PubMed Health⌕ Search

Biomedical subjects

Ira Cohen

Publications and source records attributed to Ira Cohen.

2 recordsLinked to original sources

Oxidative stress in autism: increased lipid peroxidation and reduced serum levels of ceruloplasmin and transferrin--the antioxidant proteins.

Autism is a neurological disorder of childhood with poorly understood etiology and pathology. We compared lipid peroxidation status in the plasma of children with autism, and their developmentally normal non-autistic siblings by quantifying the levels of malonyldialdehyde, an end product of fatty acid oxidation. Lipid peroxidation was found to be elevated in autism indicating that oxidative stress is increased in this disease. Levels of major antioxidant proteins namely, transferrin (iron-binding protein) and ceruloplasmin (copper-binding protein) in the serum, were significantly reduced in autistic children as compared to their developmentally normal non-autistic siblings. A striking correlation was observed between reduced levels of these proteins and loss of previously acquired language skills in children with autism. These results indicate altered regulation of transferrin and ceruloplasmin in autistic children who lose acquired language skills. It is suggested that such changes may lead to abnormal iron and copper metabolism in autism, and that increased oxidative stress may have pathological role in autism.

Autistic Disorder↗

Semisupervised learning of classifiers: theory, algorithms, and their application to human-computer interaction.

Automatic classification is one of the basic tasks required in any pattern recognition and human computer interaction application. In this paper, we discuss training probabilistic classifiers with labeled and unlabeled data. We provide a new analysis that shows under what conditions unlabeled data can be used in learning to improve classification performance. We also show that, if the conditions are violated, using unlabeled data can be detrimental to classification performance. We discuss the implications of this analysis to a specific type of probabilistic classifiers, Bayesian networks, and propose a new structure learning algorithm that can utilize unlabeled data to improve classification. Finally, we show how the resulting algorithms are successfully employed in two applications related to human-computer interaction and pattern recognition: facial expression recognition and face detection.

Journal Article↗