However, compared to linguistic stimuli, it is more common for the human brain to perceive the outside world through non-linguistic stimuli such as natural images, so only relying on linguistic stimuli cannot fully understand the information perceived by the human brain. To address this, an end-to-end mapping model between visual neural activities evoked by non-linguistic stimuli and visual contents is demanded. METHODS: Inspired by the success of the Transformer network in neural machine translation and the convolutional neural network (CNN) in computer vision, here a CNN-Transformer hybrid language decoding model is constructed in an end-to-end fashion to decode functional magnetic resonance imaging (fMRI) signals evoked by natural images into descriptive texts about the visual stimuli. Specifically, this model first encodes a semantic sequence extracted by a two-layer 1D CNN from the multi-time visual neural activity into a multi-level abstract representation, then decodes this representation, step by step, into an English sentence. RESULTS: Experimental results show that the decoded texts are semantically consistent with the corresponding ground truth annotations. Additionally, by varying the encoding and decoding layers and modifying the original positional encoding of the Transformer, we found that a specific architecture of the Transformer is required in this work. CONCLUSIONS: The study results indicate that the proposed model can decode the visual neural activities evoked by natural images into descriptive text about the visual stimuli in the form of sentences. Hence, it may be considered as a potential computer-aided tool for neuroscientists to understand the neural mechanism of visual information processing in the human brain in the future.

Negative bias effects during audiovisual emotional processing in major depression disorder.

Aberrant affective neural processing and negative emotional bias are trait-marks of major depression disorders (MDDs). However, most research on biased emotional perception in depression has only focused on unimodal experimental stimuli, the neural basis of potentially biased emotional processing of multimodal inputs remains unclear. Here, we addressed this issue by implementing an audiovisual emotional task during functional MRI scanning sessions with 37 patients with MDD and 37 gender-, age- and education-matched healthy controls. Participants were asked to distinguish laughing and crying sounds while being exposed to faces with different emotional valences as background. We combined general linear model and psychophysiological interaction analyses to identify abnormal local functional activity and integrative processes during audiovisual emotional processing in MDD patients. At the local neural level, MDD patients showed increased bias activity in the ventromedial prefrontal cortex (vmPFC) while listening to negative auditory stimuli and concurrently processing visual facial expressions, along with decreased dorsolateral prefrontal cortex (dlPFC) activity in both the positive and negative visual facial conditions. At the network level, MDD exhibited significantly decreased connectivity in areas involved in automatic emotional processes and voluntary control systems during perception of negative stimuli, including the vmPFC, dlPFC, insula, as well as the subcortical regions of posterior cingulate cortex and striatum. These findings support a multimodal emotion dysregulation hypothesis for MDD by demonstrating that negative bias effects may be facilitated by the excessive ventral bottom-up negative emotional influences along with incapability in dorsal prefrontal top-down control system.

Integrating Multilevel Functional Characteristics Reveals Aberrant Neural Patterns during Audiovisual Emotional Processing in Depression.

Emotion dysregulation is one of the core features of major depressive disorder (MDD). However, most studies in depression have focused on unimodal emotion processing, whereas emotional perception in daily life is highly dependent on multimodal sensory inputs. Here, we proposed a novel multilevel discriminative framework to identify the altered neural patterns in processing audiovisual emotion in MDD. Seventy-four participants underwent an audiovisual emotional task functional magnetic resonance imaging scanning. Three levels of whole-brain functional features were extracted for each subject, including the task-evoked activation, task-modulated connectivity, combined activation and connectivity. Support vector machine classification and prediction models were built to identify MDD from controls and evaluate clinical relevance. We revealed that complex neural networks including the emotion regulation network (prefrontal areas and limbic-subcortical regions) and the multisensory integration network (lateral temporal cortex and motor areas) had the discriminative power. Moreover, by integrating comprehensive information of local and interactive processes, multilevel models could lead to a substantial increase in classification accuracy and depression severity prediction. Together, we highlight the high representational capacity of machine learning algorithms to characterize the complex network abnormalities associated with emotional regulation and multisensory integration in MDD. These findings provide novel evidence for the neural mechanisms underlying multimodal emotion dysregulation of depression.

A neural decoding algorithm that generates language from visual activity evoked by natural images.

Transforming neural activities into language is revolutionary for human-computer interaction as well as functional restoration of aphasia. Present rapid development of artificial intelligence makes it feasible to decode the neural signals of human visual activities. In this paper, a novel Progressive Transfer Language Decoding Model (PT-LDM) is proposed to decode visual fMRI signals into phrases or sentences when natural images are being watched. The PT-LDM consists of an image-encoder, a fMRI encoder and a language-decoder. The results showed that phrases and sentences were successfully generated from visual activities. Similarity analysis showed that three often-used evaluation indexes BLEU, ROUGE and CIDEr reached 0.182, 0.197 and 0.680 averagely between the generated texts and the corresponding annotated texts in the testing set respectively, significantly higher than the baseline. Moreover, we found that higher visual areas usually had better performance than lower visual areas and the contribution curve of visual response patterns in language decoding varied at successively different time points. Our findings demonstrate that the neural representations elicited in visual cortices when scenes are being viewed have already contained semantic information that can be utilized to generate human language. Our study shows potential application of language-based brain-machine interfaces in the future, especially for assisting aphasics in communicating more efficiently with fMRI signals.

Motor Circuit and Superior Temporal Sulcus Activities Linked to Individual Differences in Multisensory Speech Perception.

Integrating multimodal information into a unified perception is a fundamental human capacity. McGurk effect is a remarkable multisensory illusion that demonstrates a percept different from incongruent auditory and visual syllables. However, not all listeners perceive the McGurk illusion to the same degree. The neural basis for individual differences in modulation of multisensory integration and syllabic perception remains largely unclear. To probe the possible involvement of specific neural circuits in individual differences in multisensory speech perception, we first implemented a behavioral experiment to examine the McGurk susceptibility. Then, functional magnetic resonance imaging was performed in 63 participants to measure the brain activity in response to non-McGurk audiovisual syllables. We revealed significant individual variability in McGurk illusion perception. Moreover, we found significant differential activations of the auditory and visual regions and the left Superior temporal sulcus (STS), as well as multiple motor areas between strong and weak McGurk perceivers. Importantly, the individual engagement of the STS and motor areas could specifically predict the behavioral McGurk susceptibility, contrary to the sensory regions. These findings suggest that the distinct multimodal integration in STS as well as coordinated phonemic modulatory processes in motor circuits may serve as a neural substrate for interindividual differences in multisensory speech perception.

A dual-channel language decoding from brain activity with progressive transfer training.

When we view a scene, the visual cortex extracts and processes visual information in the scene through various kinds of neural activities. Previous studies have decoded the neural activity into single/multiple semantic category tags which can caption the scene to some extent. However, these tags are isolated words with no grammatical structure, insufficiently conveying what the scene contains. It is well-known that textual language (sentences/phrases) is superior to single word in disclosing the meaning of images as well as reflecting people's real understanding of the images. Here, based on artificial intelligence technologies, we attempted to build a dual-channel language decoding model (DC-LDM) to decode the neural activities evoked by images into language (phrases or short sentences). The DC-LDM consisted of five modules, namely, Image-Extractor, Image-Encoder, Nerve-Extractor, Nerve-Encoder, and Language-Decoder. In addition, we employed a strategy of progressive transfer to train the DC-LDM for improving the performance of language decoding. The results showed that the texts decoded by DC-LDM could describe natural image stimuli accurately and vividly. We adopted six indexes to quantitatively evaluate the difference between the decoded texts and the annotated texts of corresponding visual images, and found that Word2vec-Cosine similarity (WCS) was the best indicator to reflect the similarity between the decoded and the annotated texts. In addition, among different visual cortices, we found that the text decoded by the higher visual cortex was more consistent with the description of the natural image than the lower one. Our decoding model may provide enlightenment in language-based brain-computer interface explorations.

The association of polymorphisms in related circadian rhythm genes and clopidogrel resistance susceptibility.

Previous studies have confirmed that a dynamic change in circadian rhythm will affect platelet activity, resulting in clopidogrel resistance (CR). We attempted to evaluate whether polymorphisms of related circadian rhythm genes are involved in CR in stable coronary artery disease (SCAD) patients. A sum of 204 SCAD patients met our requirements and were recruited, and 96 patients were considered to have CR. After clinical data collection and platelet function evaluation, genomic DNA was isolated from human peripheral blood, and 23 tagSNPs from related circadian rhythm genes were genotyped by GenomeLab SNPstream Genotyping System. After RNA isolation, relative expression of related gene mRNAs (CLOCK, CRY1, CACNA1C and PRKCG) was measured by real-time PCR. The results showed that polymorphisms in CRY1, CACNA1C and PRKCG changed the response to clopidogrel. And then, the rs1801260 polymorphism might lead to higher mRNA expression in CLOCK and potentially induce the occurrence of CR. Additionally, the TC genotype of rs3745406 might lower mRNA expression of PRKCG, resulting in CR. These findings support the hypothesized role of circadian rhythm genes in CR and indicate probable biomarkers for CR susceptibility, providing new insight into individualized medicine for coronary heart disease.

Temporal Lobe Epilepsy Shows Distinct Functional Connectivity Patterns in Different Thalamic Nuclei.

Background: The thalamus, as a key relay of neuronal information flow between subcortical structures and cortical networks, has been implicated in focal limbic seizures propagation, awareness maintenance, and seizure-related cognitive deficits. However, the specific functional alterations between different thalamic nuclei and subcortical-cortical systems in temporal lobe epilepsy (TLE) remain largely unknown. Methods: We examined thalamic functional connectivity (FC) in 26 TLE patients and 30 healthy controls matched for sex, age, and education. The anterior (ANT), ventral posterior medial, and central lateral nuclei of thalamus were employed to establish whole-brain seed-to-voxel thalamic FC maps. Secondary Pearson's correlation analysis was conducted to assess associations between the abnormal thalamic FC and the memory performance in TLE. Results: Seed-based FC analyses revealed typical distinct FC patterns within each thalamic nuclei in both controls and TLE patients. The TLE showed significantly decreased FC between different thalamic nuclei and subcortical-cortical networks, including the limbic structures, midbrain, sensorimotor network, medial prefrontal cortex, temporal-occipital fusiform gyrus, and cerebellum. Verification analyses yielded similar patterns of thalamic FC changes in TLE. Importantly, the decreased FC between the ANT and hippocampal pathway was correlated with the poorer memory performance of TLE. Conclusion: These findings suggest that the distinct thalamocortical FC patterns are damaged to some extent in TLE patients. Importantly, the specific pathology of the ANT-hippocampal pathway in TLE may be a potential factor that contributes to memory deficits. Our study may pave the way for improved treatments and cognitive function by directly targeting different thalamocortical circuits for TLE.

Deep Natural Image Reconstruction from Human Brain Activity Based on Conditional Progressively Growing Generative Adversarial Networks.

Brain decoding based on functional magnetic resonance imaging has recently enabled the identification of visual perception and mental states. However, due to the limitations of sample size and the lack of an effective reconstruction model, accurate reconstruction of natural images is still a major challenge. The current, rapid development of deep learning models provides the possibility of overcoming these obstacles. Here, we propose a deep learning-based framework that includes a latent feature extractor, a latent feature decoder, and a natural image generator, to achieve the accurate reconstruction of natural images from brain activity. The latent feature extractor is used to extract the latent features of natural images. The latent feature decoder predicts the latent features of natural images based on the response signals from the higher visual cortex. The natural image generator is applied to generate reconstructed images from the predicted latent features of natural images and the response signals from the visual cortex. Quantitative and qualitative evaluations were conducted with test images. The results showed that the reconstructed image achieved comparable, accurate reproduction of the presented image in both high-level semantic category information and low-level pixel information. The framework we propose shows promise for decoding the brain activity.

'When' and 'what' did you see? A novel fMRI-based visual decoding framework.

OBJECTIVE: Visual perception decoding plays an important role in understanding our visual systems. Recent functional magnetic resonance imaging (fMRI) studies have made great advances in predicting the visual content of the single stimulus from the evoked response. In this work, we proposed a novel framework to extend previous works by simultaneously decoding the temporal and category information of visual stimuli from fMRI activities. APPROACH: 3 T fMRI data of five volunteers were acquired while they were viewing five categories of natural images with random presentation intervals. For each subject, we trained two classification-based decoding modules that were used to identify the occurrence time and semantic categories of the visual stimuli. In each module, we adopted recurrent neural network (RNN), which has proven to be highly effective for learning nonlinear representations from sequential data, for the analysis of the temporal dynamics of fMRI activity patterns. Finally, we integrated the two modules into a complete framework. MAIN RESULTS: The proposed framework shows promising decoding performance. The average decoding accuracy across five subjects was over 19 times the chance level. Moreover, we compared the decoding performance of the early visual cortex (eVC) and the high-level visual cortex (hVC). The comparison results indicated that both eVC and hVC participated in processing visual stimuli, but the semantic information of the visual stimuli was mainly represented in hVC. SIGNIFICANCE: The proposed framework advances the decoding of visual experiences and facilitates a better understanding of our visual functions.