Recognition of EEG Signals from Imagined Vowels Using Deep Learning Methods.

The use of imagined speech with electroencephalographic (EEG) signals is a promising field of brain-computer interfaces (BCI) that seeks communication between areas of the cerebral cortex related to language and devices or machines. However, the complexity of this brain process makes the analysis and classification of this type of signals a relevant topic of research. The goals of this study were: to develop a new algorithm based on Deep Learning (DL), referred to as CNNeeg1-1, to recognize EEG signals in imagined vowel tasks; to create an imagined speech database with 50 subjects specialized in imagined vowels from the Spanish language (/a/,/e/,/i/,/o/,/u/); and to contrast the performance of the CNNeeg1-1 algorithm with the DL Shallow CNN and EEGNet benchmark algorithms using an open access database (BD1) and the newly developed database (BD2). In this study, a mixed variance analysis of variance was conducted to assess the intra-subject and inter-subject training of the proposed algorithms. The results show that for intra-subject training analysis, the best performance among the Shallow CNN, EEGNet, and CNNeeg1-1 methods in classifying imagined vowels (/a/,/e/,/i/,/o/,/u/) was exhibited by CNNeeg1-1, with an accuracy of 65.62% for BD1 database and 85.66% for BD2 database.

Evaluation of the Architect HIV Ag/Ab Combo Assay in a low-prevalence setting: The role of samples with a low S/CO ratio.

BACKGROUND: The Architect HIV Ag/Ab Combo Assay, a fourth-generation ELISA, has proven to be highly reliable for the diagnosis of HIV infection. However, its high sensitivity may lead to false-positive results. OBJECTIVES: To evaluate the diagnostic performance of Architect in a low-prevalence population and to assess the role of the sample-to-cutoff ratio (S/CO) in reducing the frequency of false-positive results. STUDY DESIGN: We conducted a retrospective study of samples analyzed by Architect between January 2015 and June 2017. Positive samples were confirmed by immunoblot (RIBA) or nucleic acid amplification tests (NAATs). Different S/CO thresholds (1, 2.5, 10, 25, and 100) were analyzed to determine sensitivity, specificity, and negative and positive predictive values (NPV, PPV). ROC analysis was used to determine the optimal S/CO. RESULTS: A total of 69,471 samples were analyzed. 709 (1.02%) were positive by Architect. Of these, 63 (8.89%) were false-positive results. Most of them (93.65%) were in samples with S/CO < 100. However, most confirmations by NAATs (12 out of 19 cases) were also recorded for these samples. The optimal S/CO was 2.5, which provided the highest area under the ROC curve (0.9998) and no false-negative results. With this S/CO, sensitivity and specificity were 100.0%, and PPV and NPV were 95.8% and 100.0%, respectively. In addition, the frequency of false-positive results decreased significantly to 4.15%. CONCLUSIONS: Although Architect generates a relatively high number of false-positive results, raising the S/CO limit too much to increase specificity can lead to false-negative results, especially in newly infected individuals.