Alternative splicing changes the Ca1.2 calcium channel electrophysiological property, but the in vivo significance of such altered channel function is lacking. Structure-function studies of heterologously expressed Ca1.2 channels could not recapitulate channel function in the native milieu of the cardiomyocyte. To address this gap in knowledge, we investigated the role of alternative exon 33 of the Ca1.2 calcium channel in heart function. Exclusion of exon 33 in Ca1.2 channels has been reported to shift the activation potential -10.4 mV to the hyperpolarized direction, and increased expression of Ca1.2 channels was observed in rat myocardial infarcted hearts. However, how a change in Ca1.2 channel electrophysiological property, due to alternative splicing, might affect cardiac function in vivo is unknown. To address these questions, we generated m exon 33-null mice. These mice contained Ca1.2 channels with a gain-of-function that included conduction of larger currents that reflects a shift in voltage dependence and a modest increase in single-channel open probability. This altered channel property underscored the development of ventricular arrhythmia, which is reflected in significantly more deaths of exon 33 mice from β-adrenergic stimulation. In vivo telemetric recordings also confirmed increased frequencies in premature ventricular contractions, tachycardia, and lengthened QT interval. Taken together, the significant decrease or absence of exon 33-containing Ca1.2 channels is potentially proarrhythmic in the heart. Of clinical relevance, human ischemic and dilated cardiomyopathy hearts showed increased inclusion of exon 33. However, the possible role that inclusion of exon 33 in Ca1.2 channels may play in the pathogenesis of human heart failure remains unclear.

Extranodal nasal natural killer (NK)/T-cell lymphoma is a very rare kind of lymphoma, Oral cavity involvement of extranodal natural killer/T-cell lymphoma, nasal type is extremely rare, and its clinicopathologic features are also poorly understood. Recently, we experienced an unusual case of Epstein-Barr virus-associated, extranodal NK/T-cell type with a unhealed palatal ulcer as the earliest clinical feature. It is a challenge for oral medicine specialists to make the early diagnosis for this special type of tumor.

Benefiting from the event-driven nature and sparse spiking communication of the brain, spiking neural networks (SNNs) are becoming a promising energy-efficient alternative to traditional artificial neural networks (ANNs). However, the performance gap between SNNs and ANNs has been a significant hindrance to deploying SNNs ubiquitously for a very long period of time. To leverage the full potential of SNNs, we study the effect of attention mechanisms in SNNs, which makes they can focus on important information. We first present our idea of attention in SNNs with a plug-and-play combined module kit, termed the Multi-dimensional Attention (MA) module. Then, a new attention SNN architecture with end-to-end training called "MA-SNN" is proposed, which infers attention weights along the temporal dimension, channel dimension, as well as spatial dimension separately or simultaneously. Based on the existing neuroscience theories, we exploit the attention weights to optimize membrane potentials, which in turn regulate the spiking response in a data-dependent way. At the cost of negligible additional parameters, MA facilitates vanilla SNNs to achieve sparser spiking activity, better performance, and energy efficiency concurrently. Experiments are conducted in event-based DVS128 Gesture/Gait action recognition and ImageNet-1k image classification. On Gesture/Gait, the spike counts are reduced by 84.9%/81.6%, the task accuracy and energy efficiency are improved by 5.9%/4.7% and 3.4×/3.2×. On ImageNet-1K, we achieve top-1 accuracy of 75.92% and 77.08% on single/4-step Res-SNN-104, which are state-of-the-art results in SNNs. Compared with counterpart Res-ANN-104, the performance gap becomes -0.95/+0.21 percent and has 31.8×/7.4× better energy efficiency. To our best knowledge, this is for the first time, that the SNN community achieves comparable or even better performance compared with its ANN counterpart in the large-scale dataset. To analyze and support the effectiveness of MA-SNN, we theoretically prove that the spiking degradation or the gradient vanishing, which usually holds in general SNNs, can be resolved by introducing the block dynamical isometry theory. We also analyze the efficiency of MA-SNN based on our proposed spiking response visualization method. Our work lights up SNN's potential as a general backbone to support various applications in the field of SNN research, with a great balance between effectiveness and efficiency.

No abstract

No abstract