The mere presence of information in the brain does not always mean that this information is available to consciousness (de-Wit, Alexander, Ekroll, & Wagemans, 2016). Experiments using paradigms such as binocular rivalry, visual masking, and the attentional blink have shown that visual information can be processed and represented by the visual system without reaching consciousness. Using multivariate pattern analysis (MVPA) and magneto-encephalography (MEG), we investigated the temporal dynamics of information processing for unconscious and conscious stimuli. We decoded stimulus information from the brain recordings while manipulating visual consciousness by presenting stimuli at threshold contrast in a backward masking paradigm. Participants’ consciousness was measured using both a forced-choice categorisation task and self-report. We show that brain activity during both conscious and non-conscious trials contained stimulus information, and that this information was enhanced in conscious trials. Overall, our results indicate that visual consciousness is characterised by enhanced neural activity representing the visual stimulus, and that this effect arises as early as 180 ms post-stimulus onset.
Aging is considered to be a complex process in almost every species' life, which can be studied at a variety of levels of abstraction as well as in different organs. Not surprisingly, biometric characteristics from facial images play a significant role in predicting human's age. Specifically, automatic age estimation in real-time situation has begun to affirm its position as an essential process in a vast variety of applications. In this paper, two approaches are addressed as solutions for such application: prediction of accurate age and age group by using the two most fundamental techniques in the domain of deep learningconvolutional neural networks (CNNs) and deep neural networks (DNNs). In summary, this work can be split into two main key contributions. By applying a novel hierarchical aggregation built on the base of neural network developed from the training dataset, in the first stage, features extraction, the convolutional activation features are extracted from the captured facial image. As soon as this part is done, the features classification step is performed, in which Softmax Regression (SR) and majority vote classifiers are applied to predict accurate age and age group respectively. The effectiveness of the designed model was showed satisfactorily in the experimental results, which emphasizes the promising of the solution and indicates another direction for future development of algorithms and models in the field of machine learning.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.