2,448 research outputs found
CoMFormer: Continual Learning in Semantic and Panoptic Segmentation
No full textContinual learning for segmentation has recently seen increasing interest.
However, all previous works focus on narrow semantic segmentation and disregard
panoptic segmentation, an important task with real-world impacts. %a In this
paper, we present the first continual learning model capable of operating on
both semantic and panoptic segmentation. Inspired by recent transformer
approaches that consider segmentation as a mask-classification problem, we
design CoMFormer. Our method carefully exploits the properties of transformer
architectures to learn new classes over time. Specifically, we propose a novel
adaptive distillation loss along with a mask-based pseudo-labeling technique to
effectively prevent forgetting. To evaluate our approach, we introduce a novel
continual panoptic segmentation benchmark on the challenging ADE20K dataset.
Our CoMFormer outperforms all the existing baselines by forgetting less old
classes but also learning more effectively new classes. In addition, we also
report an extensive evaluation in the large-scale continual semantic
segmentation scenario showing that CoMFormer also significantly outperforms
state-of-the-art methods.Comment: Under submissio
Feasibility of using training cases from International Spinal Cord Injury Core Data Set for testing of International Standards for Neurological Classification of Spinal Cord Injury items
No full textStudy design: Descriptive comparison analysis. Objective: To evaluate whether five training cases of International Spinal Cord Injury Core Data Set (ISCICDS) are appropriate for testing the facts within the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) and could thus be used for testing its training effectiveness. Methods: The authors reviewed the five training cases from the ISCICDS and determined the sensory level (SL), motor level (ML) and American Spinal Injury Association Impairment Scale (AIS) for the training cases. The key points from the training cases were compared with our interpretation of the key aspects of the ISNCSCI. Results: For determining SL, three principles of ML, sacral sparing, complete injury, classification of AIS A, B, C and D, determining motor incomplete status through sparing of motor function more than three levels below the ML, there are corresponding case scenarios in ISCICDS. However, no case scenario shows classification of AIS E and the use of voluntary anal sphincter contraction for determination of motor incomplete status. Neurological level of injury could be deduced from the SL and ML. Finally, none of the cases include information about zone of partial preservation, sensory score or motor score. Conclusion: Majority of the facts related to SL, ML and AIS are included in the five training cases of ISCICDS. Thus, using these training cases, it is feasible to test the above facts within the ISNCSCI. It is suggested that the missing fact should be included in an update of the training cases.Clinical NeurologyRehabilitationSCI(E)[email protected]
Robotics Training Algorithms for Optimizing Motor Learning in Spinal Cord Injured Subjects
No full textThe circuitries within the spinal cord are remarkably robust and plastic. Even in the absence of supraspinal control, such circuitries are capable of generating functional movements and changing their level of excitability based on a specific combination of properceptive inputs going into the spinal cord. This has led to an increase in locomotor training, such as Body Weight Support Treadmill training (BWST) for spinal cord injured (SCI) patients. However, today, little is known about the underlying physiological mechanisms responsible for the locomotor recovery achieved with this type of rehabilitative training, and the optimal rehabilitative strategy is still unknown.
This thesis describes a mouse model to study the effect of rehabilitative training on SCI. Using this model, the effects of locomotor recovery on adult spinal mice following complete spinal cord transaction is examined. Results that indicate adult spinal mice can be robotically trained to step, and when combined with the administration of quipazine (a broad serotonin agonist), there is an interaction and retention effect. Results also demonstrate that the training paradigm can be optimized in using “Assisted-as-Needed” (AAN) training. To find the optimal AAN training parameters, a learning model is developed to test the effect of various parameters of the AAN training algorithm. Simulation results from our model show that learning is training-dependent. In addition, the model predicts that improved motor learning can improve post-SCI by making the AAN training more adaptable.
The primary contributions of this thesis are twofold, in biology and engineering. We develop a mouse model using novel robotic devices and controls that can be used to study SCI and other locomotor disorders in the future by taking advantage of the many different strains of transgenic mice that are commercially available. We also further confirm that sensory integration responsible for motor control is distributed throughout the hierarchy of the neuromuscular system and can be achieved within the isolated spinal cord. Lastly, by developing a learning model, we can start looking into how variability plays a role in motor learning, the understanding of which will have profound implications in neurophysiology, machine learning and adaptive optimal controls research.</p
A systematic review of evidence on malignant spinal metastases : natural history and technologies for identifying patients at high risk of vertebral fracture and spinal cord compression
No full textBackground: Spinal metastases can lead to significant morbidity and reduction in quality of life due to spinal cord compression (SCC). Between 5% and 20% of patients with spinal metastases develop metastatic spinal cord compression during the course of their disease. An early study estimated average survival for patients with SCC to be between 3 and 7 months, with a 36% probability of survival to 12 months. An understanding of the natural history and early diagnosis of spinal metastases and prediction of collapse of the metastatic vertebrae are important.
Objective: To undertake a systematic review to examine the natural history of metastatic spinal lesions and to identify patients at high risk of vertebral fracture and SCC.
Data sources: The search strategy covered the concepts of metastasis, the spine and adults. Searches were undertaken from inception to June 2011 in 13 electronic databases [MEDLINE; MEDLINE In-Process & Other Non-Indexed Citations; EMBASE; Cochrane Database of Systematic Reviews; Cochrane Central Register of Controlled Trials (CENTRAL); Database of Abstracts of Reviews of Effects (DARE), NHS Economic Evaluation Database (NHS EED), HTA databases (NHS Centre for Reviews and Dissemination); Science Citation Index and Conference Proceedings (Web of Science); UK Clinical Research Network (UKCRN) Portfolio Database; Current Controlled Trials; ClinicalTrials.gov].
Review methods: Titles and abstracts of retrieved studies were assessed by two reviewers independently. Disagreement was resolved by consensus agreement. Full data were extracted independently by one reviewer. All included studies were reviewed by a second researcher with disagreements resolved by discussion. A quality assessment instrument was used to assess bias in six domains: study population, attrition, prognostic factor measurement, outcome measurement, confounding measurement and account, and analysis. Data were tabulated and discussed in a narrative review. Each tumour type was looked at separately.
Results: In all, 2425 potentially relevant articles were identified, of which 31 met the inclusion criteria. No study examined natural history alone. Seventeen studies reported retrospective data, 10 were prospective studies, and three were other study designs. There was one systematic review. There were no randomised controlled trials (RCTs). Approximately 5782 participants were included. Sample sizes ranged from 41 to 859. The age of participants ranged between 7 and 92 years. Types of cancers reported on were lung alone (n= 3), prostate alone (n= 6), breast alone (n= 7), mixed cancers (n= 13) and unclear (n= 1). A total of 93 prognostic factors were identified as potentially significant in predicting risk of SCC or collapse. Overall findings indicated that the more spinal metastases present and the longer a patient was at risk, the greater the reported likelihood of development of SCC and collapse. There was an increased risk of developing SCC if a cancer had already spread to the bones. In the prostate cancer studies, tumour grade, metastatic load and time on hormone therapy were associated with increased risk of SCC. In one study, risk of SCC before death was 24%, and 2.37 times greater with a Gleason score 7 than with a score of < 7 (p= 0.003). Other research found that patients with six or more bone lesions were at greater risk of SCC than those with fewer than six lesions [odds ratio (OR) 2.9, 95% confidence interval (CI) 1.012 to 8.35, p= 0.047]. For breast cancer patients who received a computerised tomography (CT) scan for suspected SCC, multiple logistic regression in one study identified four independent variables predictive of a positive test: bone metastases 2 years (OR 3.0 95% CI 1.2 to 7.6; p= 0.02); metastatic disease at initial diagnosis (OR 3.4, 95% CI 1.0 to 11.4; p= 0.05); objective weakness (OR 3.8, 95% CI 1.5 to 9.5; p= 0.005); and vertebral compression fracture on spine radiograph (OR 2.6, 95% CI 1.0 to 6.5; p= 0.05). A further study on mixed cancers, among patients who received surgery for SCC, reported that vertebral body compression fractures were associated with presurgery chemotherapy (OR 2.283, 95% CI 1.064 to 4.898; p= 0.03), cancer type [primary breast cancer (OR 4.179, 95% CI 1.457 to 11.983; p= 0.008)], thoracic involvement (OR 3.505, 95% CI 1.343 to 9.143; p= 0.01) and anterior cord compression (OR 3.213, 95% CI 1.416 to 7.293; p= 0.005).
Limitations: Many of the included studies provided limited information about patient populations and selection criteria and they varied in methodological quality, rigour and transparency. Several studies identified type of cancer (e.g. breast, lung or prostate cancer) as a significant factor in predicting SCC, but it remains difficult to determine the risk differential partly because of residual bias. Consideration of quantitative results from the studies does not easily allow generation of a coherent numerical summary, studies were heterogeneous especially with regard to population, results were not consistent between studies, and study results almost universally lacked corroboration from other independent studies.
Conclusion: No studies were found which examined natural history. Overall burden of metastatic disease, confirmed metastatic bone involvement and immediate symptomatology suggestive of spinal column involvement are already well known as factors for metastatic SCC, vertebral collapse or progression of vertebral collapse. Although we identified a large number of additional possible prognostic factors, those which currently offer the most potential are unclear. Current clinical consensus favours magnetic resonance imaging and CT imaging modalities for the investigation of SCC and vertebral fracture. Future research should concentrate on: (1) prospective randomised designs to establish clinical and quality-of-life outcomes and cost-effectiveness of identification and treatment of patients at high risk of vertebral collapse and SCC; (2) Service Delivery and Organisation research on magnetic resonance imaging (MRI) scans and scanning (in tandem with research studies on use of MRI to monitor progression) in order to understand best methods for maximising use of MRI scanners; and (3) investigation of prognostic algorithms to calculate probability of a specified event using high-quality prospective studies, involving defined populations, randomly selected and clearly identified samples, and with blinding of investigators
A Political Life
No full textDaniel Defert was Michel Foucault’s partner for more than the last 20 years of his life. A student during the early 1960s, he supported the Algerian liberation movement, participated in the activist Gauche prolétarienne and, together with Foucault, founded GIP (Groupe d’information sur les prisons), a group that gathered and published information about the conditions in French prisons. After Foucault died in 1984 of AIDS, without having been given the precise diagnosis by any of his doctors, Defert founded AIDES, to this day the largest organization for the support of HIV-positive people and people with AIDS in France. AIDES started out as a network of mutual assistance and care but also exerted political influence and fought against the discrimination of HIV-positive people and the vilification of homosexuality. The Berlin publisher Merve has just issued the German translation of Defert’s interview biography A Political Life. The conversation with Cord Riechelmann will revolve around the historical context of the political struggles which Defert and Foucault engaged in and critical questions about the relation between society and homosexuality that are still with us today.
Daniel Defert, born in 1937, is a sociologist, philosopher, activist and founder of the organization AIDES, which he presided over until 1991. He has served as co-editor of the writings of Michel Foucault. Cord Riechelmann is a biologist, philosopher, and author who writes, among others, for the Frankfurter Allgemeine Sonntagszeitung and Jungle World. Introduction by Elisa Barth (Merve
Cardiovascular response during urodynamics in individuals with spinal cord injury
No full textStudy design: Retrospective chart review. Objectives: To establish the frequency and severity of autonomic dysreflexia (AD) during urodynamics among individuals with chronic spinal cord injury (SCI) and to investigate the possible effect of the number of years since SCI on the severity of AD. Setting: SCI outpatient clinic. Methods: A retrospective chart review was undertaken of individuals with SCI who were seen at an outpatient clinic and could potentially develop an episode of AD (T6 and above). Data regarding age, gender, urodynamic examination, lower urinary tract function, cardiovascular parameters and SCI were collected. In addition, information on signs and symptoms of AD were retrieved. Results: A total of 76 individuals with SCI were examined with blood pressure (BP) monitoring. The majority had cervical SCI (79%). The mean age was 47.8 +/- 13.9 years. The median duration after SCI was 51.5 months. During urodynamics, a total of 48 (63.2%) individuals showed an increase in systolic BP>20 mm Hg, meeting the criteria for AD. Indicators for higher incidences of AD were cervical SCI, being 42 years after SCI, the presence of detrusor sphincter dyssynergia (DSD) and low bladder compliance. AD was more severe in individuals with complete (American Spinal Cord Association (ASIA) impairment scale (AIS) A) injuries, worse with greater time after SCI. Conclusion: Individuals with cervical SCI, DSD, poor bladder compliance or 42 years after SCI were associated with a higher possibility of developing AD during urodynamics. Furthermore, AD was more severe in complete (AIS A) individuals and was exacerbated with time after injury.State scholarship from China Scholarship Council; Canadian Heart and Stroke Foundation; Canadian Institute of Health Research; Canadian Foundation for Innovation; Craig Neilsen Foundation; Christopher and Dana Reeve FoundationSCI(E)PubMedARTICLE3279-2845
Pediatric spinal cord injury in a subset of European countries
No full textStudy design:Postal survey.Setting:A total of 19 countries in Europe.Objectives:Firstly, to collect information about incidence and systems of care for pediatric spinal cord injury (pedSCI); including prevention, initial care and follow-up in a subset of European countries. Secondly, to initiate a network of involved professionals for exchange of information and development of research and care programs.Methods:A short semi-structured questionnaire was sent to respondents working with spinal cord injury (SCI) in 19 countries in Europe.Results:Only in Portugal and Sweden, is the incidence of pedSCI (fatal injuries included) established, that is 27 children/million children/year and 4.6 children/million children/year, respectively. For the other countries, the estimated incidence of pedSCI (nonfatal injuries) varied from 0.9 to 21.2 children/million children/year in the age group of 0–14 years. Although the incidence varies considerably, pedSCI is rare throughout Europe. The management differs between the countries depending on the age of the child and the local organization of health care.Conclusions:The survey confirms that pedSCI is rare. In order to establish high-quality standardized care, further integration of knowledge in this area is needed throughout Europe. The contacts initiated by this survey may be used to create an international network serving as a reference for health professionals, researchers and families, thereby possibly alleviating some of the unwanted variations of care identified in this study.Spinal Cord (2006) 44, 106–112. doi:andlt;highlightandgt;10.1038andlt;/highlightandgt;/andlt;highlightandgt;sj.scandlt;/highlightandgt;.andlt;highlightandgt;3101793andlt;/highlightandgt;; published online 30 August 2005 [ABSTRACT FROM AUTHOR]</p
Training effectiveness when teaching the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) to medical students
No full textStudy design: Interventional training sessions. Objectives: To examine the effectiveness of training medical students in the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI). Setting: A Peking University teaching hospital. Methods: A total of 46 medical students were involved in the study. First, they had a 2-h self-study session with the ISNCSCI booklet, followed by a 10-item questionnaire. The booklet was allowed for use as a reference during this assessment. Two days later, the questionnaire was repeated without the use of reference. Students then had a session with case discussion, followed by a final questionnaire. Results: After the initial self-study session, the mean score (max. 10) on the questionnaire was 7.67 +/- 1.49. Two days later, the mean score of the second test was 7.96 +/- 1.15. All key points of the ISNCSCI were supplemented during the second session with case presentations. The mean score of the final test increased significantly to 9.61 +/- 0.88 (P<0.01) in comparison with the previous tests. The overall correct response rates by students in determining sensory level, motor level and American Spinal Injury Association Impairment Scale of the training case were 89.1%, 84.8% and 91.3%, respectively. Conclusion: The training effectiveness of ISNCSCI through self-study is reliable. The correct answers to key points could remain for at least 2 days without the need to use a reference. However, some specialized knowledge could not be transmitted without more detailed discussions and case presentations. Utilization of cases is a valuable method in training ISNCSCI and can improve the overall training effectiveness.Clinical NeurologyRehabilitationSCI(E)PubMed2ARTICLE10768-7715
Pediatric spinal cord injury in Sweden: incidence, etiology and outcome
No full textSTUDY DESIGN:: Retrospective descriptive study. OBJECTIVES:: To assess incidence, causes and early outcome of traumatic spinal cord injury (SCI) among children in Sweden, thereby identifying high-risk groups and situations as a basis for preventative measures and improved care. METHODS:: Data from population registers, County Habilitation Centers as well as from informal sources were used to estimate the incidence of SCI in Sweden during the years 1985-1996 among children aged 0-15 years. Contacts with the treating hospitals, reviews of medical records and/or personal interviews were used to verify primary data. In total, 92 children were thus identified. RESULTS:: The incidence was found to be 4.6/million children/year (95% CI 3.6-5.5). When excluding prehospital fatalities, the incidence was 2.4 (95% CI 1.8-3.1). The main cause of injury among fatalities was traffic accidents. Associated injuries occurred in 41% of the children. Among survivors (10-15 years), sports-related injuries (43%) were as common a cause as traffic accidents (39%). The survivors were treated in 18 different hospitals. CONCLUSION:: Pediatric SCI in Sweden is rare, presumably because of effective primary prevention. Preventative measures should be further differentiated for each age group of children in accordance with their differing risk profiles. In contrast to the effective preventative measures that have been implemented in Sweden, care of these patients is still too fractionated and decentralized for sufficient specialization to emerge.Spinal Cord (2003) 41, 328-336. doi:andlt;highlightandgt;10.1038andlt;/highlightandgt;/andlt;highlightandgt;sj.scandlt;/highlightandgt;.andlt;highlightandgt;3101478andlt;/highlightandgt; [ABSTRACT FROM AUTHOR]</p
Pediatric spinal cord injury in Sweden; how to identify a cohort of rare events
No full textSTUDY DESIGN:: Register study enhanced and verified by medical records and personal interviews and examinations. SETTINGS:: Sweden. OBJECTIVES:: To define a method of identifying a study population of rare events. To point out the relative importance of every step, an example is given of identifying persons who sustained traumatic spinal cord injury (SCI) in childhood. METHODS:: Cases were identified in seven steps that all needed to be fulfilled, from definition of selection criteria through combination of several data sources, to the use of several verification methods. RESULTS:: Initial screening by registers identified 384 possible cases, which however were found by subsequent analysis to include a large number of incorrect cases. At completion of all analytic steps, 35 living cases could be fully verified and 14 deceased cases could be partially verified. CONCLUSIONS:: Registers offer a practical initial source for study population identification. The screening of International Classification of Diseases codes defining SCI only included less than 30% of true SCIs. Subsequently, further refinement and quality control is necessary in order to ensure validity. Such further verification is time-consuming, but nevertheless necessary in order to verify a true cohort.Spinal Cord (2003) 41, 337-346. doi:andlt;highlightandgt;10.1038andlt;/highlightandgt;/andlt;highlightandgt;sj.scandlt;/highlightandgt;.andlt;highlightandgt;3101456andlt;/highlightandgt; [ABSTRACT FROM AUTHOR]</p
- …
