Human-multimodal deep learning collaboration in 'precise' diagnosis of lupus erythematosus subtypes and similar skin diseases.

BACKGROUND: Lupus erythematosus (LE) is a spectrum of autoimmune diseases. Due to the complexity of cutaneous LE (CLE), clinical skin image-based artificial intelligence is still experiencing difficulties in distinguishing subtypes of LE. OBJECTIVES: We aim to develop a multimodal deep learning system (MMDLS) for human-AI collaboration in diagnosis of LE subtypes. METHODS: This is a multi-centre study based on 25 institutions across China to assist in diagnosis of LE subtypes, other eight similar skin diseases and healthy subjects. In total, 446 cases with 800 clinical skin images, 3786 multicolor-immunohistochemistry (multi-IHC) images and clinical data were collected, and EfficientNet-B3 and ResNet-18 were utilized in this study. RESULTS: In the multi-classification task, the overall performance of MMDLS on 13 skin conditions is much higher than single or dual modals (Sen = 0.8288, Spe = 0.9852, Pre = 0.8518, AUC = 0.9844). Further, the MMDLS-based diagnostic-support help improves the accuracy of dermatologists from 66.88% ± 6.94% to 81.25% ± 4.23% (p = 0.0004). CONCLUSIONS: These results highlight the benefit of human-MMDLS collaborated framework in telemedicine by assisting dermatologists and rheumatologists in the differential diagnosis of LE subtypes and similar skin diseases.

Amyloid-ß Protein Precursor Regulates Electrophysiological Properties in the Hippocampus via Altered Kv1.4 Expression and Function in Mice.

BACKGROUND: Amyloid-ß protein precursor (AßPP) is enriched in neurons. However, the mechanism underlying AßPP regulation of neuronal activity is poorly understood. Potassium channels are critically involved in neuronal excitability. In hippocampus, A-type potassium channels are highly expressed and involved in determining neuronal spiking. OBJECTIVE: We explored hippocampal local field potential (LFP) and spiking in the presence and absence of AßPP, and the potential involvement of an A-type potassium channel. METHODS: We used in vivo extracellular recording and whole-cell patch-clamp recording to determine neuronal activity, current density of A-type potassium currents, and western blot to detect changes in related protein levels. RESULTS: Abnormal LFP was observed in AßPP-/- mice, including reduced beta and gamma power, and increased epsilon and ripple power. The firing rate of glutamatergic neurons reduced significantly, in line with an increased action potential rheobase. Given that A-type potassium channels regulate neuronal firing, we measured the protein levels and function of two major A-type potassium channels and found that the post-transcriptional level of Kv1.4, but not Kv4.2, was significantly increased in the AßPP-/- mice. This resulted in a marked increase in the peak time of A-type transient outward potassium currents in both glutamatergic and gamma-aminobutyric acid-ergic (GABAergic) neurons. Furthermore, a mechanistic experiment using human embryonic kidney 293 (HEK293) cells revealed that the AßPP deficiency-induced increase in Kv1.4 may not involve protein-protein interaction between AßPP and Kv1.4. CONCLUSION: This study suggests that AßPP modulates neuronal firing and oscillatory activity in the hippocampus, and Kv1.4 may be involved in mediating the modulation.

A reflective display based on the electro-microfluidic assembly of particles within suppressed water-in-oil droplet array.

Reflective displays have stimulated considerable interest because of their friendly readability and low energy consumption. Herein, we develop a reflective display technique via an electro-microfluidic assembly of particles (eMAP) strategy whereby colored particles assemble into annular and planar structures inside a dyed water droplet to create "open" and "closed" states of a display pixel. Water-in-oil droplets are compressed within microwells to form a pixel array. The particles dispersed in droplets are driven by deformation-strengthened dielectrophoretic force to achieve fast and reversible motion and assemble into multiple structures. This eMAP based device can display designed information in three primary colors with ≥170° viewing angle, ~0.14 s switching time, and bistability with an optimized material system. This proposed technique demonstrates the basis of a high-performance and energy-saving reflective display, and the display speed and color quality could be further improved by structure and material optimization; exhibiting a potential reflective display technology.

Controllable ingestion and release of guest components driven by interfacial molecular orientation of host liquid crystal droplets.

Controllable construction and manipulation of artificial multi-compartmental structures are crucial in understanding and imitating smart molecular elements such as biological cells and on-demand delivery systems. Here, we report a liquid crystal droplet (LCD) based three-dimensional system for controllable and reversible ingestion and release of guest aqueous droplets (GADs). Induced by interfacial thermodynamic fluctuation and internal topological defect, microscale LCDs with perpendicular anchoring condition at the interface would spontaneously ingest external components from the surroundings and transform them as radially assembled tiny GADs inside LCDs. Landau-de Gennes free-energy model is applied to describe and explain the assembly dynamics and morphologies of these tiny GADs, which presents a good agreement with experimental observations. Furthermore, the release of these ingested GADs can be actively triggered by changing the anchoring conditions at the interface of LCDs. Since those ingestion and release processes are controllable and happen very gently at room temperature and neutral pH environment without extra energy input, these microscale LCDs are very prospective to provide a unique and viable route for constructing hierarchical 3D structures with tunable components and compartments.

A skin in situ immune cell detection kit for the diagnosis and classification of cutaneous lupus erythematosus.

BACKGROUND: Although lupus can be diagnosed by first impression, medical history, physical examination, pathological analysis and laboratory tests, the accurate classification of cutaneous lupus erythematosus (LE) is still a major challenge in the clinic, which might mislead the selection of treatments and miss the right time for the administration of therapies. The goal of this study was to establish a novel kit to assist with the diagnosis and classification of cutaneous lupus. METHODS: Sixty-five patients from three hospitals were included in this study, including 50 patients with LE and other similar skin diseases. We invited two dermatology specialists to make an accurate diagnosis of the subtypes of lupus based on the patient's clinical features, laboratory examination tests, pathology manifestation analysis, medical treatments and follow-up records. Then, we used their diagnosis results as a standard to which we successively compared the consistency of each step of our diagnosis processes, including impression diagnosis, pathology diagnosis, the combined consideration of the former two diagnostic analyses, and the results of an in situ immune cell detection kit to assist in arriving at a judgement. RESULTS: By Cohen's kappa analysis, we found that the results of the in situ immune cell detection kit had the highest consistency with the diagnoses of the two specialists, both for the diagnosis (k=0.921) and for the classification of cutaneous lupus (k=0.940). In addition, this kit enhanced the LE classification accuracy by 36.3% compared with the diagnostic accuracy of impression diagnosis combined with only pathological analysis. CONCLUSIONS: This skin in situ immune cell detection kit may assist doctors in achieving a higher diagnostic performance and price ratio and enhance their diagnostic efficiency.

Amyloid-ß Protein Precursor Deficiency Changes Neuronal Electrical Activity and Levels of Mitochondrial Proteins in the Medial Prefrontal Cortex.

BACKGROUND: Neuropathological features of Alzheimer's disease are characterized by the deposition of amyloid-ß (Aß) plaques and impairments in synaptic activity and memory. However, we know little about the physiological role of amyloid-ß protein precursor (AßPP) from which Aß derives. OBJECTIVE: Evaluate APP deficiency induced alterations in neuronal electrical activity and mitochondrial protein expression. METHODS: Utilizing electrophysiological, biochemical, pharmacological, and behavioral tests, we revealed aberrant local field potential (LFP), extracellular neuronal firing and levels of mitochondrial proteins. RESULT: We show that APP knockout (APP-/-) leads to increased gamma oscillations in the medial prefrontal cortex (mPFC) at 1-2 months old, which can be restored by baclofen (Bac), a γ-aminobutyric acid type B receptor (GABABR) agonist. A higher dose and longer exposure time is required for Bac to suppress neuronal firing in APP-/- mice than in wild type animals, indicating enhanced GABABR mediated activity in the mPFC of APP-/- mice. In line with increased GABABR function, the glutamine synthetase inhibitor, L-methionine sulfonate, significantly increases GABABR levels in the mPFC of APP-/- mice and this is associated with a significantly lower incidence of death. The results suggest that APP-/- mice developed stronger GABABR mediated inhibition. Using HEK 293 as an expression system, we uncover that AßPP functions to suppress GABABR expression. Furthermore, APP-/- mice show abnormal expression of several mitochondrial proteins. CONCLUSION: APP deficiency leads to both abnormal network activity involving defected GABABR and mitochondrial dysfunction, suggesting critical role of AßPP in synaptic and network function.