Maha Khachab

Brain Imaging and Machine Learning for Brain-Computer Interface

Human-computer interfaces are in continuous development, from keyboard, mouse, touch screen, to voice dictation, gesture recognition, etc. The aim is to facilitate the interaction between the human brain and the resources offered by a machine or a computer. Recently, a wider interest has emerged in directly interfacing the brain and the computer. The development of methods that combine the nervous system with artificial devices is attracting a growing interest from clinical research, because the interaction between brain and machines may lead to novel prosthetic devices or to a more efficient use of computer resources by breaking the barriers imposed at present by the classical human-machine interfaces. Individuals with impaired motor control may be disabled in the performance of their daily activities. Their motor performance, however, can be supported by artificial motor control systems. Such motor support systems may also assist healthy individuals in performing their tasks. One can also imagine interacting with different systems in parallel, or developing newer software tools without the need to physically typing the code.

Feature selection for an improved Parkinson's disease identification based on handwriting

Parkinsons disease (PD) is a neurological disorder associated with a progressive decline in motor skills, speech, and cognitive processes. Since the diagnosis of Parkinsons disease is difficult, researchers have worked to develop a support tool based on algorithms to separate healthy controls from PD patients. Online handwriting analysis is one of the methods that can be used to diagnose PD. The purpose of this study is to find a subset of handwriting features suitable for efficiently identifying subjects with PD. Data was taken from PDMultiMC database collected in Lebanon, and consisting of 16 medicated PD patients and 16 age matched controls. Seven handwriting tasks were collected such as copying patterns, copying words in Arabic, and writing full names. For each task kinematic and spatio-temporal, pressure, energy, entropy, and intrinsic features were extracted. Feature selection was done in two stages; the first stage selected a subset using statistical analysis, and the second step selected the most relevant features of this subset by a suboptimal approach. The selected features were fed to a support vector machine classifier with RBF kernel, whose aim is to identify the subjects suffering from PD. The accuracy of the classification of PD was as high as 96.875%, with sensitivity and specificity equal to 93.75 % and 100% respectively. The results as well as the selected features suggest that handwriting can be a valuable marker as a PD diagnosis tool.

Classification of Contrast Ultrasound Images using Autoregressive Model Coupled to Gaussian Mixture Model

Contrast ultrasound images are not clear enough to be directly adopted in the diagnostic. In fact, the ultrasound agents enhance the vascular zones but unfortunately the signals backscattered from agent and tissues are still close. Therefore, it is necessary to implement image-processing techniques to enhance the contrast echo and thus have the capability of classification. In this article, we apply a new approach based on the autoregressive model coupled to the Gaussian mixture model to represent both agent and tissue behaviors. Then, we process the resultant image by a classification method based on a fixed windows size in order to obtain a satisfying differentiation of the ultrasound image into two classes. Finally, we adopt the agent to tissue ratio (ATR) factor and the Fisher criterion to compare the performance of this method with existing techniques as harmonic and B mode.

Autoregressive modeling application for vascular zones detection in the contrast echographic images

Contrast agents are used in ultrasound imaging to enhance blood region and thereby separate the perfused area and the surrounding tissues. But unfortunately the signals backscattered from agent and tissues are still close. So it is necessary to implement signal processing to enhance the contrast echo. In this article, we apply the autoregressive model to exploit the nonlinear behavior agent properties. Then, we process the obtained pictures by a classification method followed by erosion dilatation algorithm to obtain a satisfying differentiation of the ultrasound image into two classes. The Agent to Tissue Ratio (ATR) factor is used to compare the performance of the methods, and the Fisher criterion is used to study the classification feasibility.

Colectomy induces an aldosterone-mediated increase in jejunal glucose uptake in rats

Aims: The main function of the colon is water and electrolyte absorption. Total colectomy eliminates this colonic function and may alter the absorptive capacity of the small intestine for nutrients. This study examines the effect of total colectomy on jejunal glucose absorption and investigates the potential role of aldosterone in mediating the alterations in glucose uptake post‐colectomy using the aldosterone antagonist spironolactone. Main methods: Total colectomy with ileo‐rectal anastomosis was performed on anesthetized rats. Sham rats were identically handled without colon resection. Two days post‐surgery, groups of colectomized rats were injected with either a daily subcutaneous dose of spironolactone or sesame oil for 12 days. Body weight changes and food and water intake were measured in all experimental groups. Glucose absorption was measured by in‐vivo single pass perfusion in the rat jejunum of control, sham, colectomized, colectomized with spironolactone, and colectomized with sesame oil treatment. Na/K ATPase, SGK1, SGLT1 and GLUT2 expressions were determined in jejunal mucosa in control, colectomized and colectomized/spironolactone injected rats by Western blot analysis. Histological assessment was performed on jejunal sections in control and colectomized groups. Key findings: Glucose absorption significantly increased in colectomized rats with an observed increase in Na/K ATPase and SGK1 expression. No significant expression change in SGLT1 and GLUT2 was detected in the jejunum in colectomized rats. Spironolactone, however, significantly decreased the glucose uptake post‐colectomy and normalized Na/K ATPase and SGK1 expression. Significance: Our results suggest that jejunal glucose uptake increases post‐colectomy as a possible consequence of an aldosterone‐mediated function.

Improvement of the GMM‐AR classification of multiframe contrast ultrasound images using gaussian filter

Despite the use of contrast agents that enhance the visualization of vascular zones, the backscattered signals from the contrast agent and tissue are still close which prevents the direct wide ultrasonic use in diagnosis. Thus, it was necessary to implement image‐processing techniques that enhance the contrast echo and have the capability of classification. We have applied a new approach based on the autoregressive model where an image of prediction errors is calculated in the first phase. Then, a Gaussian filter is applied in order to model well afterward both agent and tissue behaviors by a Gaussian mixture model. The Agent to Tissue Ratio (ATR)factor and Fisher criterion are adopted to compare the performance of this method with existing techniques as the harmonic and B mode techniques. The experiments conducted have shown the advantages of our proposed approach where an increasing of ATR and Fisher are recorded. In fact, our ATR attains 19.20 dB which represents a good improvement in comparison with B...