Classification Accuracy Assessment Research Articles

Analysis of Mountain Rangeland Changes, Using Remotely-Sensed Data, (Case Study: Dehdez Area, Iran)

The land-cover dynamics has been quite conspicuous over the last three decades in Dehdez area, Iran. Therefore, the present study was undertaken in the Dehdez area to assess the trends of rangelands dynamics in the study area during the period 1990-2006. Two clear, cloud-free Landsat and one ASTER images were selected to classify the study area. All images were rectified to UTM zone 39, WGS84 using at least 25 well distributed ground control points and nearest neighbor resampling. Land-use/cover mapping is achieved through interpretation of Landsat TM satellite images of 1990, 1998 and ASTER image of 2006. Fieldwork was carried out to collect data for training and validating land-use/cover interpretation from satellite image of 2006, and for qualitative description of the characteristics of each land-use/cover class. In order to create a testing sample set, first of all, a set of testing points was selected randomly. A supervised classification technique with Maximum Likelihood Algorithm was applied based on 48 training samples for the image of 2006, and 42 samples for the images of 1990 and 1998 and the land-use/cover maps were produced. Error matrices were used to assess classification accuracy. The results showed rangeland covers about 30.8%, 36.7% and 45% of the total geographical area of the Dehdez area in 1990, 1998 and 2006, respectively. Overall accuracies of land-use/cover classification for 1990, 1998 and 2006 were 89.37%, 75.24% and 71.14%, respectively. Kappa values obtained were of 78.71%, 55.61% and 51.41% of accuracy for the 1990, 1998 and 2006, respectively. During 16 years span period (1990-2006) about 1738.4 ha, 383.7 ha, 32.8 ha and 890.1 ha of rangelands were converted to forest, agriculture, water and settlement. The total rich rangelands in the area, accounted for 38.5%, 44% and 42.2% in 1990, 1998 and 2006, respectively. The total poor rangeland in the area accounted for 61.5%, 56% and 57.8% in 1990, 1998 and 2006, respectively. Satellite Remote Sensing enabled the generation of a detailed rangeland map and the separation of grazing intensity levels in rangelands could be generated with the relatively little effort in areas that were difficult to access.

Relative quantity and allocation disagreement measures for category-level accuracy assessment

Accuracy assessment of classified imagery is an important task in remote sensing. A commonly used approach for assessing classification accuracy is based on the error matrix. The error matrix can be summarized by overall and category accuracy measures, e.g. the overall agreement and the user’s and producer’s accuracies. Two relatively new measures are the so-called quantity disagreement and allocation disagreement. In this article, relative versions of the category-level variants of these measures are defined. It is also shown how the new category measures are related to the overall quantity and allocation disagreement measures.

Evaluating an unmanned aerial vehicle‐based approach for assessing habitat extent and condition in fine‐scale early successional mountain mosaics

AbstractQuestionCan very high‐resolution colour orthophotography and digital surface models (DSMs) from an unmanned aerial vehicle (UAV) be effectively used for assessment of habitat extent and condition in fine‐scale disturbance‐dependent mosaics?LocationSerra de Arga mountain range, a Natura 2000 protected site in the NW region of Portugal where drastic changes in pastoral activities have occurred over recent decades.MethodsAn UAV platform was used to collect very high‐resolution (6 cm) images and to produce a DSM (10 cm). From these data, several features were extracted related to colour, band ratios, as well as texture features calculated from colour imagery and surface elevation. Based on a systematic sampling design, field data were collected for both training and validation of a supervised classifier. Extracted features and ground truth training data were combined to calibrate a pixel‐based Random forest classifier, with the purpose of devising a habitat map for the entire study area. Map validation was performed to assess classification accuracy, and feature importance metrics were calculated.ResultsValidation results revealed good mean overall accuracy (0.89), with some performance decrease in situations of high interspersion of habitat types. The priority habitat type 6230* (Nardus grasslands), defining the vegetation matrix of the test site, obtained 0.96 and 0.91, considering, respectively, producer and user accuracy. In turn, priority habitat type 4020* (Atlantic wet heathlands) recorded 0.68 and 0.77. The obtained habitat map allowed measurement of the extent, description of the spatial arrangement and provided an indication of the conservation condition of target habitat types. Test results regarding the discrimination ability of different features highlighted the importance of surface elevation textures derived from the DSM, followed by band ratios textures and other more complex texture features calculated from colour imagery.ConclusionsOverall, the developed methodology showed promising results for assessing the extent and condition of habitats of high conservation priority in fine‐scale, dynamic vegetation mosaics. Future advances in the use of UAV platforms may play an important role in monitoring protected sites and fulfil legal reporting obligations of EU member states, while reducing the costs associated with intensive in‐field assessments.

Railcar Bogie Performance Monitoring using Mutual Information and Support Vector Machines


 
 
 Railcar condition monitoring is an area of high importance and global relevance. The economic and safety concerns of equipment maintenance in North America mandate efforts in prognostics and health management. This paper presents the results from the development of a vibration based condition monitoring algorithm for freight rail, utilizing mutual information feature selection and support vector machine classification of bogie component faults. The algorithm is an implementation of a previously proposed railcar condition monitoring solution by the authors. The proposed monitoring solution is a data-driven method which was developed with measurements taken at a railroad test laboratory under controlled conditions. Vibration data was collected from multiple locations on a railcar over several test runs, each utilizing wheelsets with different levels of wear. The input of controlled wheel wear levels was aimed at varying the system outputs to resemble those of cars with different levels of mileage in revenue service. The generated data sets were processed and a feature set was extracted from the acceleration signals. The data was divided into training and validation partitions using a cross validation scheme to preserve the sequence for both sets. A mutual information (MI) estimation algorithm was used to rank the features based on their similarity to the classified fault state. Both the optimized feature set from the MI feature selection algorithm as well as the full, non-discriminate feature set were used as inputs to the support vector machine to assess classification accuracy. The results of this assessment are presented in the paper along with a presentation of the methods. The paper concludes with a proposal for a monitoring strategy aimed at specifically detecting faulty components and practicing predictive maintenance.
 
 

Mapping the land-cover distribution in arid and semiarid urban landscapes with Landsat Thematic Mapper imagery

Mapping the land-cover distribution in arid and semiarid urban landscapes using medium spatial resolution imagery is especially difficult due to the mixed-pixel problem in remotely sensed data and the confusion of spectral signatures among bare soils, sparse density shrub lands, and impervious surface areas (ISAs hereafter). This article explores a hybrid method consisting of linear spectral mixture analysis (LSMA), decision tree classifier, and cluster analysis for mapping land-cover distribution in two arid and semiarid urban landscapes: Urumqi, China, and Phoenix, USA. The Landsat Thematic Mapper (TM) imagery was unmixed into four endmember fraction images (i.e. high-albedo object, low-albedo object, green vegetation (GV), and soil) using the LSMA approach. New variables from these fraction images and TM spectral bands were used to map seven land-cover classes (i.e. forest, shrub, grass, crop, bare soil, ISA, and water) using the decision tree classifier. The cluster analysis was further used to modify the classification results. QuickBird imagery in Urumqi and aerial photographs in Phoenix were used to assess classification accuracy. Overall classification accuracies of 86.0% for Urumqi and 88.7% for Phoenix were obtained, much higher accuracies than those utilizing the traditional maximum likelihood classifier (MLC). This research demonstrates the necessity of using new variables from fraction images to distinguish between ISA and bare soils and between shrub and other vegetation types. It also indicates the different effects of spatial patterns of land-cover composition in arid and semiarid landscapes on urban land-cover classification.

Validation of Land Cover Products Using Reliability Evaluation Methods

Validation of land cover products is a fundamental task prior to data applications. Current validation schemes and methods are, however, suited only for assessing classification accuracy and disregard the reliability of land cover products. The reliability evaluation of land cover products should be undertaken to provide reliable land cover information. In addition, the lack of high-quality reference data often constrains validation and affects the reliability results of land cover products. This study proposes a validation schema to evaluate the reliability of land cover products, including two methods, namely, result reliability evaluation and process reliability evaluation. Result reliability evaluation computes the reliability of land cover products using seven reliability indicators. Process reliability evaluation analyzes the reliability propagation in the data production process to obtain the reliability of land cover products. Fuzzy fault tree analysis is introduced and improved in the reliability analysis of a data production process. Research results show that the proposed reliability evaluation scheme is reasonable and can be applied to validate land cover products. Through the analysis of the seven indicators of result reliability evaluation, more information on land cover can be obtained for strategic decision-making and planning, compared with traditional accuracy assessment methods. Process reliability evaluation without the need for reference data can facilitate the validation and reflect the change trends of reliabilities to some extent.

Simulation of seagrass bed mapping by satellite images based on the radiative transfer model

Seagrass and seaweed beds play important roles in coastal marine ecosystems. They are food sources and habitats for many marine organisms, and influence the physical, chemical, and biological environment. They are sensitive to human impacts such as reclamation and pollution. Therefore, their management and preservation are necessary for a healthy coastal environment. Satellite remote sensing is a useful tool for mapping and monitoring seagrass beds. The efficiency of seagrass mapping, seagrass bed classification in particular, has been evaluated by mapping accuracy using an error matrix. However, mapping accuracies are influenced by coastal environments such as seawater transparency, bathymetry, and substrate type. Coastal management requires sufficient accuracy and an understanding of mapping limitations for monitoring coastal habitats including seagrass beds. Previous studies are mainly based on case studies in specific regions and seasons. Extensive data are required to generalise assessments of classification accuracy from case studies, which has proven difficult. This study aims to build a simulator based on a radiative transfer model to produce modelled satellite images and assess the visual detectability of seagrass beds under different transparencies and seagrass coverages, as well as to examine mapping limitations and classification accuracy. Our simulations led to the development of a model of water transparency and the mapping of depth limits and indicated the possibility for seagrass density mapping under certain ideal conditions. The results show that modelling satellite images is useful in evaluating the accuracy of classification and that establishing seagrass bed monitoring by remote sensing is a reliable tool.

Accurate Classification of Chronic Migraine via Brain Magnetic Resonance Imaging.

The International Classification of Headache Disorders provides criteria for the diagnosis and subclassification of migraine. Since there is no objective gold standard by which to test these diagnostic criteria, the criteria are based on the consensus opinion of content experts. Accurate migraine classifiers consisting of brain structural measures could serve as an objective gold standard by which to test and revise diagnostic criteria. The objectives of this study were to utilize magnetic resonance imaging measures of brain structure for constructing classifiers: (1) that accurately identify individuals as having chronic vs episodic migraine vs being a healthy control; and (2) that test the currently used threshold of 15 headache days/month for differentiating chronic migraine from episodic migraine. Study participants underwent magnetic resonance imaging for determination of regional cortical thickness, cortical surface area, and volume. Principal components analysis combined structural measurements into principal components accounting for 85% of variability in brain structure. Models consisting of these principal components were developed to achieve the classification objectives. Tenfold cross validation assessed classification accuracy within each of the 10 runs, with data from 90% of participants randomly selected for classifier development and data from the remaining 10% of participants used to test classification performance. Headache frequency thresholds ranging from 5-15 headache days/month were evaluated to determine the threshold allowing for the most accurate subclassification of individuals into lower and higher frequency subgroups. Participants were 66 migraineurs and 54 healthy controls, 75.8% female, with an average age of 36 +/- 11 years. Average classifier accuracies were: (1) 68% for migraine (episodic + chronic) vs. healthy controls; (2) 67.2% for episodic migraine vs healthy controls; (3) 86.3% for chronic migraine vs. healthy controls; and (4) 84.2% for chronic migraine vs episodic migraine. The classifiers contained principal components consisting of several structural measures, commonly including the temporal pole, anterior cingulate cortex, superior temporal lobe, entorhinal cortex, medial orbital frontal gyrus, and pars triangularis. A threshold of 15 headache days/month allowed for the most accurate subclassification of migraineurs into lower frequency and higher frequency subgroups. Classifiers consisting of cortical surface area, cortical thickness, and regional volumes were highly accurate for determining if individuals have chronic migraine. Furthermore, results provide objective support for the current use of 15 headache days/month as a threshold for dividing migraineurs into lower frequency (i.e., episodic migraine) and higher frequency (i.e., chronic migraine) subgroups.

Rice Fields Mapping in Fragmented Area Using Multi-Temporal HJ-1A/B CCD Images

Rice is one of the most important crops in the world; meanwhile, the rice field is also an important contributor to greenhouse gas methane emission. Therefore, it is important to get an accurate estimation of rice acreage for both food production and climate change related studies. The eastern plain region is one of the major single-cropped rice (SCR) growing areas in China. Subjected to the topography and intensified human activities, the rice fields are generally fragmented and irregular. How remote sensing can meet this challenge to accurately estimate the acreage of the rice in this region using medium-resolution imagery is the topic of this study. In this study, the applicability of the Chinese HJ-1A/B satellites and a two-band enhanced vegetation index (EVI2) was investigated. Field campaigns were carried out during the rice growing season and ground-truth data were collected for classification accuracy assessments in 2012. A stepwise classification strategy utilizing the EVI2 signatures during key phenology stages, i.e., the transplanting and the vegetative to reproductive transition phases, of the SCR was proposed, and the overall classification accuracy was 91.7%. The influence of the mixed pixel and boundary effects to classification accuracy was also investigated. This work demonstrates that the Chinese HJ-1A/B data are suitable data source to estimating SCR cropping area under complex land cover composition.

Examining the Capability of Supervised Machine Learning Classifiers in Extracting Flooded Areas from Landsat TM Imagery: A Case Study from a Mediterranean Flood

This study explored the capability of Support Vector Machines (SVMs) and regularised kernel Fisher’s discriminant analysis (rkFDA) machine learning supervised classifiers in extracting flooded area from optical Landsat TM imagery. The ability of both techniques was evaluated using a case study of a riverine flood event in 2010 in a heterogeneous Mediterranean region, for which TM imagery acquired shortly after the flood event was available. For the two classifiers, both linear and non-linear (kernel) versions were utilised in their implementation. The ability of the different classifiers to map the flooded area extent was assessed on the basis of classification accuracy assessment metrics. Results showed that rkFDA outperformed SVMs in terms of accurate flooded pixels detection, also producing fewer missed detections of the flooded area. Yet, SVMs showed less false flooded area detections. Overall, the non-linear rkFDA classification method was the more accurate of the two techniques (OA = 96.23%, K = 0.877). Both methods outperformed the standard Normalized Difference Water Index (NDWI) thresholding (OA = 94.63, K = 0.818) by roughly 0.06 K points. Although overall accuracy results for the rkFDA and SVMs classifications only showed a somewhat minor improvement on the overall accuracy exhibited by the NDWI thresholding, notably both classifiers considerably outperformed the thresholding algorithm in other specific accuracy measures (e.g. producer accuracy for the “not flooded” class was ~10.5% less accurate for the NDWI thresholding algorithm in comparison to the classifiers, and average per-class accuracy was ~5% less accurate than the machine learning models). This study provides evidence of the successful application of supervised machine learning for classifying flooded areas in Landsat imagery, where few studies so far exist in this direction. Considering that Landsat data is open access and has global coverage, the results of this study offers important information towards exploring the possibilities of the use of such data to map other significant flood events from space in an economically viable way.

Sex and Smoking Status Effects on the Early Detection of Early Lung Cancer in High-Risk Smokers Using an Electronic Nose.

Volatile organic compounds (VOCs) in exhaled breath as measured by electronic nose (e-nose) have utility as biomarkers to detect subjects at risk of having lung cancer in a screening setting. We hypothesize that breath analysis using an e-nose chemo-resistive sensor array could be used as a screening tool to discriminate patients diagnosed with lung cancer from high-risk smokers. Breath samples from 191 subjects-25 lung cancer patients and 166 high-risk smoker control subjects without cancer-were analyzed. For clinical relevancy, subjects in both groups were matched for age, sex, and smoking histories. Classification and regression trees and discriminant functions classifiers were used to recognize VOC patterns in e-nose data. Cross-validated results were used to assess classification accuracy. Repeatability and reproducibility of e-nose data were assessed by measuring subject-exhaled breath in parallel across two e-nose devices. e-Nose measurements could distinguish lung cancer patients from high-risk control subjects, with a better than 80% classification accuracy. Subject sex and smoking status impacted classification as area under the curve results (ex-smoker males 0.846, ex-smoker female 0.816, current smoker male 0.745, and current smoker female 0.725) demonstrated. Two e-nose systems could be calibrated to give equivalent readings across subject-exhaled breath measured in parallel. e-Nose technology may have significant utility as a noninvasive screening tool for detecting individuals at increased risk for lung cancer. The results presented further the case that VOC patterns could have real clinical utility to screen for lung cancer in the important growing ex-smoker population.

The Impact of Positional Errors on Soft Classification Accuracy Assessment: A Simulation Analysis

Validating or accessing the accuracy of soft classification maps has rapidly developed over the past few years. This assessment employs a soft error matrix as generalized from the traditional, hard classification error matrix. However, the impact of positional error on the soft classification is uncertain and whether the well-accepted half-pixel registration accuracy is suitable for the soft classification accuracy assessment is unknown. In this paper, a simulation analysis was conducted to examine the influence of positional error on the overall accuracy (OA) and kappa in soft classification accuracy assessment under different landscape conditions (i.e., spatial characteristics and spatial resolutions). Results showed that with positional error ranging from 0 to 3 soft pixels, the OA-error varied from 0 to 44.6 percent while the kappa-error varied from 0 to 93.7 percent. Landscape conditions with smaller mean patch size (MPS) and greater fragmentation produced greater positional error impact on the accuracy measures at spatial resolutions of 1 and 2 unit distances. However, this trend did not hold for spatial resolutions of 5 and 10 unit distances. A half of a pixel was not sufficient to keep the overall accuracy error and kappa error under 10 percent. The results indicate that for soft classification accuracy assessment the requirement for registration accuracy is higher and depends greatly on the landscape characteristics. There is a great need to consider positional error for validating soft classification maps of different spatial resolutions.

Accuracy Assessment of Land Use Land Cover Classification using Google Earth

This study examines the accuracy assessment of land use land cover classification using Google Earth in the case of Kilite Awulalo, Tigray State, Ethiopia for the year 2014. For this study, Landsat-8 OLI_TIRS image of 2014 was used and analyzed using Arc GIS 10.1. Supervised classification scheme was used to classify the images. Under land use and land cover categories Agriculture land, Settlement land, Grazing land, Forest land, Bush land, Water bodies and Bare/stony land were studied. After classification of land use land cover types, 100 Random Points were generated in Arc GIS and converting random points to KML in order to open in Google Earth. Each random point’s value verified from Google Earth for accuracy assessment. Google Earth model was used to measure of how many ground truth pixels are correctly classified. For this study, Free Google Earth which was Build in Date 10/7/2013 was used. The result shows that total (overall) accuracy of land use and land cover for 2014 is 82.00% and Kappa (K) is 77.02% which is acceptable in both accuracy total (overall) and Kappa accuracy.

Land Cover Classification Accuracy Assessment Using Full-Waveform LiDAR Data

Land Cover Classification Accuracy Assessment Using Full-Waveform LiDAR Data

Physiological reactivity to nonideographic virtual reality stimuli in veterans with and without PTSD.

BackgroundPost-traumatic stress disorder (PTSD) currently is diagnosed via clinical interview in which subjective self reports of traumatic events and associated experiences are discussed with a mental health professional. The reliability and validity of diagnoses can be improved with the use of objective physiological measures.MethodsIn this study, physiological activity was recorded from 58 male veterans (PTSD Diagnosis n = 16; Trauma Exposed/No PTSD Diagnosis: n = 23; No Trauma/No PTSD Diagnosis: n = 19) with and without PTSD and combat trauma exposure in response to emotionally evocative non-idiographic virtual reality stimuli.ResultsStatistically significant differences among the Control, Trauma, and PTSD groups were present during the viewing of two virtual reality videos. Skin conductance and interbeat interval features were extracted for each of ten video events (five events of increasing severity per video). These features were submitted to three stepwise discriminant function analyses to assess classification accuracy for Control versus Trauma, Control versus PTSD, and Trauma versus PTSD pairings of participant groups. Leave-one-out cross-validation classification accuracy was between 71 and 94%.ConclusionsThese results are promising and suggest the utility of objective physiological measures in assisting with PTSD diagnosis.

Enhanced Urban Landcover Classification for Operational Change Detection Study Using Very High Resolution Remote Sensing Data

Abstract. This study presents an operational case of advancements in urban land cover classification and change detection by using very high resolution spatial and multispectral information from 4-band QuickBird (QB) and 8-band WorldView-2 (WV-2) image sequence. Our study accentuates quantitative, pixel based, image difference approach for operational change detection using very high resolution pansharpened QB and WV-2 images captured over San Francisco city, California, USA (37° 44" 30N', 122° 31" 30' W and 37° 41" 30° N ,122° 20" 30' W). In addition to standard QB image, we compiled three multiband images from eight pansharpened WV-2 bands: (1) multiband image from four traditional spectral bands, i.e., Blue, Green, Red and near-infrared 1 (NIR1) (henceforth referred as "QB equivalent WV-2"), (2) multiband image from four new spectral bands, i.e., Coastal, Yellow, Red Edge and NIR2 (henceforth referred as "new band WV-2"), and (3) multiband image consisting of four traditional and four new bands (henceforth referred as "standard WV-2"). All the four multiband images were classified using support vector machine (SVM) classifier into four most abundant land cover classes, viz, hard surface, vegetation, water and shadow. The assessment of classification accuracy was performed using random selection of 356 test points. Land cover classifications on "standard QB" image (kappa coeffiecient, κ = 0.93), "QB equivalent WV-2" image (κ = 0.97), and "new band WV-2" image (κ = 0.97) yielded overall accuracies of 96.31 %, 98.03 % and 98.31 %, respectively, while "standard WV-2" image (κ = 0.99) yielded an improved overall accuracy of 99.18 %. It is concluded that the addition of four new spectral bands to the existing four traditional bands improved the discrimination of land cover targets, due to increase in the spectral characteristics of WV-2 satellite. Consequently, to test the validity of improvement in classification process for implementation in operational change detection application, comparative assessment of transition of various landcover classes in three WV-2 images with respect to "standard QB" image was carried out using image difference method. As far as waterbody class is concerned, there was no significant transition observed in all the three WorldView-2 Images, whereas, hard surface class showed lowest transition in "standard WV-2" image and highest in case of "new band WV-2". The most significant transition was occurred in vegetation class in all of the three images, showing positive change (increase) in standard WV-2 image (0.31 Sq. Km) and negative change (decrease) in other two images (−0.12 Sq. Km for "QB equivalent WV-2" image and −31.15 Sq. Km in "new band WV-2" image) with considerable amount. Similar case was observed with the shadow class, but the difference is, transition from shadow to other classes was negative in all the three WV-2 images which can be attributed to the fact that, "standard QB" image had more shadow area (based on acquisition time and sun position) than WV-2, that means all the band combinations of WV-2 succeeded in extracting the features hidden below the shadow in "standard QB" image. These trends indicate that the overall bandwise transition in landcover classes in case of "standard WV-2" is more precise than other two images. We note that "QB equivalent WV-2" image had narrower band widths than those of "standard QB" image but the observed vegetation change is not prominent as in case of other two images, but at the same time, transition in hard surface and waterbody was discerned more efficiently than "new band WV-2" image. The addition of new bands in WV-2 enabled more effective vegetation analysis, so the vegetation transition results shown by "new band WV-2" image were at par with the "standard WV-2" image, showing the importance of these newly added bands in the WV-2 imagery, with comparatively lower transitions in other classes. In a nutshell, it can be claimed that incorporation of new bands along with even narrower Red, Green, Blue and Near Infrared-1 bands in WV-2 image holds remarkable importance which leads to enhancement in the potential of WV-2 imagery in change detection and other feature extraction studies.

Hyperspectral Data for Land use/Land cover classification

Abstract. An attempt has been made to compare the multispectral Resourcesat-2 LISS III and Hyperion image for the selected area at sub class level classes of major land use/ land cover. On-screen interpretation of LISS III (resolution 23.5 m) was compared with Spectral Angle Mapping (SAM) classification of Hyperion (resolution 30m). Results of the preliminary interpretation of both images showed that features like fallow, built up and wasteland classes in Hyperion image are clearer than LISS-III and Hyperion is comparable with any high resolution data. Even canopy types of vegetation classes, aquatic vegetation and aquatic systems are distinct in Hyperion data. Accuracy assessment of SAM classification of Hyperion compared with the common classification systems followed for LISS III there was no much significant difference between the two. However, more number of vegetation classes could be classified in SAM. There is a misinterpretation of built up and fallow classes in SAM. The advantages of Hyperion over visual interpretation are the differentiation of the type of crop canopy and also crop stage could be confirmed with the spectral signature. The Red edge phenomenon was found for different canopy type of the study area and it clearly differentiated the stage of vegetation, which was verified with high resolution image. Hyperion image for a specific area is on par with high resolution data along with LISS III data.

Sitting and standing intention can be decoded from scalp EEG recorded prior to movement execution.

Low frequency signals recorded from non-invasive electroencephalography (EEG), in particular movement-related cortical potentials (MRPs), are associated with preparation and execution of movement and thus present a target for use in brain-machine interfaces. We investigated the ability to decode movement intent from delta-band (0.1–4 Hz) EEG recorded immediately before movement execution in healthy volunteers. We used data from epochs starting 1.5 s before movement onset to classify future movements into one of three classes: stand-up, sit-down, or quiet. We assessed classification accuracy in both externally triggered and self-paced paradigms. Movement onset was determined from electromyography (EMG) recordings synchronized with EEG signals. We employed an artifact subspace reconstruction (ASR) algorithm to eliminate high amplitude noise before building our time-embedded EEG features. We applied local Fisher's discriminant analysis to reduce the dimensionality of our spatio-temporal features and subsequently used a Gaussian mixture model classifier for our three class problem. Our results demonstrate significantly better than chance classification accuracy (chance level = 33.3%) for the self-initiated (78.0 ± 2.6%) and triggered (74.7 ± 5.7%) paradigms. Surprisingly, we found no significant difference in classification accuracy between the self-paced and cued paradigms when using the full set of non-peripheral electrodes. However, accuracy was significantly increased for self-paced movements when only electrodes over the primary motor area were used. Overall, this study demonstrates that delta-band EEG recorded immediately before movement carries discriminative information regarding movement type. Our results suggest that EEG-based classifiers could improve lower-limb neuroprostheses and neurorehabilitation techniques by providing earlier detection of movement intent, which could be used in robot-assisted strategies for motor training and recovery of function.

NI-55 * PARAMETRIC MAPS OF CONTRAST AGENT DIFFUSION USING A NOVEL DIFFUSION-PERFUSION MODEL APPLIED TO DCE-MRI CAN DIFFERENTIATE NON-ENHANCING TUMOR FROM EDEMA IN HUMAN GLIOBLASTOMA

INTRODUCTION: We propose a new method for estimating the diffusion of contrast agent in the perfusion scan, dynamic contrast enhanced (DCE)-MRI. We term the new parameteres “effective diffusion coefficients (Deff)”, due to the possibility of simultaneous diffusion and bulk flow in the brain. METHODS: Twenty-six glioblastoma patients were scanned after informed consent. All scans were acquired at the time of declared progression of the tumor. MRI was performed with a 3T Siemens scanner (Erlangen, Germany). A total of 130 images were acquired with five-second time resolution. The bolus of gadolinium-diethylenetriaminepentaacetate (Gd-DTPA, gadopenetic acid, Magnevist) was injected after 10-13 images. Diffusion was estimated via a perfusion-diffusion combination model. RESULTS: To establish whether the diffusion estimates provide new information or can simply be explained by perfusion, we first assess the correlation between diffusion and the traditional DCE parameter from the Tofts model, Ktrans. Within enhancing tumor lesions, where Ktrans is typically used to assess neovascularization, the average correlation is 0.28 ± 0.07 (mean ± standard error), which shows that diffusion estimates depend little on perfusion. We then compare interpatient variability of the same two parameters in normal appearing white matter, with Deff having mean to standard deviation ratio, a form of signal to noise ratio, of 1.55, while the same metric for Ktrans was 0.87. The mean of Deff in NAWM is 3.8x10−4 mm2/s and in tumor is 6.39x10−3 mm2/s. Finally, we determine thresholds that best separate tumor voxels from non-tumor voxels and assess classification accuracy over five trials with sensitivity/specificity of Deff being 91.6/75.3 and for Ktrans, 75.2/74.9, with sensitivity being significanly higher with Deff (p = 3.2x10−9). DISCUSSION: Our preliminary results suggest that Deff provides new information about glioblastomas not achievable by the Tofts model, has greater robustness across patients than Ktrans, and is more sensitive to tumors.

B-10 * CPT-II versus TOVA: Assessing the Diagnostic Power of Continuous Performance Tests

Objective: To determine if there is a significant difference in classification accuracy between Conners' Continuous Performance Test – II (CPT-II) and the Test of Variables of Attention (TOVA) for Attention Deficit Hyperactivity Disorder (ADHD). It was hypothesized the TOVA would better predict a diagnosis of ADHD, and that response time variability would have the strongest relationship with diagnosis. Method: Participants were 50 children referred for neuropsychological evaluation; 30% were female and 70% male. Mean age was 12.59 (SD = 2.87). Participants were administered the CPT-II and TOVA within the same neuropsychological evaluation. Results: The majority of the sample (86%) received a diagnosis of ADHD as a result of the evaluation. Receiver Operating Characteristic (ROC) was used to assess classification accuracy of various aspects of the two tests. Neither standard scores for commission errors nor omission errors of the CPT-II or the TOVA resulted in a ROC significantly above chance. ROCs for response time and response time variability were also near chance for the CPT-II but slightly better for the TOVA (AUCs of .571 and .629, respectively). The TOVA's API produced an ROC with better than chance classification (AUC = .618). When the CPT-II's confidence index was transformed into a continuous variable, the resulting ROC had an area under the curve of .687, which was the highest of any variable examined. Conclusion(s): Given the number of cognitive diagnoses which can be accompanied by attentional difficulties, the poor classification accuracies demonstrated here for ADHD are not unexpected, but the CPT-II's slightly better performance for total score was unanticipated.