Efficient and Flexible Method for Reducing Moderate-Size Deep Neural Networks with Condensation.

Neural networks have been extensively applied to a variety of tasks, achieving astounding results. Applying neural networks in the scientific field is an important research direction that is gaining increasing attention. In scientific applications, the scale of neural networks is generally moderate size, mainly to ensure the speed of inference during application. Additionally, comparing neural networks to traditional algorithms in scientific applications is inevitable. These applications often require rapid computations, making the reduction in neural network sizes increasingly important. Existing work has found that the powerful capabilities of neural networks are primarily due to their nonlinearity. Theoretical work has discovered that under strong nonlinearity, neurons in the same layer tend to behave similarly, a phenomenon known as condensation. Condensation offers an opportunity to reduce the scale of neural networks to a smaller subnetwork with a similar performance. In this article, we propose a condensation reduction method to verify the feasibility of this idea in practical problems, thereby validating existing theories. Our reduction method can currently be applied to both fully connected networks and convolutional networks, achieving positive results. In complex combustion acceleration tasks, we reduced the size of the neural network to 41.7% of its original scale while maintaining prediction accuracy. In the CIFAR10 image classification task, we reduced the network size to 11.5% of the original scale, still maintaining a satisfactory validation accuracy. Our method can be applied to most trained neural networks, reducing computational pressure and improving inference speed.

Asprosin in the Paraventricular Nucleus Induces Sympathetic Activation and Pressor Responses via cAMP-Dependent ROS Production.

Asprosin is a newly discovered adipokine that is involved in regulating metabolism. Sympathetic overactivity contributes to the pathogenesis of several cardiovascular diseases. The paraventricular nucleus (PVN) of the hypothalamus plays a crucial role in the regulation of sympathetic outflow and blood pressure. This study was designed to determine the roles and underlying mechanisms of asprosin in the PVN in regulating sympathetic outflow and blood pressure. Experiments were carried out in male adult SD rats under anesthesia. Renal sympathetic nerve activity (RSNA), mean arterial pressure (MAP), and heart rate (HR) were recorded, and PVN microinjections were performed bilaterally. Asprosin mRNA and protein expressions were high in the PVN. The high asprosin expression in the PVN was involved in both the parvocellular and magnocellular regions according to immunohistochemical analysis. Microinjection of asprosin into the PVN produced dose-related increases in RSNA, MAP, and HR, which were abolished by superoxide scavenger tempol, antioxidant N-acetylcysteine (NAC), and NADPH oxidase inhibitor apocynin. The asprosin promoted superoxide production and increased NADPH oxidase activity in the PVN. Furthermore, it increased the cAMP level, adenylyl cyclase (AC) activity, and protein kinase A (PKA) activity in the PVN. The roles of asprosin in increasing RSNA, MAP, and HR were prevented by pretreatment with AC inhibitor SQ22536 or PKA inhibitor H89 in the PVN. Microinjection of cAMP analog db-cAMP into the PVN played similar roles with asprosin in increasing the RSNA, MAP, and HR, but failed to further augment the effects of asprosin. Pretreatment with PVN microinjection of SQ22536 or H89 abolished the roles of asprosin in increasing superoxide production and NADPH oxidase activity in the PVN. These results indicated that asprosin in the PVN increased the sympathetic outflow, blood pressure, and heart rate via cAMP-PKA signaling-mediated NADPH oxidase activation and the subsequent superoxide production.

Neural networks of different species, brain areas and states can be characterized by the probability polling state.

Cortical networks are complex systems of a great many interconnected neurons that operate from collective dynamical states. To understand how cortical neural networks function, it is important to identify their common dynamical operating states from the probabilistic viewpoint. Probabilistic characteristics of these operating states often underlie network functions. Here, using multi-electrode data from three separate experiments, we identify and characterize a cortical operating state (the "probability polling" or "p-polling" state), common across mouse and monkey with different behaviors. If the interaction among neurons is weak, the p-polling state provides a quantitative understanding of how the high dimensional probability distribution of firing patterns can be obtained by the low-order maximum entropy formulation, effectively utilizing a low dimensional stimulus-coding structure. These results show evidence for generality of the p-polling state and in certain situations its advantage of providing a mathematical validation for the low-order maximum entropy principle as a coding strategy.

Dynamical and Coupling Structure of Pulse-Coupled Networks in Maximum Entropy Analysis.

Maximum entropy principle (MEP) analysis with few non-zero effective interactions successfully characterizes the distribution of dynamical states of pulse-coupled networks in many fields, e.g., in neuroscience. To better understand the underlying mechanism, we found a relation between the dynamical structure, i.e., effective interactions in MEP analysis, and the anatomical coupling structure of pulse-coupled networks and it helps to understand how a sparse coupling structure could lead to a sparse coding by effective interactions. This relation quantitatively displays how the dynamical structure is closely related to the anatomical coupling structure.

Implicit Regularization of Dropout.

It is important to understand how dropout, a popular regularization method, aids in achieving a good generalization solution during neural network training. In this work, we present a theoretical derivation of an implicit regularization of dropout, which is validated by a series of experiments. Additionally, we numerically study two implications of the implicit regularization, which intuitively rationalizes why dropout helps generalization. First, we find that input weights of hidden neurons tend to condense on isolated orientations trained with dropout. Condensation is a feature in the non-linear learning process, which makes the network less complex. Second, we find that the training with dropout leads to the neural network with a flatter minimum compared with standard gradient descent training, and the implicit regularization is the key to finding flat solutions. Although our theory mainly focuses on dropout used in the last hidden layer, our experiments apply to general dropout in training neural networks. This work points out a distinct characteristic of dropout compared with stochastic gradient descent and serves as an important basis for fully understanding dropout.

Ile-Pro-Pro attenuates sympathetic activity and hypertension.

Isoleucine-proline-proline (Ile-Pro-Pro, IPP) is a natural food source tripeptide that inhibits angiotensin-converting enzyme (ACE) activity. The aim of this study was to determine the central and peripheral roles of IPP in attenuating sympathetic activity, oxidative stress and hypertension. Male Sprague-Dawley rats were subjected to sham-operated surgery (Sham) or two-kidney one-clip (2K1C) surgery to induce renovascular hypertension. Renal sympathetic nerve activity and blood pressure were recorded. Bilateral microinjections of IPP to hypothalamic paraventricular nucleus (PVN) attenuated sympathetic activity (-16.1 ± 2.5%, P < 0.001) and hypertension (-8.7 ± 1.5 mmHg, P < 0.01) in 2K1C rats by inhibiting ACE activity and subsequent angiotensin II and superoxide production in the PVN. Intravenous injections of IPP also attenuated sympathetic activity (-15.1 ± 2.1%, P < 0.001) and hypertension (-16.8 ± 2.3 mmHg, P < 0.001) via inhibiting ACE activity and oxidative stress in both PVN and arteries of 2K1C rats. The duration of the effects of the intravenous IPP was longer than those of the PVN microinjection, but the sympatho-inhibitory effect of intravenous injections occurred later than that of the PVN microinjection. Intraperitoneal injection of IPP (400 pmol/day for 20 days) attenuated hypertension and vascular remodeling via inhibiting ACE activity and oxidative stress in both PVN and arteries of 2K1C rats. These results indicate that IPP attenuates hypertension and sympathetic activity by inhibiting ACE activity and oxidative stress. The sympathoinhibitory effect of peripheral IPP is mainly caused by the ACE inhibition in PVN, and the antihypertensive effect is related to the sympathoinhibition and the arterial ACE inhibition. Long-term intraperitoneal IPP therapy attenuates hypertension, oxidative stress and vascular remodeling.

Asprosin contributes to vascular remodeling in hypertensive rats via superoxide signaling.

OBJECTIVE: Proliferation and migration of vascular smooth muscle cells (VSMCs) contribute to vascular remodeling. Asprosin, a newly discovered protein hormone, is involved in metabolic diseases. Little is known about the roles of asprosin in cardiovascular diseases. This study focused on the role and mechanism of asprosin on VSMC proliferation and migration, and vascular remodeling in a rat model of hypertension. METHODS AND RESULTS: VSMCs were obtained from the aortic media of 8-week-old male Wistar-Kyoto rats (WKY) and spontaneously hypertensive rats (SHR). Asprosin was upregulated in the VSMCs of SHR. For in vitro studies, asprosin promoted VSMC proliferation and migration of WKY and SHR, and increased Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) activity, NOX1/2/4 protein expressions and superoxide production. Knockdown of asprosin inhibited the proliferation, migration, NOX activity, NOX1/2 expressions and superoxide production in the VSMCs of SHR. The roles of asprosin in promoting VSMC proliferation and migration were not affected by hydrogen peroxide scavenger, but attenuated by superoxide scavenger, selective NOX1 or NOX2 inhibitor. Toll-like receptor 4 (TLR4) was upregulated in SHR, TLR4 knockdown inhibited asprosin overexpression-induced proliferation, migration and oxidative stress in VSMCs of WKY and SHR. Asprosin was upregulated in arteries of SHR, and knockdown of asprosin in vivo not only attenuated oxidative stress and vascular remodeling in aorta and mesentery artery, but also caused a subsequent persistent antihypertensive effect in SHR. CONCLUSIONS: Asprosin promotes VSMC proliferation and migration via NOX-mediated superoxide production. Inhibition of endogenous asprosin expression attenuates VSMC proliferation and migration, and vascular remodeling of SHR.

Data-informed deep optimization.

Motivated by the impressive success of deep learning in a wide range of scientific and industrial applications, we explore in this work the application of deep learning into a specific class of optimization problems lacking explicit formulas for both objective function and constraints. Such optimization problems exist in many design problems, e.g., rotor profile design, in which objective and constraint values are available only through experiment or simulation. They are especially challenging when design parameters are high-dimensional due to the curse of dimensionality. In this work, we propose a data-informed deep optimization (DiDo) approach emphasizing on the adaptive fitting of the the feasible region as follows. First, we propose a deep neural network (DNN) based adaptive fitting approach to learn an accurate DNN classifier of the feasible region. Second, we use the DNN classifier to efficiently sample feasible points and train a DNN surrogate of the objective function. Finally, we find optimal points of the DNN surrogate optimization problem by gradient descent. To demonstrate the effectiveness of our DiDo approach, we consider a practical design case in industry, in which our approach yields good solutions using limited size of training data. We further use a 100-dimension toy example to show the effectiveness of our approach for higher dimensional problems. Our results indicate that, by properly dealing with the difficulty in fitting the feasible region, a DNN-based method like our DiDo approach is flexible and promising for solving high-dimensional design problems with implicit objective and constraints.

Chemerin-9 in paraventricular nucleus increases sympathetic outflow and blood pressure via glutamate receptor-mediated ROS generation.

Chemerin is an adipokine involved in regulating energy homeostasis and reproductive function. Excessive sympathetic activity contributes to hypertension, chronic heart failure and chronic renal disease. Hypothalamic paraventricular nucleus (PVN) is crucial in regulating sympathetic activity and blood pressure. The present study is designed to investigate the roles of chemerin in the PVN in regulating sympathetic activity and blood pressure and underlying mechanisms. Microinjections were performed in the bilateral PVN in male adult rats under anesthesia. Renal sympathetic nerve activity (RSNA) and mean arterial pressure (MAP) were continuously recorded. The PVN microinjection of chemerin-9, an active fragment of chemerin, increased RSNA and MAP, which were abolished by chemokine-like receptor 1 (CMKLR1) antagonist α-NETA, a superoxide scavenger tempol, antioxidant N-acetylcysteine (NAC), NADPH oxidase inhibitor diphenyleneiodonium chloride (DPI) and apocynin. Immunofluorescence analyses showed that N-methyl-D-aspartate (NMDA) receptors existed in most of cells of the PVN, and some of them co-existed with chemerin. The effects of chemerin-9 on RSNA and MAP were prevented by glutamate-binding site antagonist L-phenylalanine, NMDA receptor antagonist MK-801, and calcium channel blocker verapamil or nifedipine, but only attenuated by non-NMDA receptor antagonist CNQX. Moreover, chemerin-9 increased NADPH oxidase activity and superoxide production, which were prevented by α-NETA, MK-801, or verapamil. These results indicate that chemerin-9 in the PVN increases sympathetic activity and blood pressure via CMKLR1-dependent calcium influx, and glutamate receptor-mediated NADPH oxidase activation and subsequent superoxide production.

[Clinical application of neuronavigation in transsphenoidal microsurgery of pituitary adenomas].

OBJECTIVES: To summarize the experiences in clinical application of neuronavigation in transsphenoidal microsurgery of specific pituitary adenomas, and to discuss its indications. METHODS: From January 2006 to December 2010, 138 cases of transsphenoidal microsurgery for specific pituitary adenomas under neuronavigation were reviewed. The indications for neuronavigation in transsphenoidal microsurgery includes: recurrent or regrowth of residual pituitary adenomas after former transsphenoidal surgery in 36 cases, invasive pituitary adenomas in 45 cases, extremely laterally or deeply situated microadenomas in 45 cases, poor pneumatization of the sphenoid in 4 cases, skull base anomalies due to osteodysplasia fibrosa in 3 cases, narrow space between bilateral internal carotid arteries in 4 cases, distortion of nasal septum in 1 case. RESULTS: In the recurrence group, 12 were totally removed, 9 subtotally removed; postoperative complications included hematoma within the tumor cavity in 2 cases, cerebrospinal fluid (CSF) leakage in 4 cases among which 3 developed intracranial infection and 2 communicating hydrocephalus, oculomotor paralysis in 1 case and hypopituitarism in 3 cases; 9 were cured and 8 remission. In the invasive group, 5 were totally removed, 27 subtotally removed; postoperative complications included hematoma within the tumor cavity in 1 case, CSF leakage and intracranial infection in 1 case; 2 were cured and 22 remission. None of the 30 invasive hormone-secreting adenomas were cured or remission. The 45 cases of hormone-secreting microadenomas were all totally removed, among which 38 were cured. Among the poor sphenoid pneumatization group, total and subtotal tumor removal were achieved in 2 cases respectively with only one cured. In the skull base anomaly group, 2 were totally removed and 1 subtotally removed, with only one cured. For the cases with narrow space between bilateral internal carotid arteries and distortion of nasal septum, all were totally removed and cured. CONCLUSIONS: Transsphenoidal microsurgery under neuronavigation can be applied for pituitary adenomas in above specific indications. It is an accurate, safe and effective approach for specific pituitary adenomas, which can not only expand the indication of transsphenoidal microsurgery for pituitary adenomas, but also reduce the harmful exposure of X-rays for the operating staff.

[Management of pulmonary thromboembolism in post-operative neurosurgical patients].

OBJECTIVE: To summarize the diagnosis and treatment of pulmonary thromboembolism (PTE) in post-operative neurosurgical patients. METHODS: We retrospectively analyzed the clinical data of 7 patients who experienced pulmonary thromboembolism after neurosurgical operations in our department from October 2009 to March 2010. RESULTS: Of these 7 patients, 6 were confirmed with computed tomographic pulmonary angiography (CTPA) and 1 was diagnosed according to the clinical manifestations and other diagnostic examinations. All the patients were treated initially with low-dose heparin or low-molecular-weight heparin and then with warfarin. Two patients were implanted with permanent inferior vena cava filters before anticoagulation. One received anticoagulant therapy and died of respiratory failure due to pulmonary embolism on the fourth post-operative day. Six patients were discharged after significant improvement. CONCLUSIONS: Many risk factors may cause PTE peri-operatively. Post-operative CTPA may be indicated. Anticoagulation and other management strategies may be applied to improve the outcome.

Maximum entropy principle analysis in network systems with short-time recordings.

In many realistic systems, maximum entropy principle (MEP) analysis provides an effective characterization of the probability distribution of network states. However, to implement the MEP analysis, a sufficiently long-time data recording in general is often required, e.g., hours of spiking recordings of neurons in neuronal networks. The issue of whether the MEP analysis can be successfully applied to network systems with data from short-time recordings has yet to be fully addressed. In this work, we investigate relationships underlying the probability distributions, moments, and effective interactions in the MEP analysis and then show that, with short-time recordings of network dynamics, the MEP analysis can be applied to reconstructing probability distributions of network states that is much more accurate than the one directly measured from the short-time recording. Using spike trains obtained from both Hodgkin-Huxley neuronal networks and electrophysiological experiments, we verify our results and demonstrate that MEP analysis provides a tool to investigate the neuronal population coding properties for short-time recordings.

Translation and Initial Validation of the Chinese Version of the Action Research Arm Test in People with Stroke.

PURPOSE: This study aimed to translate the English version of the Action Research Arm Test (ARAT) into Chinese and to evaluate the initial validation of the Chinese version (C-ARAT) in patients with a first stroke. METHODS: An expert group translated the original ARAT from English into Chinese using a forward-backward procedure. Forty-four patients (36 men and 8 women) aged 22-80 years with a first stroke were enrolled in this study. The participants were evaluated using 3 stroke-specific outcome measures: C-ARAT, the upper extremity section of the Fugl-Meyer assessment (UE-FMA), and the Wolf Motor Function Test (WMFT). Internal consistency was analysed using Cronbach's α coefficients and item-scale correlations. Concurrent validity was determined using Spearman's rank correlation coefficients. Floor and ceiling effects were considered to be present when more than 20% of patients fell outside the preliminarily set lower or upper boundary, respectively. RESULTS: The C-ARAT items yielded excellent internal consistency, with a Cronbach's α of 0.98 (p < 0.001) and item-total correlations ranging from 0.727 to 0.948 (p < 0.001). The C-ARAT exhibited good-to-excellent correlations with the UE-FMA and WMFT functional ability (WMFT-FA) scores, with respective ρ values of 0.824 and 0.852 (p < 0.001), and an excellent negative correlation with the WMFT performance time (WMFT-time), with a ρ value of -0.940 (p < 0.001). The C-ARAT subscales generally exhibited good-to-excellent correlations with stroke-specific assessments, with ρ values ranging from 0.773 to 0.927 (p < 0.001). However, the gross subscale exhibited moderate-to-good correlations with the UE-FMA and WMFT-FA scores, with respective ρ values of 0.665 and 0.720 (p < 0.001). No significant floor effect was observed, and a significant ceiling effect was observed only on the WMFT-time. CONCLUSIONS: The C-ARAT demonstrated excellent internal consistency and good-to-excellent concurrent validity. This test could be used to evaluate upper extremity function in stroke patients without cognitive impairment.

Responsiveness and Predictive Ability of the Chinese Version of the Action Research Arm Test in People with Cerebral Infarction.

PURPOSE: To detect the responsiveness and predictive ability of the Chinese version Action Research Arm Test (C-ARAT) in participants within the first 3 months after cerebral infarction. METHODS: Ninety-seven individuals (75 men, mean age 59.87 ± 10.94 years) with a first cerebral infarction were enrolled in this study. The participants were evaluated by two outcome measures: C-ARAT and the Barthel Activities of Daily Living Index (BI) at five time points: 0D, 3W, 3M, 6M and 1Y after enrolment. The standardised response mean (SRM) and the Wilcoxon signed rank test were used to analyse responsiveness. Predictive validity was determined by using Spearman's rank correlation coefficients. The predicted performance of C-ARAT on activities of daily living (ADLs) was measured by linear regression model. Floor and ceiling effects were estimated by counting the proportion of subjects falling outside the 5% lower or upper boundary, respectively. RESULTS: The C-ARAT showed moderate to large responsiveness in detecting changes over time (SRM = 0.58-0.84). The C-ARAT subscales showed small to large responsiveness (SRM = 0.44-0.90). The C-ARAT at 0D showed moderate to good correlation with the BI scores at 3W, 3M and 6M (ρ = 0.561-0.624, p < 0.001), and exhibited fair correlation with the BI score 1Y after enrolment (ρ = 0.384, p < 0.05). C-ARAT was a good predictor (adjusted R 2 = 0.185-0.249) of BI within 3M follow-up. The C-ARAT total score showed a notable floor effect at 0D and 3W and a notable ceiling effect at 3M, 6M and 1Y. CONCLUSION: The results of this study support the use of the C-ARAT as a measurement of upper extremity function in individuals with a first cerebral infarction.

[Transsphenoidal microsurgical results of non-invasive prolactinomas].

OBJECTIVE: To analyze the transsphenoidal microsurgical results of non-invasive prolactinomas, in order to provide reference for their treatment choice. METHODS: To review the transsphenoidal microsurgical results of 234 non-invasive prolactinomas treated in our department in recent 10 years, and to analyze the prognostic factors. There were 18 males and 216 females, aged between 13 and 58 years, averaged (31.1 +/- 8.5) years. The course ranged from half a month to 20 years, averaged (47.3 +/- 44.9) months. The preoperative serum PRL level ranged between 41 and 8406 ng/ml, averaged (400.5 +/- 888.0) ng/ml, with a median of 164.1 ng/ml. The primary symptoms were amenorrhea, galactorrhea and/or infertility in 211 cases. The tumor size was small (< 1 cm) in 100, large (> or = 1 cm) in 116 and giant (> or = 3 cm) in 18 cases. All the patients received transsphenoidal microsurgery and were followed-up for 12 to 132 months, averaged (43.8 +/- 35.0) months. RESULTS: There was no mortality. One hundred and twenty-seven (54.3%) cases had transient postoperative imbalance of water and electrolytes. One hundred and eighty-eight cases (80.3%) were cured, 12 (5.1%) experienced clinical remission, 20 (8.5%) were improved, and 14 (6.0%) were ineffective. The male patient, the giant prolactinomas and those with higher preoperative serum PRL level had a relative poor postoperative prognosis. While the other factors had no influence on prognosis, including the course, preoperative bromocriptine intake, tumor texture, tumor apoplexy and intraoperative descending extent of the diaphragm of sella. The overall operative expense for transsphenoidal microsurgery ranged from 8323.8 to 22898.5 yuan, averaged (12912.0 +/- 2361.2) yuan. CONCLUSIONS: Transsphenoidal microsurgery may be chosen as the primary therapy for non-invasive prolactinomas, with the purposes of therapeutical efficacy, facilitating the patients, re-establishing the patients' self-confidence and reducing the overall expense.

The Crucial Changes of Sit-to-Stand Phases in Subacute Stroke Survivors Identified by Movement Decomposition Analysis.

OBJECTIVE: The aim of this study was to detect the key changes during sit-to-stand (STS) movement cycle in hemiparetic stroke survivors using a five-phase kinematic and kinetic analysis. METHODS: Twenty-five subacute stroke survivors and 17 age-matched healthy adults participated in this study. The kinematic and kinetic parameters during STS cycle were measured using three-dimensional motion analysis system with force plates. The five standard phases of STS cycle were identified by six timing transitional points. RESULTS: Longer total time as well as larger changes were observed at the initial phase (phase I, 0.76 ± 0.62 VS 0.43 ± 0.09 s; p = 0.049) and at the end of hip and knee extension phase (phase IV, 0.93 ± 0.41 VS 0.63 ± 0.14 s; p = 0.008) in the stroke group than healthy group. Time to maximal knee joint moment was significantly delayed in the stroke group than in the control group (1.14 ± 1.06 VS 0.60 ± 0.09 s, p < 0.001). The maximal hip flexion was lower during the rising phase from seated position on the affected side in the stroke group than in the control group (84.22° ± 11.64°VS 94.11° ± 9.40°; p = 0.022). Ground reaction force was lower (4.61 ± 0.73 VS 5.85 ± 0.53 N, p < 0.001) in the affected side of the stroke group than in the control group. In addition, knee joint flexion was significantly lower at just-standing phase (T4) and at end point (T5) (5.12° ± 5.25° VS 8.21° ± 7.28°, p = 0.039; 0.03° ± 5.41° VS 3.07° ± 6.71°, p = 0.042) on the affected side than the unaffected side. Crucial decrease of knee joint moment at abrupt transitory (T2) and the maximal moment was also observed on the affected side in comparison with the unaffected side (0.39 ± 0.29 VS 0.77 ± 0.25 Nm/kg, p < 0.001; 0.42 ± 0.38 VS 0.82 ± 0.24 Nm/kg, p < 0.001). CONCLUSION: The findings of movement decomposition analysis provided useful information to clinical evaluation of STS performance, and may potentially contribute to the design of rehabilitation intervention program for optimum functional recovery of STS after stroke.

Effect of Virtual Reality on Postural and Balance Control in Patients with Stroke: A Systematic Literature Review.

Objective. To critically evaluate the studies that were conducted over the past 10 years and to assess the impact of virtual reality on static and dynamic balance control in the stroke population. Method. A systematic review of randomized controlled trials published between January 2006 and December 2015 was conducted. Databases searched were PubMed, Scopus, and Web of Science. Studies must have involved adult patients with stroke during acute, subacute, or chronic phase. All included studies must have assessed the impact of virtual reality programme on either static or dynamic balance ability and compared it with a control group. The Physiotherapy Evidence Database (PEDro) scale was used to assess the methodological quality of the included studies. Results. Nine studies were included in this systematic review. The PEDro scores ranged from 4 to 9 points. All studies, except one, showed significant improvement in static or dynamic balance outcomes group. Conclusions. This review provided moderate evidence to support the fact that virtual reality training is an effective adjunct to standard rehabilitation programme to improve balance for patients with chronic stroke. The effect of VR training in balance recovery is less clear in patients with acute or subacute stroke. Further research is required to investigate the optimum training intensity and frequency to achieve the desired outcome.

Reconstruction of the sellar floor following transsphenoidal surgery using gelatin foam and fibrin glue.

OBJECTIVE: To introduce a new principle of sellar reconstruction and to evaluate the effectiveness of absorbable gelatin foam and fibrin glue for sellar reconstruction. METHODS: A total of 176 consecutive patients who underwent surgery for pituitary adenomas, cysts, chordomas, or subdiaphragmatic craniopharyngiomas in the sella turcica between January 2001 and April 2003 at Peking Union Medical College Hospital were enrolled. Different techniques of sellar closure and indications for each specific condition were retrospectively reviewed. RESULTS: Seventy-seven (43.7%) patients developed a visible cerebrospinal fluid (CSF) leakage during surgery. Intra-operative CSF leakage were repaired simply with gelatin foam and fibrin glue in 62 (35.2%) patients, and with autologous fat graft and sellar floor reconstruction in 15 (8.5%) patients. Postoperative CSF rhinorrhea occurred only in 1 case. There were no visual deterioration, allergic rhinitis, meningitis, pneumocranium, granulomas, or other complications associated with the reconstruction procedure. CONCLUSION: The procedure of using gelatin foam and fibrin glue and principle of cranial base reconstruction is safe and effective in preventing postoperative complications following transsphenoidal surgery.

Effect of preoperative use of long-acting octreotide on growth hormone secreting pituitary adenoma and transsphenoidal surgery.

OBJECTIVE: To investigate whether somatostatin analog octreotide long acting release (LAR) shrinks growth hormone (GH) secreting adenomas, and improves the results of subsequent transsphenoidal surgery. METHODS: Seventeen previously untreated active acromegalic patients with pituitary adenomas were treated with LAR (30 mg intramuscular injection every 28 days) for 3 months prior to transsphenoidal surgery. Clinical reaction, mean GH secretion, and tumor volume were measured under basal conditions and after LAR treatment. RESULTS: Presurgical treatment improved acromegaly symptoms and induced a significant reduction of GH under the 5 ng/mL limit in microadenoma (P < 0.05), while only 18.2% (2/11) in macroadenoma. Meanwhile, tumor shrinkage occurred in 58.8% (10/17) patients, with 1 case in the microadenoma group. All marked shrinkage (> 25%) occurred in the macroadenoma group. Statistical analysis showed tumor shrinkage caused by LAR was greater in macroadenoma group than that in microadenoma group (P < 0.05). During operation, adenoma was soft in 15 cases, with the exception of 2 cases in which the soft tumor was divided by fibrous septa, but all tumor removal was smooth. CONCLUSIONS: A short term administration of preoperative LAR may induce a significant decrease in GH-secretion level and adenoma volume. Presurgical use of octreotide LAR improves surgical results especially in macroadenomas.

[Gene therapy of rat prolactinomas mediated by adenoviral vectors with rat tyrosine hydroxylase gene].

OBJECTIVE: To investigate the potential of gene therapy of rat prolactinomas mediated by adenoviral vectors with a gene encoding rat tyrosine hydroxylase. METHODS: Recombinant replication-deficient adenovirus named Ad-GFP-TH with rat TH-cDNA and control adenovirus named Ad-GFP were constructed by homologous recombination in bacterial cells. The rat pituitary prolactinoma cell line MMQ are chosen as the target cells to study the effect of gene therapy on their growth and prolactin secretion mediated by Ad-GFP-TH. RESULTS: Recombinant Ad-GFP-TH and Ad-GFP were successfully reconstructed. Transfection of MMQ cells with Ad-GFP-TH not only restrained their growth but also decreased their PRL secretion. CONCLUSION: Gene therapy may serve for a potential treatment for prolactinomas, especially invasive prolactinomas.