The pathology of Alzheimer's Disease (AD) is characterized by aggregates of amyloid beta (Aβ) peptides and neurofibrillary tau tangles. Increased blood-brain barrier (BBB) permeability and reduced Aβ clearance, which signal neurovascular dysfunction, have also been proposed as early markers of AD. Despite intense scrutiny, the mechanisms of AD remain elusive and novel treatments that address core symptoms of dementia are limited. New alternative methods (NAMs) aim to develop in-vitro translational models that recapitulate human pathology more accurately than previous models and could contribute to the development of new therapies. Here, we developed a NAM model of the cortical neurovascular unit (NVU) using brain cells derived from human induced pluripotent stem cells (hiPSCs) from a patient with AD and a healthy individual. Differentiated neurons, astrocytes, pericytes, microglia, and brain-like microvascular endothelial cells were cultured in a microphysiological system to create a brain-chip model to evaluate NVU-related endpoints. Compared to control, AD brain-chips had reduced claudin-5 and ZO-1 expression and increased paracellular permeability. AD brain-chips also had decreased activity of the efflux transporter P-glycoprotein (P-gp), but its expression was unchanged. In AD brain-chips, levels of Aβ42, total tau, and p-tau 181 were decreased in protein lysates from the brain channel, while levels of total tau and p-tau 181 were increased in protein lysates from the vascular channel. Finally, AD brain-chips had increased levels of the proinflammatory markers IL-6 and MCP-1 in effluent from both brain and vascular channels. In this brain-chip model, we showed Aβ-independent NVU dysfunction that was related to neuroinflammation and vascular tau accumulation. This study demonstrates the utility of the brain-chip model to evaluate changes in NVU functions induced by AD-like pathology and highlights donor-specific responses associated with the use of hiPSC-derived models.

Predictive processing asserts that the brain learns a generative model of the world, which it uses to make sensory-updated predictions about reality. While traditional views emphasize the cerebral cortex, prediction is a fundamental brain principle, which underscores the vital role of older subcortical structures. This review offers a framework for understanding the brain as an integrated system of semi-independent cortical and subcortical functional units that collectively enable predictive processing. The cerebral cortex is positioned as the primary driver of subconscious predictions, whereas the thalamus, hippocampal complex, amygdala, basal ganglia, and cerebellum contribute critical indirect roles by translating the predictions into conscious, cohesive, and coordinated experiences and behaviours. Specifically, the thalamus controls and establishes selective attention by synchronizing multiple cortical regions, enabling attended predictions to be expressed into conscious perception and cognition; the hippocampal complex captures novelty and constructs episodic simulations, which represent highly abstract or hypothetical predictions that contribute to the conscious cognitive experience; and the amygdala appraises motivational value and activates emotional states, which predict survival-critical events and prime the brain for action, contributing to a subjective emotional experience. During this translation, the basal ganglia and cerebellum contribute sculpting roles, with the basal ganglia chunking predictions into repertoires, facilitating the cohesive expression of actions, and potentially perceptual, cognitive, and emotional experiences, while the cerebellum generates and adjusts temporal predictions, enabling the coordinated expression of actions and experiences. This integrative framework highlights the essential, often-overlooked contributions of subcortical units to predictive processing, providing a unified model for future research.

Purpose: The aim of this study was an exploration of the multiscale vibratory response of the spine following orthopedic surgery in patients with idiopathic scoliosis and postoperative traumatic fatigue injury. Methods: In this paper, the postoperative macroscopic spine model in the modal, time and frequency domains to obtain the vibration response of the patient's entire spine were analyzed. Subsequently, the stresses in the cortical bone mesoscopic bone units around the surgically damaged interface were calculated using submodeling algorithms. The pore stresses and pore flow velocities of the osteocytes were then derived from the stresses of the mesoscopic bone units to predict fatigue damage at the fusion surface. Results: The findings indicated that the first three orders of intrinsic frequency exerted the most significant influence on the spine model. The maximum stress of the bone unit was observed at the X3 bone plate on the left side of the fusion surface, and the maximum pore pressure and flow velocity of the bone cells occurred at the X4 on the right side of the fusion surface. The medical implants used in spinal orthopedics, titanium cages and pedicle nails, change the mobility of the adjacent segments and also create a stress shielding effect that impacts the fusion of bone tissues. Conclusions: Microscopic bone cell synapses experience greater pore pressures and pore flow velocities in the vibration environment compared to those under the static environment, which may promote cell growth. Vibration at low loads typically does not induce fatigue damage to cancellous bone at the fusion surface of medical implants.

Cortical beta band oscillations (13-30 Hz) are associated with sensorimotor control, but their precise role remains unclear. Evidence suggests that for low-threshold motor neurons (MNs), these oscillations are conveyed to muscles via the fastest corticospinal fibers. However, their transmission to MNs of different sizes may vary due to differences in the relative strength of corticospinal and reticulospinal projections across the MN pool. Consequently, it remains uncertain whether corticospinal beta transmission follows similar pathways and maintains consistent strength across the entire MN pool. To investigate this, we examined beta activity in MNs innervating the tibialis anterior muscle across the full range of recruitment thresholds in a study involving 12 participants of both sexes. We characterized beta activity at both the cortical and motor unit (MU) levels, while participants performed contractions from mild to submaximal levels. Corticomuscular coherence remained unchanged across contraction forces after normalizing for the net MU spike rate, suggesting that beta oscillations are transmitted with similar strength to MNs, regardless of size. To further explore beta transmission, we estimated corticospinal delays using the cumulant density function, identifying peak correlations between cortical and muscular activity. Once compensated for variable peripheral axonal propagation delay across MNs, the corticospinal delay remained stable, and its value (∼14 ms) indicated projections through the fastest corticospinal fibers for all MNs. These findings demonstrate that corticospinal beta band transmission is determined by the fastest pathway connecting in the corticospinal tract, projecting similarly across the entire MN pool.

Edge visual systems demand high energy-efficiency vision processors like neuromorphic hardware leveraging spikebased computations. But their disability of directly interacting with non-spike information in the real world requests additional components to execute image pre-processing, spike encoding and decoding, severely increasing overall system cost, energy and latency. To overcome such drawback, this brief proposes a tiny neuromorphic vision processor which emulates functional regions along the ventral pathway in the visual cortex. It performs image pre-processing, spike encoding, spike-based feature extraction and classification, spike decoding as well as decision making on a single chip. To reduce hardware resources, our processor builds on a reconfigurable cortical neuron (RCN) unit, which runs different neuron models for different visual cortex regions in a timemultiplexing fashion. It also embeds biological learning circuits to better adapt the processor to dynamic edge scenarios. Our neuromorphic processor was prototyped on a very-low-cost Xilinx Zynq-7010 device. On the MNIST dataset, it exhibited a real-time inference speed of 696 frame/s covering image pre-processing to final decision and a high on-chip learning accuracy of 97.12%, while only delivering a power consumption as low as 118 mW.

The human perception system features many dynamic functional mechanisms that efficiently process the large amount of sensory information available in the surrounding environment. In this system, sensory adaptation operates as a core mechanism that seamlessly filters familiar and inconsequential external stimuli at sensory endpoints. Such adaptive filtering minimizes redundant data movement between sensory terminals and cortical processing units and contributes to a lower communication bandwidth requirement and lower energy consumption at the system level. Recreating the behavior of sensory adaptation using electronic devices has garnered significant research interest owing to its promising prospects in next-generation intelligent perception platforms. Herein, the recent progress in bioinspired adaptive device engineering is systematically examined, and its valuable applications in electronic skins, wearable electronics, and machine vision are highlighted. The rapid development of bioinspired adaptive sensors can be attributed not only to the recent advances in emerging neuromorphic electronic elements, including piezoelectric and triboelectric sensors, memristive devices, and neuromorphic transistors, but also to the improved understanding of biological sensory adaptation. Existing challenges hindering device performance optimization, multimodal adaptive sensor development, and system-level integration are also discussed, providing insights for the development of high-performance neuromorphic sensing systems.

Cortical encoding of hierarchical linguistic information when syllabic rhythms are obscured by echoes

We introduce Ultra-Flexible Tentacle Electrodes (UFTEs), packing many independent fibers with the smallest possible footprint without limitation in recording depth using a combination of mechanical and chemical tethering for insertion. We demonstrate a scheme to implant UFTEs simultaneously into many brain areas at arbitrary locations without angle-of-insertion limitations, and a 512-channel wireless logger. Immunostaining reveals no detectable chronic tissue damage even after several months. Mean spike signal-to-noise ratios are 1.5-3x compared to the state-of-the-art, while the highest signal-to-noise ratios reach 89, and average cortical unit yields are ~1.75/channel. UFTEs can track the same neurons across sessions for at least 10 months (longest duration tested). We tracked inter- and intra-areal neuronal ensembles (neurons repeatedly co-activated within 25 ms) simultaneously from hippocampus, retrosplenial cortex, and medial prefrontal cortex in freely moving rodents. Average ensemble lifetimes were shorter than the durations over which we can track individual neurons. We identify two distinct classes of ensembles. Those tuned to sharp-wave ripples display the shortest lifetimes, and the ensemble members are mostly hippocampal. Yet, inter-areal ensembles with members from both hippocampus and cortex have weak tuning to sharp wave ripples, and some have unusual months-long lifetimes. Such inter-areal ensembles occasionally remain inactive for weeks before re-emerging.

The sensorimotor gating is a nervous system function that modulates the acoustic startle response (ASR). Prepulse inhibition (PPI) phenomenon is an operational measure of sensorimotor gating, defined as the reduction of ASR when a high intensity sound (pulse) is preceded in milliseconds by a weaker stimulus (prepulse). Brainstem nuclei are associated with the mediation of ASR and PPI, whereas cortical and subcortical regions are associated with their modulation. However, it is still unclear how the modulatory units can influence PPI. In the present work, we developed a computational model of a neural circuit involved in the mediation (brainstem units) and modulation (cortical and subcortical units) of ASR and PPI. The activities of all units were modeled by the leaky-integrator formalism for neural population. The model reproduces basic features of PPI observed in experiments, such as the effects of changes in interstimulus interval, prepulse intensity, and habituation of ASR. The simulation of GABAergic and dopaminergic drugs impaired PPI by their effects over subcortical units activity. The results show that subcortical units constitute a central hub for PPI modulation. The presented computational model offers a valuable tool to investigate the neurobiology associated with disorder-related impairments in PPI.

Patients with mutations in the thyroid hormone (TH) cell transporter monocarboxylate transporter 8 (MCT8) gene develop severe neuropsychomotor retardation known as Allan-Herndon-Dudley syndrome (AHDS). It is assumed that this is caused by a reduction in TH signaling in the developing brain during both intrauterine and postnatal developmental stages, and treatment remains understandably challenging. Given species differences in brain TH transporters and the limitations of studies in mice, we generated cerebral organoids (COs) using human induced pluripotent stem cells (iPSCs) from MCT8-deficient patients. MCT8-deficient COs exhibited (i) altered early neurodevelopment, resulting in smaller neural rosettes with thinner cortical units, (ii) impaired triiodothyronine (T3) transport in developing neural cells, as assessed through deiodinase-3-mediated T3 catabolism, (iii) reduced expression of genes involved in cerebral cortex development, and (iv) reduced T3 inducibility of TH-regulated genes. In contrast, the TH analogs 3,5-diiodothyropropionic acid and 3,3',5-triiodothyroacetic acid triggered normal responses (induction/repression of T3-responsive genes) in MCT8-deficient COs, constituting proof of concept that lack of T3 transport underlies the pathophysiology of AHDS and demonstrating the clinical potential for TH analogs to be used in treating patients with AHDS. MCT8-deficient COs represent a species-specific relevant preclinical model that can be utilized to screen drugs with potential benefits as personalized therapeutics for patients with AHDS.

BackgroundDual-energy CT (DECT)-derived bone mineral density (BMD) of the distal radius and other CT-derived metrics related to bone health have been suggested for opportunistic osteoporosis screening and risk evaluation for sustaining distal radius fractures (DRFs). MethodsThe distal radius of patients who underwent DECT between 01/2016 and 08/2021 was retrospectively analyzed. Cortical Hounsfield Unit (HU), trabecular HU, cortical thickness, and DECT-based BMD were acquired from a non-fractured, metaphyseal area in all examinations. Receiver-operating characteristic (ROC) analysis was conducted to determine the area under the curve (AUC) values for predicting DRFs based on DECT-derived BMD, HU values, and cortical thickness. Logistic regression models were then employed to assess the associations of these parameters with the occurrence of DRFs. ResultsIn this study, 263 patients (median age: 52 years; interquartile range: 36–64; 132 women; 192 fractures) were included. ROC curve analysis revealed a higher area under the curve (AUC) value for DECT-derived BMD compared to cortical HU, trabecular HU, and cortical thickness (0.91 vs. 0.61, 0.64, and 0.69, respectively; p <.001). Logistic regression models confirmed the association between lower DECT-derived BMD and the occurrence of DRFs (Odds Ratio, 0.83; p <.001); however, no influence was observed for cortical HU, trabecular HU, or cortical thickness. ConclusionsDECT can be used to assess the BMD of the distal radius without dedicated equipment such as calibration phantoms to increase the detection rates of osteoporosis and stratify the individual risk to sustain DRFs. In contrast, assessing HU-based values and cortical thickness does not provide clinical benefit.

Model-based trajectory classification of anchored molecular motor-biopolymer interactions

Extracellular recordings were made from 642 units in the primary visual cortex (V1) of a highly visual marsupial, the Tammar wallaby. The receptive field (RF) characteristics of the cells were objectively estimated using the non-linear input model (NIM), and these were correlated with spike shapes. We found that wallaby cortical units had 68% regular spiking (RS), 12% fast spiking (FS), 4% triphasic spiking (TS), 5% compound spiking (CS) and 11% positive spiking (PS). RS waveforms are most often associated with recordings from pyramidal or spiny stellate cell bodies, suggesting that recordings from these cell types dominate in the wallaby cortex. In wallaby, 70-80% of FS and RS cells had orientation selective RFs and had evenly distributed linear and nonlinear RFs. We found that 47% of wallaby PS units were non-orientation selective and they were dominated by linear RFs. Previous studies suggest that the PS units represent recordings from the axon terminals of non-orientation selective cells originating in the lateral geniculate nucleus (LGN). If this is also true in wallaby, as strongly suggested by their low response latencies and bursty spiking properties, the results suggest that significantly more neurons in wallaby LGN are already orientation selective. In wallaby, less than 10% of recorded spikes had triphasic (TS) or sluggish compound spiking (CS) waveforms. These units had a mixture of orientation selective and non-oriented properties, and their cellular origins remain difficult to classify.

Neural plasticity of the brain or its ability to reorganize following injury has likely coincided with the successful clinical correction of severe deformity by facial transplantation since 2005. In this study, we present the cortical reintegration outcomes following syngeneic hemifacial vascularized composite allograft (VCA) in a small animal model. Specifically, changes in the topographic organization and unit response properties of the rodent whisker-barrel somatosensory system were assessed following hemifacial VCA. Clear differences emerged in the barrel-cortex system when comparing naïve and hemiface transplanted animals. Neurons in the somatosensory cortex of transplanted rats had decreased sensitivity albeit increased directional sensitivity compared with naïve rats and evoked responses in transplanted animals were more temporally dispersed. In addition, receptive fields were often topographically mismatched with the indication that the mismatched topography reorganized within adjacent barrel (same row-arc bias following hemifacial transplant). These results suggest subcortical changes in the thalamus and/or brainstem play a role in hemifacial transplantation cortical plasticity and demonstrate the discrete and robust data that can be derived from this clinically relevant small animal VCA model for use in optimizing postsurgical outcomes.NEW & NOTEWORTHY Robust rodent hemifacial transplant model was used to record functional changes in somatosensory cortex after transplantation. Neurons in the somatosensory cortex of face transplant recipients had decreased sensitivity to stimulation of whiskers with increased directional sensitivity vs. naive rats. Transplant recipient cortical unit response was more dispersed in temporary vs. naive rats. Despite histological similarities to naive cortices, transplant recipient cortices had a mix of topographically appropriate and inappropriate whiskered at barrel cortex relationships.

With growing molecular evidence for correlations between spatial arrangement of blood vasculature and fundamental immunological functions, carried out in distinct compartments of the subdivided lymph node, there is an urgent need for three-dimensional models that can link these aspects. We reconstructed such models at a 1.84 µm resolution by the means of X-ray phase-contrast imaging with a 2D Talbot array in a short time without any staining. In addition reconstructions are verified in immunohistochemistry staining as well as in ultrastructural analyses. While conventional illustrations of mammalian lymph nodes depict the hilus as a definite point of blood and lymphatic vessel entry and exit, our method revealed that multiple branches enter and emerge from an area that extends up to one third of the organ's surface. This could be a prerequisite for the drastic and location-dependent remodeling of vascularization, which is necessary for lymph node expansion during inflammation. Contrary to corrosion cast studies we identified B-cell follicles exhibiting a two times denser capillary network than the deep cortical units of the T-cell zone. In addition to our observation of high endothelial venules spatially surrounding the follicles, this suggests a direct connection between morphology and B-cell homing. Our findings will deepen the understanding of functional lymph node composition and lymphocyte migration on a fundamental basis.

β Oscillations (13–30 Hz) are ubiquitous in the human motor nervous system. Yet, their origins and roles are unknown. Traditionally, β activity has been treated as a stationary signal. However, recent studies observed that cortical β occurs in “bursting events,” which are transmitted to muscles. This short-lived nature of β events makes it possible to study the main mechanism of β activity found in the muscles in relation to cortical β. Here, we assessed whether muscle β activity mainly results from cortical projections. We ran two experiments in healthy humans of both sexes (N = 15 and N = 13, respectively) to characterize β activity at the cortical and motor unit (MU) levels during isometric contractions of the tibialis anterior muscle. We found that β rhythms observed at the cortical and MU levels are indeed in bursts. These bursts appeared to be time-locked and had comparable average durations (40–80 ms) and rates (approximately three to four bursts per second). To further confirm that cortical and MU β have the same source, we used a novel operant conditioning framework to allow subjects to volitionally modulate MU β. We showed that volitional modulation of β activity at the MU level was possible with minimal subject learning and was paralleled by similar changes in cortical β activity. These results support the hypothesis that MU β mainly results from cortical projections. Moreover, they demonstrate the possibility to decode cortical β activity from MU recordings, with a potential translation to future neural interfaces that use peripheral information to identify and modulate activity in the central nervous system.SIGNIFICANCE STATEMENT We show for the first time that β activity in motor unit (MU) populations occurs in bursting events. These bursts observed in the output of the spinal cord appear to be time-locked and share similar characteristics of β activity at the cortical level, such as the duration and rate at which they occur. Moreover, when subjects were exposed to a novel operant conditioning paradigm and modulated MU β activity, cortical β activity changed in a similar way as peripheral β. These results provide evidence for a strong correspondence between cortical and peripheral β activity, demonstrating the cortical origin of peripheral β and opening the pathway for a new generation of neural interfaces.

Modern diffusion and functional magnetic resonance imaging (dMRI/fMRI) provide non-invasive high-resolution images from which multi-layered networks of whole-brain structural and functional connectivity can be derived. Unfortunately, the lack of observed correspondence between the connectivity profiles of the two modalities challenges the understanding of the relationship between the functional and structural connectome. Rather than focusing on correspondence at the level of connections we presently investigate correspondence in terms of modular organization according to shared canonical processing units. We use a stochastic block-model (SBM) as a data-driven approach for clustering high-resolution multi-layer whole-brain connectivity networks and use prediction to quantify the extent to which a given clustering accounts for the connectome within a modality. The employed SBM assumes a single underlying parcellation exists across modalities whilst permitting each modality to possess an independent connectivity structure between parcels thereby imposing concurrent functional and structural units but different structural and functional connectivity profiles. We contrast the joint processing units to their modality specific counterparts and find that even though data-driven structural and functional parcellations exhibit substantial differences, attributed to modality specific biases, the joint model is able to achieve a consensus representation that well accounts for both the functional and structural connectome providing improved representations of functional connectivity compared to using functional data alone. This implies that a representation persists in the consensus model that is shared by the individual modalities. We find additional support for this viewpoint when the anatomical correspondence between modalities is removed from the joint modeling. The resultant drop in predictive performance is in general substantial, confirming that the anatomical correspondence of processing units is indeed present between the two modalities. Our findings illustrate how multi-modal integration admits consensus representations well-characterizing each individual modality despite their biases and points to the importance of multi-layered connectomes as providing supplementary information regarding the brain's canonical processing units.

Recent advances in multi-electrode array technology have made it possible to monitor large neuronal ensembles at cellular resolution in animal models. In humans, however, current approaches restrict recordings to a few neurons per penetrating electrode or combine the signals of thousands of neurons in local field potential (LFP) recordings. Here we describe a new probe variant and set of techniques that enable simultaneous recording from over 200 well-isolated cortical single units in human participants during intraoperative neurosurgical procedures using silicon Neuropixels probes. We characterized a diversity of extracellular waveforms with eight separable single-unit classes, with differing firing rates, locations along the length of the electrode array, waveform spatial spread and modulation by LFP events such as inter-ictal discharges and burst suppression. Although some challenges remain in creating a turnkey recording system, high-density silicon arrays provide a path for studying human-specific cognitive processes and their dysfunction at unprecedented spatiotemporal resolution.

Postnatal fracture healing of atrophic long bone diaphyseal nonunions remains a challenge for orthopedic surgeons. Paucity of autologous spongiosa has potentiated the use of tissue engineered bone grafts to improve success rates of bone marrow engraftment used in plate reosteosynthesis. Herein, the development and in vitro validation of a "sandwich-type" biofabricated diaphyseal cross-sectional unit, with an outer mechanically robust bioprinted cortical bone shell, encompassing an engineered bone marrow, are reported. Channelized silk fibroin blend sponges derived from Bombyx mori and Antheraea assama help in developing compartmentalized endosteum, exhibiting specialized osteoblasts (endosteal niche) and discontinuous endothelium (vascular niche). The cellular cross-talk between these two niches triggered via integrin-mediated cell adhesion, enables in preserving quiescence state of CD34+ /CD38- hematopoietic stem cells and their recycling in the engineered marrow. The outer cortical bone strut is developed through multimaterial microextrusion bioprinting strategy. Osteogenically primed mesenchymal stem cells-laden silk fibroin-nano-hydroxyapatite bioink is bioprinted alongside paramagnetic Fe-doped bioactive glass-polycaprolactone blend thermoplastic ink, reinforcing it for mechanical stability. Pulsed magnetic field actuation positively influences the osteogenic commitment and maturation of the bioprinted constructs via mechanotransductory route. Therefore, the assembled engineered marrow and bioprinted cortical shell hold promise as potential orthobiologic substitutes toward atrophic nonunion repairs.

To observe the protective effect of electroacupuncture (EA) on neurovascular unit, neurological function in rats with cerebral ischemia (CI), so as to explore its mechanisms underlying improvement of ischemic cerebral tissue. Male SD rats, SPF grade, were randomly and equally divided into sham operation group, model group, EA group Ⅰ and EA group Ⅱ，27 rats in each group. The CI model was established by occlusion of the middle cerebral artery (MCAO). EA (2 Hz/20 Hz, 0.5 mA) was applied to "Quchi"(LI11), "Hegu"(LI4), "Zusanli"(ST36) and "Shuigou"(GV26) for rats of the EA group Ⅰ, and to "Baihui"(GV26), "Fengfu"(GV16), "Neiguan"(PC26) and "Xinshu"(BL15) for rats of the EA group Ⅱ for 20 min, once a day for 14 days. The modified neurologic severity score (mNSS) was calculated according to the state of locomotor, sensory, and reflex parameters. Transmission electron microscope (TEM) was used to observe the neuronal structure of the ischemic cerebral area. The CD34 positive cells (for microvessels) of the ischemic brain tissue were detected by using immunohistochemistry, and the expression levels of cerebral phosphatidylinositol-3 kinase (PI3K) and protein kinase B (Akt) mRNAs were detected by quantitative real-time-PCR, respectively. Along with the extension of time, the mNSS at 4 h, and 3, 7 and 14 d after CI were apparently decreased, and the number of CD34 positive cells from 3 d to 14 d after CI, and the expression of PI3K mRNA and Akt mRNA from 3 d to 7 d were significantly increased in the model，EAⅠand EA Ⅱ group （P<0.01, P<0.05）. Compared with the sham operation group, the mNSS at 4 h, and 3, 7 and 14 d, and CD34-positive number and PI3K mRNA and Akt mRNA expression levels at 3, 7 and 14 d were significantly increased in the model group (P<0.01, P>0.05). In comparison with the model group, the mNSS at 3, 7 and 14 d were obviously decreased (P<0.01), and the CD34-positive number and PI3K and Akt mRNA expression levels at 3, 7 and 14 d considerably increased in both EA group Ⅰ and Ⅱ (P<0.01, P<0.05). The therapeutic effect of EA group Ⅱ was significantly superior to that of EA group Ⅰ in lowering mNSS at 14 d, up-regulating the CD34-positive number at 7 and 14 d,and PI3K mRNA at 3, 7 and 14 d and Akt mRNA at 3 and 7 d (P<0.05, P<0.01). Results of TEM showed an irregular shape of neurons with nuclear pyknosis, non-uniform chromatin, more organelle loss, swollen mitochondrial Golgi complex and expansion of rough endoplasmic reticulum, being relatively milder in the EA group Ⅰ, particularly in the EA group Ⅱ. EA therapy can improve the neurological function in cerebral ischemia rats, which may be related to its effects in protecting the neurovascular unit and up-regulating PI3K/Akt signal pathway. The effects of EA at GV26, GV16, PC26 and BL15 are better than those of EA at LI11, LI4, ST36 and GV26.