NeuroSeg-III: efficient neuron segmentation in two-photon Ca2+ imaging data using self-supervised learning.

Two-photon Ca2+ imaging technology increasingly plays an essential role in neuroscience research. However, the requirement for extensive professional annotation poses a significant challenge to improving the performance of neuron segmentation models. Here, we present NeuroSeg-III, an innovative self-supervised learning approach specifically designed to achieve fast and precise segmentation of neurons in imaging data. This approach consists of two modules: a self-supervised pre-training network and a segmentation network. After pre-training the encoder of the segmentation network via a self-supervised learning method without any annotated data, we only need to fine-tune the segmentation network with a small amount of annotated data. The segmentation network is designed with YOLOv8s, FasterNet, efficient multi-scale attention mechanism (EMA), and bi-directional feature pyramid network (BiFPN), which enhanced the model's segmentation accuracy while reducing the computational cost and parameters. The generalization of our approach was validated across different Ca2+ indicators and scales of imaging data. Significantly, the proposed neuron segmentation approach exhibits exceptional speed and accuracy, surpassing the current state-of-the-art benchmarks when evaluated using a publicly available dataset. The results underscore the effectiveness of NeuroSeg-III, with employing an efficient training strategy tailored for two-photon Ca2+ imaging data and delivering remarkable precision in neuron segmentation.

MultiChannelSleepNet: A Transformer-Based Model for Automatic Sleep Stage Classification With PSG.

Automatic sleep stage classification plays an essential role in sleep quality measurement and sleep disorder diagnosis. Although many approaches have been developed, most use only single-channel electroencephalogram signals for classification. Polysomnography (PSG) provides multiple channels of signal recording, enabling the use of the appropriate method to extract and integrate the information from different channels to achieve higher sleep staging performance. We present a transformer encoder-based model, MultiChannelSleepNet, for automatic sleep stage classification with multichannel PSG data, whose architecture is implemented based on the transformer encoder for single-channel feature extraction and multichannel feature fusion. In a single-channel feature extraction block, transformer encoders extract features from time-frequency images of each channel independently. Based on our integration strategy, the feature maps extracted from each channel are fused in the multichannel feature fusion block. Another set of transformer encoders further capture joint features, and a residual connection preserves the original information from each channel in this block. Experimental results on three publicly available datasets demonstrate that our method achieves higher classification performance than state-of-the-art techniques. MultiChannelSleepNet is an efficient method to extract and integrate the information from multichannel PSG data, which facilitates precision sleep staging in clinical applications.

Design and Fabrication of a High-Temperature SOI Pressure Sensor with Optimized Crossbeam Membrane.

This paper presents a SOI piezoresistive pressure sensor with the crossbeam membrane. The roots of the crossbeam were widened, which solved the problem of the poor dynamic performance of small-range pressure sensors working at a high temperature of 200 °C. A theoretical model was established to optimize the proposed structure, which combined the finite element and the curve fitting. Using the theoretical model, the structural dimensions were optimized to obtain the optimal sensitivity. During optimization, the sensor nonlinearity was also taken into consideration. The sensor chip was fabricated by MEMS bulk-micromachining technology, and Ti/Pt/Au metal leads were prepared to improve the sensor ability of high-temperature resistance over a long time. The sensor chip was packaged and tested, and the experimental results show the sensor achieved an accuracy of 0.241% FS, nonlinearity of 0.180% FS, hysteresis of 0.086% FS and repeatability of 0.137% FS at the high temperature. Given the good reliability and performance at the high temperature, the proposed sensor provides a suitable alternative for the measurement of pressure at high temperatures.

Hippocampal ripples correlate with memory performance in humans.

Memory performance evaluation has generally been based on behavioral tests in the past decades. However, its neural correlates remain largely unknown, particularly in humans. Here we addressed this question using intracranial electroencephalography in patients with refractory epilepsy, performing an episodic memory test. We used the presurgical Wechsler Memory Scale (WMS) test to assess the memory performance of each patient. We found that hippocampal ripples significantly exhibited a transient increase during visual stimulation or before verbal recall. This increase in hippocampal ripples positively correlated with memory performance. By contrast, memory performance is negatively correlated with hippocampal interictal epileptic discharges (IEDs) or epileptic ripples in the memory task. However, these correlations were not present during quiet wakefulness. Thus, our findings uncover the neural correlates of memory performance in addition to traditional behavioral tests.

Femtosecond Laser Processing Assisted SiC High-Temperature Pressure Sensor Fabrication and Performance Test.

Due to material plastic deformation and current leakage at high temperatures, SOI (silicon-on-insulator) and SOS (silicon-on-sapphire) pressure sensors have difficulty working over 500 °C. Silicon carbide (SiC) is a promising sensor material to solve this problem because of its stable mechanical and electrical properties at high temperatures. However, SiC is difficult to process which hinders its application as a high-temperature pressure sensor. This study proposes a piezoresistive SiC pressure sensor fabrication method to overcome the difficulties in SiC processing, especially deep etching. The sensor was processed by a combination of ICP (inductive coupled plasma) dry etching, high-temperature rapid annealing and femtosecond laser deep etching. Static and dynamic calibration tests show that the accuracy error of the fabricated sensor can reach 0.33%FS, and the dynamic signal response time is 1.2 µs. High and low temperature test results show that the developed sensor is able to work at temperatures from -50 °C to 600 °C, which demonstrates the feasibility of the proposed sensor fabrication method.

Experimental Investigation on Ablation of 4H-SiC by Infrared Femtosecond Laser.

Femtosecond laser ablation has become one of the important structural processing methods for the third-generation semiconductor material, silicon carbide (SiC), and it is gradually being employed in the manufacture of microelectromechanical systems and microelectronic devices. Experimental study has been performed on infrared single and multiple pulses (1035 nm) femtosecond laser ablation of SiC at various processing parameters. Diameters of laser ablation spots on 4H-SiC were measured to estimate the absorption threshold for material modification and structural transformation, which were 2.35 J/cm2 and 4.97 J/cm2, respectively. In the multiple-pulse scribing ablation for microgrooves, the ablation threshold dropped to 0.70 J/cm2 due to the accumulation effect when the effective pulse number reached 720. The calculated average of the thermally stimulated ablation depth of 4H-SiC is 22.4 nm, which gradually decreased with the raising of the effective pulse number. For obtaining square trenches with precise and controllable depths and a smooth bottom in 4H-SiC, the effects of processing parameters on the material removal rate and surface roughness are discussed. The ablation rate per pulse is almost constant, even if the effective pulse number varies. The reduction of laser spot overlapping ratio in x direction has a greater weakening effect on the material removal rate than that in y direction. The precise amount of material removal can still be controlled, while modulating the surface roughness of the ablated features by changing the hatch rotation angle. This research will help to achieve controllable, accurate, and high-quality machining results in SiC ablation, using infrared femtosecond laser.

Design and Manufacturing of a High-Sensitivity Cutting Force Sensor Based on AlSiCO Ceramic.

On-line cutting force measurement is an effective way to monitor processing quality, improve processing accuracy, and protect the tool. In high-speed and ultra-precision machining, status monitoring is particularly necessary to ensure machining accuracy. However, the cutting force is very small in high speed and ultra-precision machining. Therefore, high-sensitivity cutting force sensors are needed. Current commercial cutting force sensors have defects such as large volume, low compatibility, and high price. In particular, the sensitivity of cutting force sensor needs to be improved for high-speed and ultra-precision machining status monitoring. This paper provides a possible solution by embedding the sensor in the tool and selecting sensitive materials with high piezoresistive coefficient. In this paper, the structural design of the sensor and the fabrication of the sensitive material SiAlCO ceramic are carried out, and then the sensor is packaged and tested. The test results show that the cutting force sensor's sensitivity was as high as 219.38 mV/N, which is a feasible way to improve cutting force sensor's compatibility and sensitivity.

Design and Fabrication of Bulk Micromachined 4H-SiC Piezoresistive Pressure Chips Based on Femtosecond Laser Technology.

Silicon carbide (SiC) has promising potential for pressure sensing in a high temperature and harsh environment due to its outstanding material properties. In this work, a 4H-SiC piezoresistive pressure chip fabricated based on femtosecond laser technology was proposed. A 1030 nm, 200 fs Yb: KGW laser with laser average powers of 1.5, 3 and 5 W was used to drill blind micro holes for achieving circular sensor diaphragms. An accurate per lap feed of 16.2 µm was obtained under laser average power of 1.5 W. After serialized laser processing, the machining depth error of no more than 2% and the surface roughness as low as 153 nm of the blind hole were measured. The homoepitaxial piezoresistors with a doping concentration of 1019 cm-3 were connected by a closed-loop Wheatstone bridge after a rapid thermal annealing process, with a specific contact resistivity of 9.7 × 10-5 Ω cm2. Our research paved the way for the integration of femtosecond laser micromachining and SiC pressure sensor chips manufacturing.

Differential Expression Hallmarks of Interneurons in Different Types of Focal Cortical Dysplasia.

Focal cortical dysplasia (FCD) is the main cause of medically intractable pediatric epilepsy. Previous studies have suggested that alteration of cortical interneurons and abnormal cytoarchitecture have been linked to initiation and development for seizure. However, whether each individual subpopulation of cortical interneurons is linked to distinct FCD subtypes remains largely unknown. Here, we retrospectively analyzed both control samples and epileptic specimens pathologically diagnosed with FCD types Ia, IIa, or IIb. We quantified three major interneuron (IN) subpopulations, including parvalbumin (PV)-, somatostatin (Sst)-, and vasoactive intestinal peptide (Vip)-positive INs across all the subgroups. Additionally, we calculated the ratio of the subpopulations of INs to the major INs (mINs) by defining the total number of the PV-, Sst-, and Vip-INs as mINs. Compared with the control, the density of the PV-INs in FCD type IIb was significantly lower, and the ratio of PV/mINs was lower in the superficial part of the cortex of the FCD type Ia and IIb groups. Interestingly, we found a significant increase in the ratio of Vip/mINs only in FCD type IIb. Overall, these results suggest that in addition to a reduction in PV-INs, the increase in Vip/mINs may be related to the initiation of epilepsy in FCD type IIb. Furthermore, the increase in Vip/mINs in FCD type IIb may, from the IN development perspective, indicate that FCD type IIb forms during earlier stages of pregnancy than FCD type Ia.

Upregulated SHP-2 expression in the epileptogenic zone of temporal lobe epilepsy and various effects of SHP099 treatment on a pilocarpine model.

Temporal lobe epilepsy (TLE) is defined as the sporadic occurrence of spontaneous recurrent seizures, and its pathogenesis is complex. SHP-2 (Src homology 2-containing protein tyrosine phosphatase 2) is a widely expressed cytosolic tyrosine phosphatase protein that participates in the regulation of inflammation, angiogenesis, gliosis, neurogenesis and apoptosis, suggesting a potential role of SHP-2 in TLE. Therefore, we investigated the expression patterns of SHP-2 in the epileptogenic brain tissue of intractable TLE patients and the various effects of treatment with the SHP-2-specific inhibitor SHP099 on a pilocarpine model. Western blotting and immunohistochemistry results confirmed that SHP-2 expression was upregulated in the temporal neocortex of patients with TLE. Double-labeling experiments revealed that SHP-2 was highly expressed in neurons, astrocytes, microglia and vascular endothelial cells in the epileptic foci of TLE patients. In the pilocarpine-induced C57BL/6 mouse model, SHP-2 upregulation in the hippocampus began one day after status epilepticus, reached a peak at 21 days and then maintained a significantly high level until day 60. Similarly, we found a remarkable increase in SHP-2 expression at 1, 7, 21 and 60 days post-SE in the temporal neocortex. In addition, we also showed that SHP099 increased reactive gliosis, the release of IL-1ß, neuronal apoptosis and neuronal loss, while reduced neurogenesis and albumin leakage. Taken together, the increased expression of SHP-2 in the epileptic zone may be involved in the process of TLE.

Elevated Expression of TRPC4 in Cortical Lesions of Focal Cortical Dysplasia II and Tuberous Sclerosis Complex.

Focal cortical dysplasia type II (FCD II) and tuberous sclerosis complex (TSC) are well-known causes of chronic refractory epilepsy in children. Canonical transient receptor potential channels (TRPCs) are non-selective cation channels that are commonly activated by phospholipase C (PLC) stimulation. Previous studies found that TRPC4 may participate in the process of epileptogenesis. This study aimed to examine the expression and distribution of TRPC4 in FCD II (n = 24) and TSC (n = 11) surgical specimens compared with that in age-matched autopsy control samples (n = 12). We found that the protein levels of TRPC4 and its upstream factor, PLC delta 1 (PLCD1), were elevated in FCD II and TSC samples compared to those of control samples. Immunohistochemistry assays revealed that TRPC4 staining was stronger in malformed cells, such as dysmorphic neurons, balloon cells and giant cells. Moderate-to-strong staining of the upstream factor PLCD1 was also identified in abnormal neurons. Moreover, double immunofluorescence staining revealed that TRPC4 was colocalised with glutamatergic and GABAergic neuron markers. Taken together, our results indicate that overexpression of TRPC4 protein may be involved in the epileptogenesis of FCD II and TSC.

Activation of LILRB2 signal pathway in temporal lobe epilepsy patients and in a pilocarpine induced epilepsy model.

Temporal lobe epilepsy (TLE) is a frequent form of focal intractable epilepsy in adults, but the specific mechanism underlying the epileptogenesis of TLE is still unknown. Human leukocyte immunoglobulin-like receptor B2 (LILRB2) (the murine homolog gene called paired immunoglobulin-like receptor B, or PirB), participates in the process of synaptic plasticity and neurite growth in the central nervous system (CNS), suggesting a potential role of LILRB2 in epilepsy. However, the expression pattern of LILRB2 and the downstream molecular signal in intractable TLE remains poorly understood. In the present study, western blotting and immunohistochemistry results showed that LILRB2 expression was upregulated in the temporal neocortex of patients with TLE. Moreover, protein levels of LILRB2 negatively correlated with the frequency of seizures in TLE patients. In the pilocarpine-induced C57BL/6 mouse model, PirB upregulation in the hippocampus began 12h after status epilepticus (SE), reached a peak at 7days and then maintained a significantly high level until day 60. Similarly, we found a remarkable increase in PirB expression at 1day, 7days and30days post-SE in the temporal cortex. Double-labeled immunofluorescence showed that LILRB2/PirB were highly expressed in neurons and astrocytes but not microglia. In addition, protein levels of POSH, SHROOM3, ROCK1 and ROCK2, the important downstream factors of the LILRB2 pathway, were significantly increased in the epileptic foci of TLE patients and located on the NeuN-positive neurons and GFAP-positive astrocytes. Taken together, our results indicate that LILRB2/PirB may be involved in the process of TLE.

Down-Regulated Expression of Liver X Receptor beta in Cortical Lesions of Patients with Focal Cortical Dysplasia.

Focal cortical dysplasia (FCD) is strongly associated with medically intractable epilepsy. Studies suggest that liver X receptor beta (LXRß) may participate in the pathogenesis of FCD. The present study investigated the expression pattern of LXRß in FCD and the distribution of LXRß in different neural precursor cells. Twenty-five surgical specimens from FCD patients and 11 age-matched control samples from autopsies were included in our study. Protein levels and distribution were detected by western blot, immunohistochemistry, and immunofluorescence staining. We found that (1) the level of LXRß protein was markedly reduced in FCD. (2) LXRß staining was weaker in the dysplastic cortices of FCD and was mainly observed in neuronal microcolumns, and malformed cells. (3) LXRß was co-localized with radial glial cells (RGCs) markers and oligodendrocyte precursor cells (OPCs) markers in malformed cells. (4) RGCs marker and OPCs marker were down-regulated while LXRß downstream factors were up-regulated in FCD specimens. Taken together, our results indicate that LXRß may interact with ß-catenin to regulate the generation of OPCs and the transformation of RGCs. LXRß therefore potentially contributes to the pathogenesis of FCD.

Expression and cellular distribution of transient receptor potential vanilloid 4 in cortical tubers of the tuberous sclerosis complex.

Cortical tubers in patients with tuberous sclerosis complex (TSC) are highly associated with intractable epilepsy. Recent evidence has shown that transient receptor potential vanilloid 4 (TRPV4) has direct effects on both neurons and glial cells. To understand the role of TRPV4 in pathogenesis of cortical tubers, we investigated the expression patterns of TRPV4 in cortical tubers of TSC compared with normal control cortex (CTX). We found that TRPV4 was clearly up-regulated in cortical tubers at the protein levels. Immunostaining indicated that TRPV4 was specially distributed in abnormal cells, including dysplastic neurons (DNs) and giant cells (GCs). In addition, double immunofluorescent staining revealed that TRPV4 was localized on neurofilament proteins (NF200) positive neurons and glial fibrillary acidic portein (GFAP) positive reactive astrocytes. Moreover, TRPV4 co-localized with both glutamatergic and GABAergic neurons. Furthermore, protein levels of protein kinase C (PKC), but not protein kinase A (PKA), the important upstream factors of the TRPV4, were significantly increased in cortical tubers. Taken together, the overexpression and distribution patterns of TRPV4 may be linked with the intractable epilepsy caused by TSC.

Increased Expression of Transient Receptor Potential Vanilloid 4 in Cortical Lesions of Patients with Focal Cortical Dysplasia.

AIM: Focal cortical dysplasia (FCD) represents a well-known cause of medically intractable epilepsy. Studies found that transient receptor potential vanilloid receptor 4 (TRPV4) may participate in the occurrence of seizures. This study investigated the expression patterns of TRPV4 in FCD and the cascade that regulate functional state of TRPV4 in cortical neurons. METHODS: Thirty-nine surgical specimens from FCD patients and 10 age-matched control samples from autopsies were included in this study. Protein expression and distribution were detected by Western blot, immunohistochemistry, and immunofluorescence staining. Calcium imaging was used to detect the TRPV4-mediated Ca(2+) influx in cortical neurons. RESULTS: (1) The protein levels of TRPV4 and of an upstream factor, protein kinase C (PKC), were markedly elevated in FCD. (2) TRPV4 staining was stronger in the dysplastic cortices of FCD and mainly observed in neuronal microcolumns and malformed cells. (3) The activation of TRPV4 was central for [Ca(2+)]i elevation in cortical neurons, and this activity of TRPV4 in cortical neurons was regulated by the PKC, but not the PKA, pathway. CONCLUSION: The overexpression and altered cellular distribution of TRPV4 in FCD suggest that TRPV4 may potentially contribute to the epileptogenesis of FCD.