Neuroergonomic and psychometric evaluation of full-face crew oxygen masks respiratory tolerance: a proof-of-concept study.

INTRODUCTION: Preventing in-flight hypoxia in pilots is typically achieved by wearing oxygen masks. These masks must be as comfortable as possible to allow prolonged and repeated use. The consequences of mask-induced facial contact pressure have been extensively studied, but little is known about mask-induced breathing discomfort. Because breathlessness is a strong distractor and engages cerebral resources, it could negatively impact flying performances. METHODS: Seventeen volunteers (age 20-32) rated respiratory discomfort while breathing with no mask and with two models of quick-donning full-face crew oxygen masks with regulators (mask A, mask B). Electroencephalographic recordings were performed to detect a putative respiratory-related cortical activation in response to inspiratory constraint (experiment 1, n=10). Oxygen consumption was measured using indirect calorimetry (experiment 2, n=10). RESULTS: With mask B, mild respiratory discomfort was reported significantly more frequently than with no mask or mask A (experiment 1: median respiratory discomfort on visual analogue scale 0.9 cm (0.5-1.4), experiment 1; experiment 2: 2 cm (1.7-2.9)). Respiratory-related cortical activation was present in 1/10 subjects with no mask, 1/10 with mask A and 6/10 with mask B (significantly more frequently with mask B). Breathing pattern, sigh frequency and oxygen consumption were not different. CONCLUSIONS: In a laboratory setting, breathing through high-end aeronautical full-face crew oxygen masks can induce mild breathing discomfort and activate respiratory-related cortical networks. Whether or not this can occur in real-life conditions and have operational consequences remains to be investigated. Meanwhile, respiratory psychometric and neuroergonomic approaches could be worth integrating to masks development and evaluation processes.

Modifications of the endosomal compartment in peripheral blood mononuclear cells and fibroblasts from Alzheimer's disease patients.

Identification of blood-based biomarkers of Alzheimer's disease (AD) remains a challenge. Neuropathological studies have identified enlarged endosomes in post-mortem brains as the earliest cellular change associated to AD. Here the presence of enlarged endosomes was investigated in peripheral blood mononuclear cells from 48 biologically defined AD patients (25 with mild cognitive impairment and 23 with dementia (AD-D)), and 23 age-matched healthy controls using immunocytochemistry and confocal microscopy. The volume and number of endosomes were not significantly different between AD and controls. However, the percentage of cells containing enlarged endosomes was significantly higher in the AD-D group as compared with controls. Furthermore, endosomal volumes significantly correlated to [C(11)]PiB cortical index measured by positron emission tomography in the AD group, independently of the APOE genotype, but not to the levels of amyloid-beta, tau and phosphorylated tau measured in the cerebrospinal fluid. Importantly, we confirmed the presence of enlarged endosomes in fibroblasts from six unrelated AD-D patients as compared with five cognitively normal controls. This study is the first, to our knowledge, to report morphological alterations of the endosomal compartment in peripheral cells from AD patients correlated to amyloid load that will now be evaluated as a possible biomarker.

Klf9 is necessary and sufficient for Purkinje cell survival in organotypic culture.

During their phase of developmental programmed cell death (PCD), neurons depend on target-released trophic factors for survival. After this period, however, they critically change as their survival becomes target-independent. The molecular mechanisms underlying this major transition remain poorly understood. Here, we investigated, which transcription factors (TFs) might be responsible for the closure of PCD. We used Purkinje cells as a model since their PCD is restricted to the first postnatal week in the mouse cerebellum. Transcriptome analysis of Purkinje cells during or after PCD allowed the identification of Krüppel like factor 9 (Klf9) as a candidate for PCD closure, given its high increase of expression at the end of the 1st postnatal week. Klf9 function was tested in organotypic cultures, through lentiviral vector-mediated manipulation of Klf9 expression. In absence of trophic factors, the Purkinje cell survival rate is of 40%. Overexpression of Klf9 during PCD dramatically increases the Purkinje cell survival rate from 40% to 88%, whereas its down-regulation decreases it to 14%. Accordingly, in organotypic cultures of Klf9 knockout animals, Purkinje cell survival rate is reduced by half as compared to wild-type mice. Furthermore, the absence of Klf9 could be rescued by Purkinje cell trophic factors, Insulin growth factor-1 and Neurotrophin3. Altogether, our results ascribe a clear role of Klf9 in Purkinje cell survival. Thus, we propose that Klf9 might be a key molecule involved in turning off the phase of Purkinje PCD.

Classification of human chromosome 21 gene-expression variations in Down syndrome: impact on disease phenotypes.

Down syndrome caused by chromosome 21 trisomy is the most common genetic cause of mental retardation in humans. Disruption of the phenotype is thought to be the result of gene-dosage imbalance. Variations in chromosome 21 gene expression in Down syndrome were analyzed in lymphoblastoid cells derived from patients and control individuals. Of the 359 genes and predictions displayed on a specifically designed high-content chromosome 21 microarray, one-third were expressed in lymphoblastoid cells. We performed a mixed-model analysis of variance to find genes that are differentially expressed in Down syndrome independent of sex and interindividual variations. In addition, we identified genes with variations between Down syndrome and control samples that were significantly different from the gene-dosage effect (1.5). Microarray data were validated by quantitative polymerase chain reaction. We found that 29% of the expressed chromosome 21 transcripts are overexpressed in Down syndrome and correspond to either genes or open reading frames. Among these, 22% are increased proportional to the gene-dosage effect, and 7% are amplified. The other 71% of expressed sequences are either compensated (56%, with a large proportion of predicted genes and antisense transcripts) or highly variable among individuals (15%). Thus, most of the chromosome 21 transcripts are compensated for the gene-dosage effect. Overexpressed genes are likely to be involved in the Down syndrome phenotype, in contrast to the compensated genes. Highly variable genes could account for phenotypic variations observed in patients. Finally, we show that alternative transcripts belonging to the same gene are similarly regulated in Down syndrome but sense and antisense transcripts are not.

Transcriptional disruptions in Down syndrome: a case study in the Ts1Cje mouse cerebellum during post-natal development.

To understand the aetiology and the phenotypic severity of Down syndrome, we searched for transcriptional signatures in a substructure of the brain (cerebellum) during post-natal development in a segmental trisomy 16 model, the Ts1Cje mouse. The goal of this study was to investigate the effects of trisomy on changes in gene expression across development time. The primary gene-dosage effect on triplicated genes (approximately 1.5) was observed at birth [post-natal day 0 (P0)], at P15 and P30. About 5% of the non-triplicated genes were significantly differentially expressed between trisomic and control cerebellum, while 25% of the transcriptome was modified during post-natal development of the cerebellum. Indeed, only 165, 171 and 115 genes were dysregulated in trisomic cerebellum at P0, P15 and P30, respectively. Surprisingly, there were only three genes dysregulated in development and in trisomic animals in a similar or opposite direction. These three genes (Dscr1, Son and Hmg14) were, quite unexpectedly, triplicated in the Ts1Cje model and should be candidate genes for understanding the aetiology of the phenotype observed in the cerebellum.

The cerebellar transcriptome during postnatal development of the Ts1Cje mouse, a segmental trisomy model for Down syndrome.

The central nervous system of persons with Down syndrome presents cytoarchitectural abnormalities that likely result from gene-dosage effects affecting the expression of key developmental genes. To test this hypothesis, we have investigated the transcriptome of the cerebellum of the Ts1Cje mouse model of Down syndrome during postnatal development using microarrays and quantitative PCR (qPCR). Genes present in three copies were consistently overexpressed, with a mean ratio relative to euploid of 1.52 as determined by qPCR. Out of 63 three-copy genes tested, only five, nine and seven genes had ratios >2 or <1.2 at postnatal days 0 (P0), P15 and P30, respectively. This gene-dosage effect was associated with a dysregulation of the expression of some two-copy genes. Out of 8258 genes examined, the Ts1Cje/euploid ratios differed significantly from 1.0 for 406 (80 and 154 with ratios above 1.5 and below 0.7, respectively), 333 (11 above 1.5 and 55 below 0.7) and 246 genes (59 above 1.5 and 69 below 0.7) at P0, P15 and P30, respectively. Among the two-copy genes differentially expressed in the trisomic cerebellum, six homeobox genes, two belonging to the Notch pathway, were severely repressed. Overall, at P0, transcripts involved in cell differentiation and development were over-represented among the dysregulated genes, suggesting that cell differentiation and migration might be more altered than cell proliferation. Finally, global gene profiling revealed that transcription in Ts1Cje mice is more affected by the developmental changes than by the trisomic state, and that there is no apparent detectable delay in the postnatal development of the cerebellum of Ts1Cje mice.

Neural-network construction and selection in nonlinear modeling.

We study how statistical tools which are commonly used independently can advantageously be exploited together in order to improve neural network estimation and selection in nonlinear static modeling. The tools we consider are the analysis of the numerical conditioning of the neural network candidates, statistical hypothesis tests, and cross validation. We present and analyze each of these tools in order to justify at what stage of a construction and selection procedure they can be most useful. On the basis of this analysis, we then propose a novel and systematic construction and selection procedure for neural modeling. We finally illustrate its efficiency through large-scale simulations experiments and real-world modeling problems.

No free lunch with the sandwich [sandwich estimator].

In nonlinear regression theory, the sandwich estimator of the covariance matrix of the model parameters is known as a consistent estimator, even when the parameterized model does not contain the regression. However, in the latter case, we emphasize the fact that the consistency of the sandwich holds only if the inputs of the training set are the values of independent identically distributed random variables. Thus, in the frequent practical modeling situation involving a training set whose inputs are deliberately chosen and imposed by the designer, we question the opportunity to use the sandwich estimator rather than the simple estimator based on the inverse squared Jacobian.

Construction of confidence intervals for neural networks based on least squares estimation.

We present the theoretical results about the construction of confidence intervals for a nonlinear regression based on least squares estimation and using the linear Taylor expansion of the nonlinear model output. We stress the assumptions on which these results are based, in order to derive an appropriate methodology for neural black-box modeling; the latter is then analyzed and illustrated on simulated and real processes. We show that the linear Taylor expansion of a nonlinear model output also gives a tool to detect the possible ill-conditioning of neural network candidates, and to estimate their performance. Finally, we show that the least squares and linear Taylor expansion based approach compares favorably with other analytic approaches, and that it is an efficient and economic alternative to the nonanalytic and computationally intensive bootstrap methods.

Nonlinear internal model control using neural networks: application to processes with delay and design issues.

We propose a design procedure of neural internal model control systems for stable processes with delay. We show that the design of such nonadaptive indirect control systems necessitates only the training of the inverse of the model deprived from its delay, and that the presence of the delay thus does not increase the order of the inverse. The controller is then obtained by cascading this inverse with a rallying model which imposes the regulation dynamic behavior and ensures the robustness of the stability. A change in the desired regulation dynamic behavior, or an improvement of the stability, can be obtained by simply tuning the rallying model, without retraining the whole model reference controller. The robustness properties of internal model control systems being obtained when the inverse is perfect, we detail the precautions which must be taken for the training of the inverse so that it is accurate in the whole space visited during operation with the process. In the same spirit, we make an emphasis on neural models affine in the control input, whose perfect inverse is derived without training. The control of simulated processes illustrates the proposed design procedure and the properties of the neural internal model control system for processes without and with delay.

On cross validation for model selection.

In response to Zhu and Rower (1996), a recent communication (Goutte, 1997) established that leave-one-out cross validation is not subject to the "no-free-lunch" criticism. Despite this optimistic conclusion, we show here that cross validation has very poor performances for the selection of linear models as compared to classic statistical tests. We conclude that the statistical tests are preferable to cross validation for linear as well as for nonlinear model selection.