Multilayer perceptron with local constraint as an emerging method in spatial data analysis

The use of Geographic Information Systems has revolutionalized the handling and the visualization of geo-referenced data and has underlined the critic role of spatial analysis. The usual tools for such a purpose are geostatistics which are widely used in Earth science. Geostatistics are based upon several hypothesis which are not always verified in practice. On the other hand, Artificial Neural Network (ANN) a priori can be used without special assumptions and are known to be flexible. This paper proposes to discuss the application of ANN in the case of the interpolation of a geo-referenced variable.

Central and Cortical Gray Mater Segmentation of Magnetic Resonance Images of the Fetal Brain

Motivation. The study of human brain development in itsearly stage is today possible thanks to in vivo fetalmagnetic resonance imaging (MRI) techniques. Aquantitative analysis of fetal cortical surfacerepresents a new approach which can be used as a markerof the cerebral maturation (as gyration) and also forstudying central nervous system pathologies [1]. However,this quantitative approach is a major challenge forseveral reasons. First, movement of the fetus inside theamniotic cavity requires very fast MRI sequences tominimize motion artifacts, resulting in a poor spatialresolution and/or lower SNR. Second, due to the ongoingmyelination and cortical maturation, the appearance ofthe developing brain differs very much from thehomogenous tissue types found in adults. Third, due tolow resolution, fetal MR images considerably suffer ofpartial volume (PV) effect, sometimes in large areas.Today extensive efforts are made to deal with thereconstruction of high resolution 3D fetal volumes[2,3,4] to cope with intra-volume motion and low SNR.However, few studies exist related to the automatedsegmentation of MR fetal imaging. [5] and [6] work on thesegmentation of specific areas of the fetal brain such asposterior fossa, brainstem or germinal matrix. Firstattempt for automated brain tissue segmentation has beenpresented in [7] and in our previous work [8]. Bothmethods apply the Expectation-Maximization Markov RandomField (EM-MRF) framework but contrary to [7] we do notneed from any anatomical atlas prior. Data set &Methods. Prenatal MR imaging was performed with a 1-Tsystem (GE Medical Systems, Milwaukee) using single shotfast spin echo (ssFSE) sequences (TR 7000 ms, TE 180 ms,FOV 40 x 40 cm, slice thickness 5.4mm, in plane spatialresolution 1.09mm). Each fetus has 6 axial volumes(around 15 slices per volume), each of them acquired inabout 1 min. Each volume is shifted by 1 mm with respectto the previous one. Gestational age (GA) ranges from 29to 32 weeks. Mother is under sedation. Each volume ismanually segmented to extract fetal brain fromsurrounding maternal tissues. Then, in-homogeneityintensity correction is performed using [9] and linearintensity normalization is performed to have intensityvalues that range from 0 to 255. Note that due tointra-tissue variability of developing brain someintensity variability still remains. For each fetus, ahigh spatial resolution image of isotropic voxel size of1.09 mm is created applying [2] and using B-splines forthe scattered data interpolation [10] (see Fig. 1). Then,basal ganglia (BS) segmentation is performed on thissuper reconstructed volume. Active contour framework witha Level Set (LS) implementation is used. Our LS follows aslightly different formulation from well-known Chan-Vese[11] formulation. In our case, the LS evolves forcing themean of the inside of the curve to be the mean intensityof basal ganglia. Moreover, we add local spatial priorthrough a probabilistic map created by fitting anellipsoid onto the basal ganglia region. Some userinteraction is needed to set the mean intensity of BG(green dots in Fig. 2) and the initial fitting points forthe probabilistic prior map (blue points in Fig. 2). Oncebasal ganglia are removed from the image, brain tissuesegmentation is performed as described in [8]. Results.The case study presented here has 29 weeks of GA. Thehigh resolution reconstructed volume is presented in Fig.1. The steps of BG segmentation are shown in Fig. 2.Overlap in comparison with manual segmentation isquantified by the Dice similarity index (DSI) equal to0.829 (values above 0.7 are considered a very goodagreement). Such BG segmentation has been applied on 3other subjects ranging for 29 to 32 GA and the DSI hasbeen of 0.856, 0.794 and 0.785. Our segmentation of theinner (red and blue contours) and outer cortical surface(green contour) is presented in Fig. 3. Finally, torefine the results we include our WM segmentation in theFreesurfer software [12] and some manual corrections toobtain Fig.4. Discussion. Precise cortical surfaceextraction of fetal brain is needed for quantitativestudies of early human brain development. Our workcombines the well known statistical classificationframework with the active contour segmentation forcentral gray mater extraction. A main advantage of thepresented procedure for fetal brain surface extraction isthat we do not include any spatial prior coming fromanatomical atlases. The results presented here arepreliminary but promising. Our efforts are now in testingsuch approach on a wider range of gestational ages thatwe will include in the final version of this work andstudying as well its generalization to different scannersand different type of MRI sequences. References. [1]Guibaud, Prenatal Diagnosis 29(4) (2009). [2] Rousseau,Acad. Rad. 13(9), 2006, [3] Jiang, IEEE TMI 2007. [4]Warfield IADB, MICCAI 2009. [5] Claude, IEEE Trans. Bio.Eng. 51(4) (2004). [6] Habas, MICCAI (Pt. 1) 2008. [7]Bertelsen, ISMRM 2009 [8] Bach Cuadra, IADB, MICCAI 2009.[9] Styner, IEEE TMI 19(39 (2000). [10] Lee, IEEE Trans.Visual. And Comp. Graph. 3(3), 1997, [11] Chan, IEEETrans. Img. Proc, 10(2), 2001 [12] Freesurfer,http://surfer.nmr.mgh.harvard.edu.

Artificial Snow Optimization in Winter Sport Destinations Using a Multi-agent Simulation

This paper presents the Juste-Neige system for predicting the snow height on the ski runs of a resort using a multi-agent simulation software. Its aim is to facilitate snow cover management in order to i) reduce the production cost of artificial snow and to improve the profit margin for the companies managing the ski resorts; and ii) to reduce the water and energy consumption, and thus to reduce the environmental impact, by producing only the snow needed for a good skiing experience. The software provides maps with the predicted snow heights for up to 13 days. On these maps, the areas most exposed to snow erosion are highlighted. The software proceeds in three steps: i) interpolation of snow height measurements with a neural network; ii) local meteorological forecasts for every ski resort; iii) simulation of the impact caused by skiers using a multi-agent system. The software has been evaluated in the Swiss ski resort of Verbier and provides useful predictions.

Is PaCO2 management optimal under controlled mechanical ventilation (CMV) during neuro-resuscitation?

INTRODUCTION. Both hypocapnia and hypercapnia can be deleterious to brain injured patients. Strict PaCO2 control is difficult to achieve because of patient's instability and unpredictable effects of ventilator settings changes. OBJECTIVE. The aim of this study was to evaluate our ability to comply with a protocol of controlled mechanical ventilation (CMV) aiming at a PaCO2 between 35 and 40 mmHg in patients requiring neuro-resuscitation. METHODS. Retrospective analysis of consecutive patients (2005-2011) requiring intracranial pressure (ICP) monitoring for traumatic brain injury (TBI), subarachnoid haemorrhage (SAH), intracranial haemorrhage (ICH) or ischemic stroke (IS). Demographic data, GCS, SAPS II, hospital mortality, PaCO2 and ICP values were recorded. During CMV in the first 48 h after admission, we analyzed the time spent within the PaCO2 target in relation to the presence or absence of intracranial hypertension (ICP[20 mmHg, by periods of 30 min) (Table 1). We also compared the fraction of time (determined by linear interpolation) spent with normal, low or high PaCO2 in hospital survivors and non-survivors (Wilcoxon, Bonferroni correction, p\0.05) (Table 2). PaCO2 samples collected during and after apnoea tests were excluded. Results given as median [IQR]. RESULTS. 436 patients were included (TBI: 51.2 %, SAH: 20.6 %, ICH: 23.2 %, IS: 5.0 %), age: 54 [39-64], SAPS II score: 52 [41-62], GCS: 5 [3-8]. 8744 PaCO2 samples were collected during 150611 h of CMV. CONCLUSIONS. Despite a high number of PaCO2 samples collected (in average one sample every 107 min), our results show that patients undergoing CMV for neuro- resuscitation spent less than half of the time within the pre-defined PaCO2 range. During documented intracranial hypertension, hypercapnia was observed in 17.4 % of the time. Since non-survivors spent more time with hypocapnia, further analysis is required to determine whether hypocapnia was detrimental per se, or merely reflects increased severity of brain insult.

Alteration in neuromuscular function after a 5 km running time trial.

The aim of this study was to characterize the effect of a 5 km running time trial on the neuromuscular properties of the plantar flexors. Eleven well-trained triathletes performed a series of neuromuscular tests before and immediately after the run on a 200 m indoor track. Muscle activation (twitch interpolation) and normalized EMG activity were assessed during maximal voluntary contraction (MVC) of plantar flexors. Maximal soleus H-reflexes and M-waves were evoked at rest (i.e. H (MAX) and M (MAX), respectively) and during MVC (i.e. H (SUP) and M (SUP), respectively). MVC significantly declined (-27%; P < 0.001) after the run, due to decrease in muscle activation (-8%; P < 0.05) and M (MAX)-normalized EMG activity (-13%; P < 0.05). Significant reductions in M-wave amplitudes (M (MAX): -13% and M (SUP): -16%; P < 0.05) as well as H (MAX)/M (MAX) (-37%; P < 0.01) and H (SUP)/M (SUP) (-25%; P < 0.05) ratios occurred with fatigue. Following exercise, the single twitch was characterized by lower peak torque (-16%; P < 0.001) as well as shorter contraction (-19%; P < 0.001) and half-relaxation (-24%; P < 0.001) times. In conclusion, the reduction in plantar flexors strength induced by a 5 km running time trial is caused by peripheral adjustments, which are attributable to a failure of the neuromuscular transmission and excitation-contraction coupling. Fatigue also decreased the magnitude of efferent motor outflow from spinal motor neurons to the plantar flexors and part of this suboptimal neural drive is the result of an inhibition of soleus motoneuron pool reflex excitability.

Impact of reduced cerebral perfusion pressure on outcome after severe traumatic brain injury is dependent on brain tissue oxygen pressure

INTRODUCTION. Reduced cerebral perfusion pressure (CPP) may worsen secondary damage and outcome after severe traumatic brain injury (TBI), however the optimal management of CPP is still debated. STUDY HYPOTHESIS: We hypothesized that the impact of CPP on outcome is related to brain tissue oxygen tension (PbtO2) level and that reduced CPP may worsen TBI prognosis when it is associated with brain hypoxia. DESIGN. Retrospective analysis of prospective database. METHODS. We analyzed 103 patients with severe TBI who underwent continuous PbtO2 and CPP monitoring for an average of 5 days. For each patient, duration of reduced CPP (\60 mm Hg) and brain hypoxia (PbtO2\15 mm Hg for[30 min [1]) was calculated with linear interpolation method and the relationship between CPP and PbtO2 was analyzed with Pearson's linear correlation coefficient. Outcome at 30 days was assessed with the Glasgow Outcome Score (GOS), dichotomized as good (GOS 4-5) versus poor (GOS 1-3). Multivariable associations with outcome were analyzed with stepwise forward logistic regression. RESULTS. Reduced CPP (n=790 episodes; mean duration 10.2 ± 12.3 h) was observed in 75 (74%) patients and was frequently associated with brain hypoxia (46/75; 61%). Episodes where reduced CPP were associated with normal brain oxygen did not differ significantly between patients with poor versus those with good outcome (8.2 ± 8.3 vs. 6.5 ± 9.7 h; P=0.35). In contrast, time where reduced CPP occurred simultaneously with brain hypoxia was longer in patients with poor than in those with good outcome (3.3±7.4 vs. 0.8±2.3 h; P=0.02). Outcome was significantly worse in patients who had both reduced CPP and brain hypoxia (61% had GOS 1-3 vs. 17% in those with reduced CPP but no brain hypoxia; P\0.01). Patients in whom a positive CPP-PbtO2 correlation (r[0.3) was found also were more likely to have poor outcome (69 vs. 31% in patients with no CPP-PbtO2 correlation; P\0.01). Brain hypoxia was an independent risk factor of poor prognosis (odds ratio for favorable outcome of 0.89 [95% CI 0.79-1.00] per hour spent with a PbtO2\15 mm Hg; P=0.05, adjusted for CPP, age, GCS, Marshall CT and APACHE II). CONCLUSIONS. Low CPP may significantly worsen outcome after severe TBI when it is associated with brain tissue hypoxia. PbtO2-targeted management of CPP may optimize TBI therapy and improve outcome of head-injured patients.

Spinal modulations accompany peripheral fatigue during prolonged tennis playing.

To examine the time course of alteration in neural process (spinal loop properties) during prolonged tennis playing, 12 competitive players performed a series of neuromuscular tests every 30 min during a 3-h match protocol. Muscle activation (twitch interpolation) and normalized EMG activity were assessed during maximal voluntary contraction (MVC) of plantar flexors. Spinal reflexes and M-waves were evoked at rest (i.e., H(max) and M(max) , respectively) and during MVC (i.e., H(sup) , V-wave, M(sup) , respectively). MVC torque declined significantly (P<0.001) across the match protocol, due to decrease (P<0.001) in muscle activation and in normalized EMG activity. The impairment in MVC was significantly correlated (r=0.77; P<0.05) with the decline in muscle activation. H(max) /M(max) (P<0.001), H(sup) /M(sup) (P<0.01) and V/M(sup) (P<0.05) ratios were depressed with fatigue and decreased by ∼80%, 46% and 61% at the end of exercise, respectively. Simultaneously, peak twitch torque and M-wave amplitude were significantly (P<0.01) altered with exercise, suggesting peripheral alterations. During prolonged tennis playing, the compromised voluntary strength capacity is linked to a reduced neural input to the working muscles. This central activation deficit partly results from a modulation in spinal loop properties.

Alteration in neuromuscular function after a 5 km running time trial

The aim of this study was to characterize the effect of a 5 km running time trial on the neuromuscular properties of the plantar flexors. Eleven well-trained triathletes performed a series of neuromuscular tests before and immediately after the run on a 200 m indoor track. Muscle activation (twitch interpolation) and normalized EMG activity were assessed during maximal voluntary contraction (MVC) of plantar flexors. Maximal soleus H-reflexes and M-waves were evoked at rest (i.e. H (MAX) and M (MAX), respectively) and during MVC (i.e. H (SUP) and M (SUP), respectively). MVC significantly declined (-27%; P < 0.001) after the run, due to decrease in muscle activation (-8%; P < 0.05) and M (MAX)-normalized EMG activity (-13%; P < 0.05). Significant reductions in M-wave amplitudes (M (MAX): -13% and M (SUP): -16%; P < 0.05) as well as H (MAX)/M (MAX) (-37%; P < 0.01) and H (SUP)/M (SUP) (-25%; P < 0.05) ratios occurred with fatigue. Following exercise, the single twitch was characterized by lower peak torque (-16%; P < 0.001) as well as shorter contraction (-19%; P < 0.001) and half-relaxation (-24%; P < 0.001) times. In conclusion, the reduction in plantar flexors strength induced by a 5 km running time trial is caused by peripheral adjustments, which are attributable to a failure of the neuromuscular transmission and excitation-contraction coupling. Fatigue also decreased the magnitude of efferent motor outflow from spinal motor neurons to the plantar flexors and part of this suboptimal neural drive is the result of an inhibition of soleus motoneuron pool reflex excitability.

Carotid Dose Spring in Definitive Irradiation of T1N0 Squamous Cell Laryngeal Carcinoma Using Helical Tomotherapy

Purpose/Objective(s): To implement a carotid dose sparing protocol using helical Tomotherapy in T1N0 squamous cell laryngeal carcinoma.Materials/Methods: Between July and August 2010, 7 men with stage T1N0 laryngeal carcinoma were included in this study. Age ranged from 47 - 74 years. Staging included endoscopic examination, CT-scan and MRI when indicated. Planned irradiation dose was 70 Gy in 35 fractions over 7 weeks. A simple treatment planning algorithm for carotid sparing was used: maximum point dose to the carotids 35 Gy, to the spinal cord 30 Gy, and 100% PTV volume to be covered with 95% of the prescribed dose. Carotid volume of interest extended to 1 cm above and below of the PTV. Doses to the carotid arteries, to the critical organs, and to the planned target volume (PTV) with our standard laryngeal irradiation protocol was compared. Daily megavoltage scans were obtained before each fraction. When necessary, the Planned Adaptive software (TomoTherapy Inc., Madison, WI) was used to evaluatethe need for a re-planning, which has never been indicated. Dose data were extracted using the VelocityAI software (Atlanta, GA), and data normalization and dose-volume histogram (DVH) interpolation were realized using the Igor Pro software (Portland, OR).Results:A significant (p\0.05) carotid dose sparing compared to our standard protocol with an average maximum point dose of 38.3 Gy (standard deviation [SD] 4.05 Gy), average mean dose of 18.59 Gy (SD 0.83 Gy) was achieved. In all patients, 95% of the carotid volume received less than 28.4 Gy (SD 0.98 Gy). The average maximum point dose to the spinal cord was 25.8 Gy (SD 3.24 Gy). PTV was fully covered with more than 95% of the prescribed dose for all patients with an average maximum point dose of 74.1 Gy and the absolute maximum dose in a single patient of 75.2 Gy. To date, the clinical outcomes have been excellent. Three patients (42%) developed stage 1 mucositis that was conservatively managed, and all the patients presented a mild to moderate dysphonia. All adverse effects resolved spontaneously in the month following the end of treatment. Early local control rate is 100% considering a 4 - 5 months post treatment follow-up.Conclusions: Helical Tomotherapy allows a clinically significant decrease of carotid irradiation dose compared to standard irradiation protocols with an acceptable spinal cord dose tradeoff. Moreover, this technique allows the PTV to be homogenously covered with a curative irradiation dose. Daily control imaging brings added security margins especially when working with high dose gradients. Further investigations and follow-up are underway to better evaluate the late clinical outcomes especially the local control rate, late laryngeal and vascular toxicity, and expected potential impact on cerebrovascular events.

Wavelet analysis residual kriging vs. neural network residual kriging

This paper deals with the problem of spatial data mapping. A new method based on wavelet interpolation and geostatistical prediction (kriging) is proposed. The method - wavelet analysis residual kriging (WARK) - is developed in order to assess the problems rising for highly variable data in presence of spatial trends. In these cases stationary prediction models have very limited application. Wavelet analysis is used to model large-scale structures and kriging of the remaining residuals focuses on small-scale peculiarities. WARK is able to model spatial pattern which features multiscale structure. In the present work WARK is applied to the rainfall data and the results of validation are compared with the ones obtained from neural network residual kriging (NNRK). NNRK is also a residual-based method, which uses artificial neural network to model large-scale non-linear trends. The comparison of the results demonstrates the high quality performance of WARK in predicting hot spots, reproducing global statistical characteristics of the distribution and spatial correlation structure.

Combining heat stress and moderate hypoxia reduces cycling time to exhaustion without modifying neuromuscular fatigue characteristics.

PURPOSE: This study investigated the isolated and combined effects of heat [temperate (22 °C/30 % rH) vs. hot (35 °C/40 % rH)] and hypoxia [sea level (FiO2 0.21) vs. moderate altitude (FiO2 0.15)] on exercise capacity and neuromuscular fatigue characteristics. METHODS: Eleven physically active subjects cycled to exhaustion at constant workload (66 % of the power output associated with their maximal oxygen uptake in temperate conditions) in four different environmental conditions [temperate/sea level (control), hot/sea level (hot), temperate/moderate altitude (hypoxia) and hot/moderate altitude (hot + hypoxia)]. Torque and electromyography (EMG) responses following electrical stimulation of the tibial nerve (plantar-flexion; soleus) were recorded before and 5 min after exercise. RESULTS: Time to exhaustion was reduced (P < 0.05) in hot (-35 ± 15 %) or hypoxia (-36 ± 14 %) compared to control (61 ± 28 min), while hot + hypoxia (-51 ± 20 %) further compromised exercise capacity (P < 0.05). However, the effect of temperature or altitude on end-exercise core temperature (P = 0.089 and P = 0.070, respectively) and rating of perceived exertion (P > 0.05) did not reach significance. Maximal voluntary contraction torque, voluntary activation (twitch interpolation) and peak twitch torque decreased from pre- to post-exercise (-9 ± 1, -4 ± 1 and -6 ± 1 % all trials compounded, respectively; P < 0.05), with no effect of the temperature or altitude. M-wave amplitude and root mean square activity were reduced (P < 0.05) in hot compared to temperate conditions, while normalized maximal EMG activity did not change. Altitude had no effect on any measured parameters. CONCLUSION: Moderate hypoxia in combination with heat stress reduces cycling time to exhaustion without modifying neuromuscular fatigue characteristics. Impaired oxygen delivery or increased cardiovascular strain, increasing relative exercise intensity, may have also contributed to earlier exercise cessation.

Direct Fourier Tomographic Reconstruction Image-To-Image Filter

We present an open-source ITK implementation of a directFourier method for tomographic reconstruction, applicableto parallel-beam x-ray images. Direct Fourierreconstruction makes use of the central-slice theorem tobuild a polar 2D Fourier space from the 1D transformedprojections of the scanned object, that is resampled intoa Cartesian grid. Inverse 2D Fourier transform eventuallyyields the reconstructed image. Additionally, we providea complex wrapper to the BSplineInterpolateImageFunctionto overcome ITKâeuro?s current lack for image interpolatorsdealing with complex data types. A sample application ispresented and extensively illustrated on the Shepp-Loganhead phantom. We show that appropriate input zeropaddingand 2D-DFT oversampling rates together with radial cubicb-spline interpolation improve 2D-DFT interpolationquality and are efficient remedies to reducereconstruction artifacts.