File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Raman spectroscopy detection of platelet for Alzheimer's disease with predictive probabilities

TitleRaman spectroscopy detection of platelet for Alzheimer's disease with predictive probabilities
Authors
Keywordspredictive probability
Alzheimers disease
Raman spectroscopy
Issue Date2014
Citation
Laser Physics, 2014, v. 24, n. 8, article no. 085702 How to Cite?
AbstractAlzheimer's disease (AD) is a common form of dementia. Early and differential diagnosis of AD has always been an arduous task for the medical expert due to the unapparent early symptoms and the currently imperfect imaging examination methods. Therefore, obtaining reliable markers with clinical diagnostic value in easily assembled samples is worthy and significant. Our previous work with laser Raman spectroscopy (LRS), in which we detected platelet samples of different ages of AD transgenic mice and non-transgenic controls, showed great effect in the diagnosis of AD. In addition, a multilayer perception network (MLP) classification method was adopted to discriminate the spectral data. However, there were disturbances, which were induced by noise from the machines and so on, in the data set; thus the MLP method had to be trained with large-scale data. In this paper, we aim to re-establish the classification models of early and advanced AD and the control group with fewer features, and apply some mechanism of noise reduction to improve the accuracy of models. An adaptive classification method based on the Gaussian process (GP) featured, with predictive probabilities, is proposed, which could tell when a data set is related to some kind of disease. Compared with MLP on the same feature set, GP showed much better performance in the experimental results. What is more, since the spectra of platelets are isolated from AD, GP has good expansibility and can be applied in diagnosis of many other similar diseases, such as Parkinson's disease (PD). Spectral data of 4 month and 12 month AD platelets, as well as control data, were collected. With predictive probabilities, the proposed GP classification method improved the diagnostic sensitivity to nearly 100%. Samples were also collected from PD platelets as classification and comparison to the 12 month AD. The presented approach and our experiments indicate that utilization of GP with predictive probabilities in platelet LRS detection analysis turns out to be more accurate for early and differential diagnosis of AD and has a wide application prospect. © 2014 Astro Ltd.
Persistent Identifierhttp://hdl.handle.net/10722/296099
ISSN
2023 Impact Factor: 1.2
2023 SCImago Journal Rankings: 0.291
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorWang, L. J.-
dc.contributor.authorDu, X. Q.-
dc.contributor.authorDu, Z. W.-
dc.contributor.authorYang, Y. Y.-
dc.contributor.authorChen, P.-
dc.contributor.authorTian, Q.-
dc.contributor.authorShang, X. L.-
dc.contributor.authorLiu, Z. C.-
dc.contributor.authorYao, X. Q.-
dc.contributor.authorWang, J. Z.-
dc.contributor.authorWang, X. H.-
dc.contributor.authorCheng, Y.-
dc.contributor.authorPeng, J.-
dc.contributor.authorShen, A. G.-
dc.contributor.authorHu, J. M.-
dc.date.accessioned2021-02-11T04:52:50Z-
dc.date.available2021-02-11T04:52:50Z-
dc.date.issued2014-
dc.identifier.citationLaser Physics, 2014, v. 24, n. 8, article no. 085702-
dc.identifier.issn1054-660X-
dc.identifier.urihttp://hdl.handle.net/10722/296099-
dc.description.abstractAlzheimer's disease (AD) is a common form of dementia. Early and differential diagnosis of AD has always been an arduous task for the medical expert due to the unapparent early symptoms and the currently imperfect imaging examination methods. Therefore, obtaining reliable markers with clinical diagnostic value in easily assembled samples is worthy and significant. Our previous work with laser Raman spectroscopy (LRS), in which we detected platelet samples of different ages of AD transgenic mice and non-transgenic controls, showed great effect in the diagnosis of AD. In addition, a multilayer perception network (MLP) classification method was adopted to discriminate the spectral data. However, there were disturbances, which were induced by noise from the machines and so on, in the data set; thus the MLP method had to be trained with large-scale data. In this paper, we aim to re-establish the classification models of early and advanced AD and the control group with fewer features, and apply some mechanism of noise reduction to improve the accuracy of models. An adaptive classification method based on the Gaussian process (GP) featured, with predictive probabilities, is proposed, which could tell when a data set is related to some kind of disease. Compared with MLP on the same feature set, GP showed much better performance in the experimental results. What is more, since the spectra of platelets are isolated from AD, GP has good expansibility and can be applied in diagnosis of many other similar diseases, such as Parkinson's disease (PD). Spectral data of 4 month and 12 month AD platelets, as well as control data, were collected. With predictive probabilities, the proposed GP classification method improved the diagnostic sensitivity to nearly 100%. Samples were also collected from PD platelets as classification and comparison to the 12 month AD. The presented approach and our experiments indicate that utilization of GP with predictive probabilities in platelet LRS detection analysis turns out to be more accurate for early and differential diagnosis of AD and has a wide application prospect. © 2014 Astro Ltd.-
dc.languageeng-
dc.relation.ispartofLaser Physics-
dc.subjectpredictive probability-
dc.subjectAlzheimers disease-
dc.subjectRaman spectroscopy-
dc.titleRaman spectroscopy detection of platelet for Alzheimer's disease with predictive probabilities-
dc.typeArticle-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1088/1054-660X/24/8/085702-
dc.identifier.scopuseid_2-s2.0-84904985784-
dc.identifier.volume24-
dc.identifier.issue8-
dc.identifier.spagearticle no. 085702-
dc.identifier.epagearticle no. 085702-
dc.identifier.eissn1555-6611-
dc.identifier.isiWOS:000339322800041-
dc.identifier.issnl1054-660X-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats