Neural network analysis of preoperative variables and outcome in epilepsy surgery

1999 ◽  
Vol 90 (6) ◽  
pp. 998-1004 ◽  
Author(s):  
Jeffrey E. Arle ◽  
Kenneth Perrine ◽  
Orrin Devinsky ◽  
Werner K. Doyle
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Epilepsy Surgery ◽  
Multivariate Techniques ◽  
Chi Square ◽  
Content Type ◽  
Individual Variables ◽  
Postoperative Seizure ◽  
Chi Square Analysis ◽  
Fine Print

Object. Because appropriate patient selection is essential for achieving successful outcomes after epilepsy surgery, the need for more robust methods of predicting postoperative seizure control has been created. Standard multivariate techniques have been only 75 to 80% accurate in this regard. Recent use of artificial intelligence techniques, including neural networks, for analyzing multivariate clinical data has been successful in predicting medical outcome.Methods. The authors applied neural network techniques to 80 consecutive patients undergoing epilepsy surgery in whom data on demographic, seizure, operative, and clinical variables to predict postoperative seizures were collected.Neural networks could be used to predict postoperative seizures in up to 98% of cases. Student's t-tests or chi-square analysis performed on individual variables revealed that only the preoperative medication index was significantly different (p = 0.02) between the two outcome groups. Six different combinations of input variables were used to train the networks. Neural network accuracies differed in their ability to predict seizures: using all data (96%); all data minus electroencephalography concordance and operative side (93%); all data except intra- or postoperative variables such as tissue pathological category (98%); all data excluding pathological category, intelligence quotient (IQ) data, and Wada results (84%); only demographics and tissue pathological category (65%); and only IQ data (63%).Conclusions. Analysis of the results reveals that several networks that are trained with the usual accepted variables characterizing the typical evaluation of epilepsy patients can predict postoperative seizures with greater than 95% accuracy.

scholarly journals Neural network analysis of preoperative variables and outcome in epilepsy surgery

1999 ◽  
Vol 6 (2) ◽  
pp. E6
Author(s):  
Jeffrey E. Arle ◽  
Kenneth Perrine ◽  
Orrin Devinsky ◽  
Werner K. Doyle
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Epilepsy Surgery ◽  
Multivariate Techniques ◽  
Chi Square ◽  
Individual Variables ◽  
Input Variables ◽  
Student’S T ◽  
Postoperative Seizure ◽  
Chi Square Analysis

Because appropriate patient selection is essential for achieving successful outcomes after epilepsy surgery, the need for more robust methods of predicting postoperative seizure control has been created. Standard multivariate techniques have been only 75 to 80% accurate in this regard. Recent use of artificial intelligence techniques, including neural networks, for analyzing multivariate clinical data, has been successful in predicting medical outcome. The authors applied neural network techniques to 80 consecutive patients undergoing epilepsy surgery in whom demographic, seizure, operative, and clinical variables to predict postoperative seizures data were obtained. Neural networks were able to predict postoperative seizures in up to 98% of cases. Student's t tests or chi-square analysis performed on individual variables revealed that only the preoperative medication index was significantly different (p = 0.02) between the two outcome groups. Six different combinations of input variables were used to train the networks. Neural network accuracies differed in their ability to predict seizures using all data (96%); all data minus electroencephalography concordance and operative side (93%); all data except intra- or postoperative variables such as tissue pathology (98%); all data excluding pathology, intelligence quotient (IQ) data, and Wada results (84%); only using demographics and tissue pathology (65%); and only using IQ data (63%). Analysis of the results reveals that several networks that are trained with the usual accepted variables characterizing the typical evaluation of epilepsy patients can predict postoperative seizures with greater than 95% accuracy.

Head injuries due to falls caused by seizures: a group at high risk for traumatic intracranial hematomas

1997 ◽  
Vol 86 (3) ◽  
pp. 433-437 ◽  
Author(s):  
Thomas J. Zwimpfer ◽  
Jennifer Brown ◽  
Irene Sullivan ◽  
Richard J. Moulton
Keyword(s):  
Head Injury ◽  
Alcohol Intoxication ◽  
Head Injuries ◽  
Chi Square ◽  
Focal Neurological Deficit ◽  
Content Type ◽  
Intracranial Hematomas ◽  
Chi Square Analysis ◽  
Mass Lesions ◽  
Fine Print

✓ This prospective review of adult patients with head injuries examines the incidence of head injuries due to falls caused by seizures, the incidence and severity of intracranial hematomas, and the morbidity and mortality rates in this patient population. A head injury was attributed to a fall caused by a seizure if the seizure was witnessed to have caused the fall, or the patient had a known seizure history, appeared postictal or was found convulsing after the fall, and no other cause for the fall was evident. A total of 1760 adult head-injured patients were consecutively admitted to the authors' service between 1986 and 1993. Five hundred eighty-two head injuries (33.1%) were due to falls and 22 (3.8%) of these were caused by seizures. Based on the prevalence rates for epilepsy in the general population of 0.5 to 2%, these results indicate that epileptics are several times more likely to suffer a head injury due to a fall. Mass lesions were found in 20 (90.9%) of these 22 patients and the remaining two patients suffered mild diffuse head injuries. There was a high incidence of extraaxial mass lesions: 17 (85%) of the 20 intracranial hematomas were either epidural (five cases) or acute subdural (12 cases) hematomas. Eighteen (81.8%) of the 22 patients required evacuation of a hematoma. Both the incidence of intracranial hematomas (90.9% vs. 39.8%; p < 0.001, chi-square analysis) and the rate of hematoma evacuation (81.8% vs. 32.3%; p < 0.001) was significantly greater in patients injured in falls due to seizures (22 cases) than in the group injured in falls from all other causes (560 cases). The higher incidence of hematomas and the need for evacuation were not explained by differences in age, severity of head injury, or incidence of alcohol intoxication. Despite the greater incidence of mass lesions and the need for operative treatment in patients injured because of seizures, their mortality rate was similar to that of patients injured in falls from other causes. On the basis of their review of patients admitted to a neurosurgical center with complaints of head injury, the authors conclude that patients with head injuries due to a fall caused by a seizure should undergo computerized tomography scanning early in their management. Until a mass lesion has been excluded, any decrease in level of consciousness or focal neurological deficit should not be attributed to the seizure itself.

Prediction of posterior fossa tumor type in children by means of magnetic resonance image properties, spectroscopy, and neural networks

1997 ◽  
Vol 86 (5) ◽  
pp. 755-761 ◽  
Author(s):  
Jeffrey E. Arle ◽  
Craig Morriss ◽  
Zhiyue J. Wang ◽  
Robert A. Zimmerman ◽  
Peter G. Phillips ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Magnetic Resonance ◽  
Mr Imaging ◽  
Diagnostic Accuracy ◽  
Posterior Fossa ◽  
Tumor Size ◽  
Tumor Type ◽  
Content Type ◽  
Fine Print

✓ Recent studies have explored characteristics of brain tumors by means of magnetic resonance spectroscopy (MRS) to increase diagnostic accuracy and improve understanding of tumor biology. In this study, a computer-based neural network was developed to combine MRS data (ratios of N-acetyl-aspartate, choline, and creatine) with 10 characteristics of tumor tissue obtained from magnetic resonance (MR) studies, as well as tumor size and the patient's age and sex, in hopes of further improving diagnostic accuracy. Data were obtained in 33 children presenting with posterior fossa tumors. The cases were analyzed by a neuroradiologist, who then predicted the tumor type from among three categories (primitive neuroectodermal tumor, astrocytoma, or ependymoma/other) based only on the data obtained via MR imaging. These predictions were compared with those made by neural networks that had analyzed different combinations of the data. The neuroradiologist correctly predicted the tumor type in 73% of the cases, whereas four neural networks using different datasets as inputs were 58 to 95% correct. The neural network that used only the three spectroscopy ratios had the least predictive ability. With the addition of data including MR imaging characteristics, age, sex, and tumor size, the network's accuracy improved to 72%, consistent with the predictions of the neuroradiologist who was using the same information. Use of only the analog data (leaving out information obtained from MR imaging), resulted in 88% accuracy. A network that used all of the data was able to identify 95% of the tumors correctly. It is concluded that a neural network provided with imaging data, spectroscopic data, and a limited amount of clinical information can predict pediatric posterior fossa tumor type with remarkable accuracy.

Online versus offline: preferred retail choice for premium brand purchase

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Rituparna Basu ◽  
Neena Sondhi
Keyword(s):  
Predictive Model ◽  
Logit Analysis ◽  
Chi Square ◽  
Content Type ◽  
Sales Personnel ◽  
Study Results ◽  
Shopping Experience ◽  
Brick And Mortar ◽  
Chi Square Analysis ◽  
Retail Format

PurposeThis exploratory study aims to examine the prevalent triggers that motivate a premium brand purchase in an online vs offline retail format.Design/methodology/approachA binary logit analysis is used to build a predictive model to assess the likelihood of the premium brand consumer seeking an online or an offline platform. Demographic and usage-based profile of the two set of consumers is established through a chi-square analysis.FindingsThree hundred and forty six urban consumers of premium branded apparels residing in two Indian Metros were studied. A predictive model with 89.6% accuracy was validated for distinguishing premium brand buyers who shop at brick-and-mortar store or online platforms. Quality and finish were factors sought by the online buyer, whereas autotelic need, pleasurable shopping experience and social approval were important triggers for an in-store purchase.Research limitations/implicationsThe study posits divergent demographics and motivational drivers that led to an online vs offline purchase. Though interesting and directional, the study results need to be examined across geographies and categories for establishing the generalizability of the findings.Practical implicationsThe study findings indicate that premium brand manufacturers can devise an omni-channel strategy that is largely tilted toward the online platform, as the quality conscious and brand aware consumer is confident and thus open to an online purchase. The implication for the physical outlet on the other hand is to ensure exclusive store atmospherics and knowledgeable but non-intrusive sales personnel.Originality/valueThe study is unique as it successfully builds a predictive model to forecast online vs offline purchase decisions among urban millennials.

Using an artificial neural network for estimating sustainable development goals index

2020 ◽  
Vol 31 (4) ◽  
pp. 1023-1037 ◽  
Author(s):  
Seyed-Hadi Mirghaderi
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Network ◽  
Sustainable Development ◽  
Artificial Neural Networks ◽  
Sustainable Development Goals ◽  
Content Type ◽  
The Sustainable Development ◽  
Development Goals ◽  
Artificial Neural

PurposeThis paper aims to develop a simple model for estimating sustainable development goals index using the capabilities of artificial neural networks.Design/methodology/approachSustainable development has three pillars, including social, economic and environmental pillars. Three clusters corresponding to the three pillars were created by extracting sub-indices of three 2018 global reports and performing cluster analysis on the correlation matrix of sub-indices. By setting the sustainable development goals index as the target variable and selecting one indicator from each cluster as input variables, 20 artificial neural networks were run 30 times.FindingsArtificial neural networks with seven nodes in one hidden layer can estimate sustainable development goals index by using just three inputs, including ecosystem vitality, human capital and gross national income per capita. There is an excellent similarity (>95%) between the results of the artificial neural network and the sustainable development goals index.Practical implicationsInstead of calculating 232 indicators for determining the value of sustainable development goals index, it is possible to use only three sub-indices, but missing 5% of precision, by using the proposed artificial neural network model.Originality/valueThe study provides additional information on the estimating of sustainable development and proposes a new simple method for estimating the sustainable development goals index. It just uses three sub-indices, which can be retrieved from three global reports.

Laser-sealed arteriotomy: a reliable aneurysm model

1987 ◽  
Vol 67 (2) ◽  
pp. 284-287 ◽  
Author(s):  
Matthew R. Quigley ◽  
Kenneth Heiferman ◽  
Hau C. Kwaan ◽  
Danko Vidovich ◽  
Peter Nora ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Carotid Artery ◽  
Systolic Blood Pressure ◽  
Systemic Hypertension ◽  
Chi Square ◽  
Content Type ◽  
Small Arteries ◽  
Rigid Supports ◽  
Aneurysm Model ◽  
Fine Print

✓ Laser-assisted vascular anastomosis (LAVA) is associated with a significant aneurysm problem when it is applied to small arteries. The etiology of this phenomenon was investigated by creating arteriotomies of different lengths and orientation in the rat carotid artery and sealing them with the milliwatt CO2 laser. It was found that increasing the arteriotomy length from 0.5 to 1.0 mm significantly raised aneurysm occurrence (4/17 vs. 25/28, chi-square: p < 0.001) regardless of orientation. Systemic hypertension (systolic blood pressure ≥ 170 mm Hg) also significantly affected the aneurysm rate among the 0.5-mm arteriotomy group, raising aneurysm occurrence from 23.5% (4/17) to 100% (14/14) (p < 0.001). Assuming that the stay-sutures used for LAVA's act as rigid supports, the rate of aneurysm occurrence must be related to the distance between sutures. This phenomenon has been exploited to create a reliable aneurysm model.

Risk factors for diabetes mellitus and hypertension among artisans in Ogun state, Nigeria

2019 ◽  
Vol 50 (4) ◽  
pp. 695-710 ◽  
Author(s):  
Cathrine A. Oladoyinbo ◽  
Adenike Mercy Abiodun ◽  
Mariam Oluwatoyin Oyalowo ◽  
Irene Obaji ◽  
Abisola Margaret Oyelere ◽  
...  
Keyword(s):  
Physical Activity ◽  
Risk Factors ◽  
Abdominal Obesity ◽  
Dietary Habits ◽  
Activity Level ◽  
Chi Square ◽  
Content Type ◽  
Factors Associated ◽  
Ogun State ◽  
Chi Square Analysis

Purpose This study was designed to assess the risk factors associated with hypertension (HTN) and diabetes among artisans in Ogun State, Nigeria. Evidences suggest increasing prevalence, incidences and morbidity of diabetes and HTN in Nigeria. However, the purpose of this study is to plan and prioritize effective intervention programs, there is need to provide data on the prevalence and risk factors for HTN and diabetes among local groups. Design/methodology/approach In total 300 apparently healthy artisans who have never been diagnosed of diabetes or HTN were randomly selected from five communities. A structured questionnaire was used in obtaining information on the personal characteristics of the respondents. An adapted dietary habit and lifestyle questionnaire were used to assess the dietary habits and lifestyle of the respondents. The WHO global activity questionnaire was adapted and used to gather information on the physical activity level of the respondents. Random blood glucose, blood pressure and anthropometric measurements were assessed using standard instruments. Chi-square (χ2), correlations and multinomial logistic regression analysis were performed to identify significant determinants of diabetes and HTN. Findings Mean age was 34.8 ± 9.9 and prevalence of diabetes and pre-diabetes were 1 and 4.7 per cent, respectively, while HTN and pre-HTN were 48.0 and 30.3 per cent, respectively. About half (55.7 per cent) of the respondents skip at least a meal daily and 31 per cent snack daily. Most (61.4 per cent) consume alcohol and 65.7 per cent engage in high physical activity. Abdominal obesity was significantly higher among women (p = 0.004). Using the chi-square analysis, age, abdominal obesity and educational status were factors found to be significantly associated with diabetes (p = 0.002; p = 0.007; p = 0.004) while age, gender, abdominal obesity and alcohol consumption had significant association with HTN. Although not statistically significant, respondents were 0.8, 1.0 and 1.1 times more likely to be diabetic with increasing body mass index, waist circumference (WC) and age (odd ratio (OR) = 0.78; confidence intervals (CI): 0.51-1.18; OR = 1.04; CI: 0.89-1.21; OR = 1.06; CI: 0.96-1.18, respectively). Abdominal obesity was significantly associated with HTN (OR = 1.08; CI: 1.03-1.13; p = 0.001). Also, older respondents were 1.1 times more likely of becoming hypertensive (OR = 1.07; CI: 1.02-1.11; p = 0.003). Increased risk of diabetes and HTN was found among respondents with increasing age and WC. Research limitations/implications This study was cross-sectional in design; it cannot be used to establish a cause-effect relationship between diabetes, HTN and the observed variables (anthropometric characteristics, dietary habits and lifestyle risk factors). Because of the few numbers (1 per cent) of respondents identified to be diabetic several important risk factors could not be included in the model. Practical implications An understanding of the risk factors associated with diabetes and HTN among sub-groups in the population will help to plan effective interventions targeted at specific groups. Originality/value The findings of this study show the associated risk factors for diabetes and HTN among artisans in Ogun State.

scholarly journals Left isn't always right: placement of pictorial and textual package elements

2013 ◽  
Vol 115 (8) ◽  
pp. 1211-1225 ◽  
Author(s):  
Tobias Otterbring ◽  
Poja Shams ◽  
Erik Wästlund ◽  
Anders Gustafsson
Keyword(s):  
Visual Attention ◽  
Information Acquisition ◽  
Detection Time ◽  
Chi Square ◽  
Content Type ◽  
Design Elements ◽  
Left Hand ◽  
Previous Design ◽  
The Right ◽  
Chi Square Analysis

PurposeThe purpose of this study is to investigate how the positioning of textual and pictorial design elements on a package affects visual attention (detection time) toward these element types.Design/methodology/approachThe study has a 3×2 (stimulus×location) between‐subjects design. One pictorial and two textual package elements, located on the top right‐ or top left‐hand side of a package, were used as stimuli. Visual attention was measured by eye‐tracking. A total of 199 university students participated. The data were analysed using a two‐way ANOVA and a Pearson's chi‐square analysis with standardised residuals.FindingsThe results show that in order to receive the most direct attention, textual elements should be on the left‐hand side of a package, whereas pictorial elements should be on the right‐hand side. This is inconsistent with previous design directions (based on recall), suggesting the opposite element organisation.Originality/valuePrevious research has focused on recall (whether respondents remember having seen package elements) or preference (whether respondents prefer a package based on element positioning). The focus of the present study determined whether respondents actually saw the different elements on a package, and how long it took them to detect such elements. Detection time for certain element types can be viewed as a new and complementary way of evaluating the position of package elements. The paper also addresses whether preference is a result of easy information acquisition.

A neural observer for sensorless speed control of servomotors

2014 ◽  
Vol 31 (8) ◽  
pp. 1668-1678 ◽  
Author(s):  
Jenq-Ruey Horng ◽  
Ming-Shyan Wang ◽  
Tai-Rung Lai ◽  
Sergiu Berinde
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Design Methodology ◽  
Computation Time ◽  
Elman Neural Network ◽  
Content Type ◽  
On Line ◽  
Future Work ◽  
Practical Implications ◽  
Neural Speed

Purpose – Extensive efforts have been conducted on the elimination of position sensors in servomotor control. The purpose of this paper is to aim at estimating the servomotor speed without using position sensors and the knowledge of its parameters by artificial neural networks (ANNs). Design/methodology/approach – A neural speed observer based on the Elman neural network (NN) structure takes only motor voltages and currents as inputs. Findings – After offline NNs training, the observer is incorporated into a DSP-based drive and sensorless control is achieved. Research limitations/implications – Future work will consider to reduce the computation time for NNs training and to adaptively tune parameters on line. Practical implications – The experimental results of the proposed method are presented to show the effectiveness. Originality/value – This paper achieves sensorless servomotor control by ANNs which are seldom studied.

Prevalence of cerebral aneurysms in patients with fibromuscular dysplasia: a reassessment

1998 ◽  
Vol 88 (3) ◽  
pp. 436-440 ◽  
Author(s):  
Harry J. Cloft ◽  
David F. Kallmes ◽  
Michelle H. Kallmes ◽  
Jonas H. Goldstein ◽  
Mary E. Jensen ◽  
...  
Keyword(s):  
Carotid Artery ◽  
Fibromuscular Dysplasia ◽  
Internal Carotid ◽  
Cerebral Aneurysms ◽  
Chi Square ◽  
Content Type ◽  
Significant Difference ◽  
Chi Square Test ◽  
Using Data ◽  
Fine Print

Object. The aim of this study was to determine the prevalence of cerebral saccular aneurysms in patients with carotid artery and/or vertebral artery (VA) fibromuscular dysplasia (FMD). Methods. A metaanalysis was performed using data from 17 previously reported series of patients with internal carotid artery (ICA) and/or VA FMD that included information on the prevalence of cerebral aneurysms. In addition, the authors retrospectively evaluated their own series of 117 patients with ICA and/or VA FMD to determine the prevalence of cerebral aneurysms. The metaanalysis of the 17 earlier series, which included 498 patients, showed a 7.6 ± 2.5% prevalence of incidental, asymptomatic aneurysms in patients with ICA and/or VA FMD. In the authors' series of patients with FMD, 6.3 ± 4.9% of patients harbored an incidental, asymptomatic aneurysm. When the authors' series was combined with those included in the metaanalysis, the prevalence was found to be 7.3 ± 2.2%. The prevalence of aneurysms in the general population would have to be greater than 5.6% for there to be no statistically significant difference (chi-square test, p < 0.05) when compared with this 7.3% prevalence in patients with FMD. Conclusions. The prevalence of intracranial aneurysms in patients with cervical ICA and/or VA FMD is approximately 7%, which is not nearly as high as the 21 to 51% prevalence that has been previously reported.

Sign in / Sign up

Export Citation Format

Share Document