Conventional machine learning algorithms are widely used in the processing and classification of biomedical data in the literature. However, in recent years, it becomes a necessity to use a multi-layered learning network due to the increase in the resolution and quantity of the data. The concept of deep learning comprises the whole set of applications that separate sub-classes with high performance, especially in the case of the existence of multi-dimensional data and insufficient human-decision making processes. These methods have the ability to examine the inter-departmental interactions of data in depth. Deep learning is superior to conventional machine learning methods because too many observational data are learned based on increasingly complex patterns. In addition, when the data size is enhanced, superiority difference over conventional machine learning methods is remarkably increased.
Electroencephalography (EEG) data can be expressed in high-dimensional, complex and spatial domains. It is suitable for analysis using deep learning-based networks as they have the potential to work with large datasets. Deep learning models developed for the analysis of EEG signals are expected to automate many repetitive cognitive tasks.
This review article covers the general introduction of deep learning architectures used in the field of biomedical and neuroscience, the applications of these architectures on EEG-based analytical tasks, possible difficulties encountered and possible solutions. As a result of a detailed study, it has been seen that deep learning methods extract complex patterns in the clinical or cognitive data and present a strong method to determine the abstract relationships between different types of data. 
This study addresses the needs of deep learning-based applications in clinical practice and neuroscientific research and aims to provide researchers with an overview of the deep learning-based analysis methodology of EEG signals and potential challenges in future research.

This study aimed to investigate the radiation attenuation properties of seven brain regions namely Frontal, Occipital, Parietal, Temporal cortexes, Hippocampus, Thalamus and Cerebellum in terms of their elemental compositions. Monte Carlo N-Particle Transport Code System-extended (MCNPX) version 2.6.0 (Los Alamos National Lab, USA) general purpose Monte Carlo code has been employed in order to calculate the mass attenuation coefficients of those aforementioned brain regions. A satisfactory agreement has been obtained on the mass attenuation coefficients (µ/ρ) calculated by MCNPX and XCOM for those brain regions under investigation. The results underlined that Cerebellum has the highest mass attenuation coefficients in terms of the radiological energy values. This can be explained by the elemental mass fraction value of Chlorine (Cl) in Cerebellum. It could be deduced that the intensity of diagnostic radiation can be more attenuated in Cerebellum than rest of the brain regions during the brain CT or brain PET examinations. The data from the present paper would be useful for the use of standard simulation geometry and mass attenuation coefficients for medical physics as well as the applications of radiation physics.

Wavelet theory is a widely used feature extraction method for raw electroencephalogram (EEG) signal processing. The nature of the EEG signal is non-stationary, therefore applying wavelet transform on EEG signals is a valuable process for extraction promising features. On the other hand, determining the proper wavelet family is a challenging step to get the best fitted features for high classification accuracy. In this paper, therefore, we focused on a comparative study of different Discrete Wavelet Transform (DWT) methods to find the most convenient wavelet function of wavelet families for a non-stationary EEG signal analysis to be used to classify mental tasks. For the classification process, four different mental tasks were selected to and we grouped each with another one to set dual tasked sets including all possible combinations. Feature extraction steps are performed using wavelet functions haar, coiflets (order 1), biorthogonal (order 6.8), reverse biorthogonal (order 6.8), daubechies (order 2) and, daubechies (order 4). Later, a specific feature reduction formula is applied to the extracted feature vector. Generated feature vector is then split into train and test data before the classification. Artificial neural network was used for classification of the extracted feature sets. From the result of the repeated analysis for each DWT methods, Coiflets performed relatively better compared to other wavelet families.